SerialPower and > Seventeen thousand RFID cattle

#1
[ This thread was moved from the serial power and data thread in the finished project section - Hippy ]


Hello Jurjen and interesting forum members. I happened on your thread while reflecting on a project challenge before me. I joined your forum.

Your SerialPower (Remote device power delivery plus multi-point data traffic) concept seems a little like RS485 (or C-Bus) and this is in the part of my challenge.

The theme of my problem is this...

Say you have seventeen thousand (or more) cattle in a "feedLot", where they are contained in what are known as "home-pens" for the next one hundred days...during this time they are feed a special diet to progressively increase their body weight...(for the obvious reason).

Each home pen will contain say 100, near identical animals... a decanting truck loads a feed trough at one end and a water trough is at the other.

Every beast has an RFID ear-tag...this being of the 134.2khz HDX type (see http://en.wikipedia.org/wiki/ISO_11784_&_11785 ) for a brief description.

As you see there are a lot of pens and a lot more cows...this is a big undertaking

A "massive number" of "small" RFID readers will be retrofitted along the cattle side lip of the communal feeding troughs ...one place for each beast of the 100 beasts in each pen...in essence this is a parallel processing problem, as they all eat at the same time....

During this time the "feedLot" desire to know two things, daily:
1. are their individual eating habits regular...(an indirect analogue for animal health)...judged by time of arrival at the feed-trough and durant of their stay.
2. is the rate of weight gain appropiate for the beasts to reach a target weight and body mass composition (marbled or lean meat etc)...minutes eating.

If the beasts could be convinced to eat one after another in single file at one trough...the solution to monitoring "time and spacial location" would be trivial.

So, consider this...I have seventeen thousand "mini readers" and I need to know the ID of each reader and the ID of each beast visiting that reader and the date and time (in quanta of minutes) spent feeding. They come and go, and a spreadsheet will "smudge" the feeding recordings.

Typical of the inferences to be gained from the "spread-sheeting" of this data is the early identification of a "sick" animal should he consistently be last to arrive at the feed trough, reinforced by confirming he regularly uses a high numbered feed trough RFID reader... people who understand animal behavior will use the data in the spread-sheet as they will for what they want.

Do not dismiss this as a project that is not serious because of the huge number of reading stations...it is under serious consideration, think of it as perhaps not a classic case but an extreme case and possible docking station for your SerialPower concept.

It may well be the "two-wire" concept is unable to supply sufficient DC power to each distant reader...probably a heavy gauge "AC or DC power bus" will be required....this is not a problem it just means more copper. Would that destroy or cripple your concept...I think not.

I may well be that to address 17,000 RFID readers (up to perhaps 60,000 in a future time and place) will take too long if each were to be polled sequentially...

Remember that "real-time" information is not necessary and the information in each reader could be solicited and decanted over several hours while the cattle are asleep... even a baud rate of 2400 is too fast! ...try thinking down near 300 (revisit some earlier more primitive, yet less noise prone time?)

It seems each reader will need to hold a buffer of "one minute present or absent events) surely this is a easy thing for picaxe to do?...the real problem is decanting the day's eating records in a timely way.... rocket speed does not seem to be needed but clear thinking does.

I am far from being out of the woods just yet... there are issues concerning establishing the best "RFID detection volume" and relative orientation of tag and reader coils etc. I have sound experience in using and designing RFID cattle/sheep readers... my interest lies in capturing multiple abreast fast moving animals...data collisions come into play, however my theme is "the massive array" and chaos behavior ...this feedlot trough-reader project may yet be "just a bridge too far for me" , but we shall see....perhaps as my new friends you can conceive how picaxe can help.

Jurjen and fellow form friends...I will stop here, for a while and ask your thoughts as to whether SerialPower has the potential "horsepower" to be considered a possibility...I realize it is still a "work-in-progress".... but it seems you have made good progress...are you still enthusiastic to keep going?

Regards

Ray
 
Last edited by a moderator:

Paix

Senior Member
#2
@Ray, Welcome to the forum and sorry that it has taken me some time to respond. I don't have the answer to your prayers, but have a few thoughts to say the least.

I am intrigued by the logistics of your project. Can you help to fill in a few blanks to help complete the picture please?

Water at one end and food at the other suggests that mild exercise is generated for healthy animals as they alternate between eating and drinking. I appreciate that where there are large numbers of live stock, that there will inevitably be incidences of potential dead stock. Early identification and analysis can mitigate that problem and thus reduce a significant cost factor.

Taking your figure and moving it to 20,000 will better scale to the projected 60,000 for our mutual convenience. If drowning, few more litres of water won't make a great difference to the outcome.

HomePen = 100 beasts, 100 feed stations and 100 RFID mini readers. No favourite chairs, but places for all?
FeedLot = 200 HomePens (and rising).

How many hours in the day are available for feeding?

The HDX RFID tags means a 112 bit tag message which is 14 x 8bit characters. Can you point us to details of the RFID mini readers in question. (I know nothing about RFID readers, but don't tell anyone). How many bytes of information from the

How long does it take to capture a a tag reading? If readings were taken every 15 seconds then if you could provide clean reads, it would presumably have to be a fairly nimble beast to swap seats before the next reading interval. So, 400 reads per minute or one every 150mS on the local bus. Would you say two consecutive reads to indicate the onset of feeding? You might be happy with that down to 30 seconds between readings which would give 200 reads per minute or one every 300mS or nominally 3 reads per second. At 2400 baud thats around 80 characters positions per message.

If the reading from the mini reader was in the form:
Reader ID, Tag ID
It would probably work out somewhere between 20 to 24 characters worth of binary data at 2400 baud, so a reasonable amount of time for additional processing.

I would record the time in one minute increments and record that and a serial number, which is faster to increment than seconds and more likely to be unique, while preserving the sequence of readings. Reset either each minute or each hour and stored as a word variable.

I would be inclined to run separate power and a pair of of MAX487, MAX1487 or similar 128 node capable RS485 interface chips on each node. If a suitable Full duplex chip can be identified, that would cut the chip count, but the logic is that to save signalling turn around time, the RFD mini readers should be addressed by a dedicated PICAXE 08M2 transmitting a sequence of reader Ids at the rate of around 3 per second. The receiver would give up it's data (or not) upon receipt and a separate system would be independently reading the returned reader/tag data and prefacing it with Home Pen ID, Time and Serial information prior to popping it into storage.

That takes care of one Home Pen and almost puts the scope of things into context.

Later in the day, each Home Pen logging receiver could be put into an overnight mode and await a request for the composite data of the day. The Feed Lot administrative controller and receiving system having determined the date to prepend onto the data from each Home Pen on the Feed Lot.

Any minor discrepancies in clock time are restricted to differenced between individual Home Pens and so are both relative and of no consequence.
There is a lot of devil in the detail of course and doubtless much cleverer people than I will have a few thoughts about the actual communication, but that's my two penny worth of an outline.

The scheme can be seen to scale with established network performance using proven RS485 transceiver devices, which have essentially been developed incrementally over the last 30 to 35 years, but are not able to run power over the data lines. In fact I'm looking at a six wire solution, two pairs and power.

The question is, does beef play nice with high tech devices?

Ian
 
#3
Hello Ian...thank you for responding...

I had a sneaking feeling my situation was like one who came out to see Hayle's comet, only to find it passed by one week ago!

There was always the possibility some of you had "gone to god"...I am pleased you have the gist of my problem and have deciphered my cryptic description.

Let me try to fill in some blanks as best I can ... let me do it in small "chunks"

Yes, cattle do die while in the "home-pen"....not often though...they are well pampered and cattle handlers often do a tour of inspection.

Obviously if one cow fails to "come over" to eat, then electronically a red-flag is to be raised...he has either "jumped the fence" or is lying dead or near so.

As each beast can be worth up to $1000, it is a good thing to "go and get him as soon as possible"...this is the function of the trend line in the feed-time "clock-in" spreadsheet...ie the compilation of the data streaming down the SerialPower/485/? pathway.

The urgency in identifying "sickness behavior" is pivotal to optimizing feedlot profit as an early sick cow can "dribble" into the feed-trough or water-trough and before you know it the "black-death" has infected his remaining 99 companions....not a happy picture.

As to your water reference, yes some do die in the water trough...all a bit messy... I understand your water metaphor.

While there is "no reserved chairs" for that special guest, it is a fact that consistently the fittest or dominant or queen cow is/are first to run over to the feed trough as the feed-truck" begins decanting....conversely the weakest cow, be he the runt or just not well will be shunned or bullied by his peers and jostled to the last seat at the table... bizarre stuff this animal behavior...almost predictable. The composing of the analyzing software would alone be an interesting aspect to this problem...still.

Concerning feed times, there is one session in the morning about 6-7 am, it lasts about 1-2 hours: and another session at about 3-4pm...then they wander off and commune with friends etc. In between feeds, another truck with a rotary broom comes along and sweeps up the dregs (and spittle)...on and on it goes.

You can see there is roughly between 6pm to 6am where the feed trough is dormant....plenty of time to "slowly retrieve" the day's data logging...???

Concerning the HDX RFID process, it is a two step activity...for typically 50milliseconds a charge field is emitted from the reader loop-aerial...inside each ear-tag is a 35mm diameter pickup coil and a crude form of diode detector (like dare I say a crystal-set"...hmmm, this dates me)...

any how a capacitor is charged up and this will supply the power to let the tag send back a reply.

At the termination of the "power field" a circuit inside the tag senses this "end of pulse event" and says ok... now is the time to reply with this particular animal's unique serial number...

having allowed about 3milliseconds for the resonating energy in the large stationary coil/aerial do exponentially die away (critical damping = highly desirable here) before it reverts to its receive mode....it now waits to "hear" the distant (about one-metre away) ear tag reply.

The reply comes bask as a "Frequency-Shit-keying" stream of data...this reply carrier jumps between 124.2khz and 134.2khz to represent digital one and zero.
There are 16 cycles of 124.2khz then 16 cycles of 134.2khz to represent one and zero.

It take about 20millisecond (depends on the composition of ones and zeros in the returning telegram)...but all cows are done within 20millisecond.

The clever little chip inside the tag now short circuits the "crystal-set-storage-capacitor" to stop weird behavior (it counts the cycles passed and says "now dump".

So in answer to you, it typically takes 50+20 milliseconds to "read a tag"..... from this you will see that the prf or read rate is about 1/70 = 14 per second.

Concerning data collisions, you will understand that if two tags are in the "detection volume" then both will reply (almost) at the same time...in which case the distant receiver gets confused as the ones and zeros intertwine...oh and yes each tag is CRC tested for validity....and this just gets so confused that the end result is that "no-cow" is deemed to be present....this is not a good thing...hmmmm...it is another problem that pickaxe will not easily or ever solve...no worries, there are other ways.

Let me stop here for a while...am I better refining the scene for you?....it is a long story and a fairway yet to go...let us not early jump into the land of code, far better to define the problem that seeks attention.

Thanks
Ray
 

Paix

Senior Member
#4
Hello Ray,
I'm certainly getting there and hopefully gleaning the information that others will need. I'm a lightweight, but obviously have my areas of expertise. An enquiring mind is probably a good start. Telegraphist, Radio Amateur and Programmer.

I think that you might be wise to ask Hippy if he will move these posts to a new thread as otherwise we are likely to cause pollution to Jerjen's long standing "SerialPower" true two-wire data+power network thread. I have great sympathy for his cause, but feel there are enough imponderables without adding unproven capability that may not scale. "RS485 on the Hoof"

I'm off to bed now, before the sun comes up, but will be back later in the day. I see that in fact the tag reading phase is really restricted to two periods in the day totaling not more than five hours. So there should be considerable scope for having data to act upon within a couple of hours of the end of each feeding session.

I had been very optimistic about the RFID readers features and hoped that they could be triggered remotely, rather than a free running transmit / receive cycle and that a captured tag reading might be stored awaiting interrogation of the reader output. I did say that I had been rather optimistic. I rather gather that your readers are bespoke, rather than commercial kit? Can you describe the standard line in/out interface of the device for me?
Is it at TTL or RS232 levels.
A PRF of 14 interrogations per second, I can understand that when the CRC check fails the reading is presumably discarded? what's the standard way of discarding valid duplicates, as you may only want one valid reading for each beast in a fifteen second period reaching the data bus?

With a range of about a metre I guess that two or three tags could be responding. My initial thought had been that the readers might be able to activated in sequence so that only one was on air at any given moment and that the volume coverage might be somehow restricted to the vicinity of the diner's appropriate ear, say one to two foot horizontally. This of course would result in a very low and probably totally unrealistic PRF. I have a lot to learn I see.

What is the standard practice with a large number of readers in this sort of proximity? In an ideal world only one receiver would be actively transmiting/receiving in any one time frame. Multiple returns are one thing, but getting blatted by the next door reader's transmit phase wouldn't improve matters.

I shall have a sleep on it and think some, then hopefully read your responses to my additional queries.

The water metaphore, had to do with me drowning, not cows. I can't count to 170,000 so it doesn't really matter if we rounded it out to 20,000 was my point.
I look forward to hearing from you. Keep an eye open in case the thread moves.

Ian
 
#5
Yes, thank you Ian....I will try and "find" the moderator for this topic ...hello Hippy are you there?... shortly I will try navigating around this site and find/speak to you separately.

Ian (Paix) has wisely suggested we move off this thread of Jurjen Kranenborg...

While his concept of "SerialPower" speaks to my needs, it is an imposition on my part to contaminate his good works.

Would it be an inconvenience to you open up a thread for me in another place....hopefully along side or near to Jurjen's work.

Put me in a corner somewhere so I can talk with some interested forum members...

I regret there will be more words exchanged before we get down to the business of code writing and as such it may be a boring discourse for some who have other needs.

So will you relocate me to a new thread and give it the name as Ian(Paix) suggests.... "RS485 and RFID on the HOOF".

ps. Is Jurgen still available to talk with?

Ian, will you forward this to Hippy in case I fail to find him.

Thank you
Ray
 
#6
Hello Ian, allow me to fill in (or paper over) some more more matters that may help in forming your considered opinions ... already I see we are tracking parallel even now.

Have a look here to see the environment to which I talk. http://www.iapnutrition.com.au/nsw_feedlot

Regarding the RFID "Ear Tags". the term HDX stands for Half Duplex...it listens (charges up it's crystal--set battery!) for 50mS then FM transmits the animal code for 20mS.

This HDX is now adopted in Australia as the standard... It was a wise choice on the part of deciding powers here to opt for this system...contending systems have problems, but were cheaper to manufacture...the rest of the word will eventually follow in Australia's footsteps on this matter

There is/was a competing concept that was a full duplex concept, where energy is withdrawn from an illuminating high flux magnetic field once per cycle of the incoming 134.2khz "charge field"...and then again "crystal-set" style charging of a now "smaller storage capacitor" which is again the effective "battery power" to run the electronics inside the tag...

At the end process, the unique digital animal code switches a resistor in parallel with the 35mm diameter Tag's pick up coil...and this mimics a queer form of amplitude modulation...the incoming 134.2khz flux field is robbed of energy at the digital rate (ie resistive loading absorbs or "loads" the source of the incoming field).

First, they are not crystal controlled; and the FSK modulation is obviously modulated by a digital number "burnt into a Ram-chip"...this is the unique animal number

Second, the FSK of HDX is achieved by switching an additional capacitor across the resonating "pickup/transmitting 35mm diameter coil...no capacitor = 134.2khz, with capacitor = 124.2khz....(beginning to sound a bit sloppy...yes/no?)

This may surprise you as being just a little primitive...no worries here, as it is easy to FM demodulate two widely displaced frequency shifts (134.2-124.2khz)
Remember a narrow band deviation of say +/-100hz would require a lot of precision...narrow band FM would be good for all the classic reasons of noise and detect-ability at a distance... however, we/they were not designing the RFID tags like a moon-landing exercise...this is an agricultural theme and "millions and millions" of these tags must be reliably made...there is a lot to be said for the "keep-it-simple" axiom.

To dwell just a tad longer on the HDX concept with its Frequency-Shift modulation, there is one secret ingredient and that is the phenomena call "FM Capture Effect"....this is the elephant in the room...the jewel in the crown...this is the "thing" that wins the day over the cheaper opposing AM modulating concepts....

Your background in Radio will let you understand and appreciate it...sorry Ian , I wax lyrical at times...this is a beautiful thing....it works to mitigate data collisions when two or more RFID tags are in close proximity...it works for us, and needs no binary horsepower, it need no batteries or wires....it is a fundamental and works whether you like it or not.

So, because the ear tag uses a "simple" LC resonant circuit for transmitting and receiving, it is subject to issues that cause de-tuning such as mechanical abuse of the tag (by the cow crushing it against a fence, or the person who with pliers inserts the tag into the cow's ear or...whatever) the resonating frequencies for coming and going can be...well all over the place, +/- 1-3 khz de-tuning is not totally uncommon....however tag manufacture is/are getting more robust.

A last issue to set the scene for you again relates to the free-running LC oscillator concept in the HDX system, this concerns detuning due to the presence of metal rails or gates or motor cycles or...well any metal thing that was not there at time of installation...

Some, perhaps most, come to think of it, most Fixed Tag Readers employ some form of "auto-Tuning or Auto-Resonating scheme" so as to ensure maximum circulating
current in their large aperture search coil ...remember the aim is to ensure lots of "magnetic flux" to charge up the little ear tags crystal set.....

While I talk of aerials and coils or antenna, in reality this whole tag concept/system is about loosely coupled tuned circuits....your radio background will again forewarn you of the consequences when tuned circuits come into each others spheres.....again it's all about basics...there are a lot of balls in the air here...

Sorry for wandering off the subject of conveying the digital message to some remote place...(and hopefully by way of the picAxe tool)....but these other matters must not be ignored as they vector straight to the heart of the problem at hand.

For example, to close here on, there is the issue of spectral electromagnetic contamination from "wires" carrying high-speed digital switched currents when in the vicinity of micro-volt sensitive 134.2khz receivers....highly desirable not to be sending RS485/SerialPower/etc signals backward and forward during the 20mSeconds windows when the receivers are "straining their ears to listen".

By the way, all seventeen thousand feedTrough readers will be synchronized, that is they tun on and off together ... whilst the individual readers are not required to be carrier synchronized (well not just yet any how)

Ian, we use TI readers see www.ti.com/rfid .... these are the industry bench-markers...but were never meant for such uses as we here invisage, they are best at single file animal RFID reading, and do it well...more information to follow.

We have a friend here who will work with us to design a more useful reader package http://www.priority1design.com.au/ , and perhaps it will contain the pickAxe and perhaps the work of Jurjen's SerialPower.... who knows, its still early days and my presence here with you is one aspect of my research....

Your and other members thoughts are gratefully received... I have no intention to rob you of intellectual property, let me be open with you as a kindred spirit.

Regards
Ray
 

SAborn

Senior Member
#7
Are you using the TMS3705 chip from TI or the complete TI boards.
If using the TMS3705 and wish to use the picaxe to decode the data and transmit the data, the problem is the picaxe is just tooooo slow.
I have worked with the TMS3705 and know first hand the picaxe will not cut the mustard for this application, a native Pic is about the limit of what will be needed.

Have you looked at sheep feed lots, as i know there is several in Victoria (and elsewhere) that are using RFID already, and to the point of weighing the feed and monitoring the amount eaten to weight gain (some eat a lot and gain little weight so are culled early)

What antenna system will you be using, as you quoted a 1m read range, this will be a reasonable size coil i would expect, and require around 1000 volts within the coil to have a 1m read range, its a lot of energy over that many coils, and what is the regulations on transmitting that amount of energy.
I know the Australian standard (and NZ but not all countries) will allow up to 1000v within the coil but not sure on mass use of coils in a small area.

What is the total length of transmittion of data, as this will be one of the biggest problems.

This is certainly a major project, and as much as i like the picaxe i dont feel its the right chip for the job at hand here.
The problem some picaxe chips have with sending serial data and timing issues would make this project a nightmare, as it may require calibration adjustments made to almost all chips to eliminate errors, and for this reason i find the picaxe has a weak link in this area when mass chips are used in repeat circuits, as the last thing you want to do is having to adjust each chip to suit, compared to just plug the programmed chip in and it works.
 

Paix

Senior Member
#8
Hello Ray, I was doing a little research about Australian Beef Feed Lots this afternoon, but you have since given me a lot more information to think about.

What I had picked up on from the a couple of posts ago when you first mentioned data collisions, was that the conversation related to multiple tags and a single reader. In my researches this afternoon I noticed that the standard was somewhere between 500mm and 800mm for a beast at the trough, and so on a bad day there was going to be a lot of transmitter overlap and if you take just half a dozen FM transmitters on the same frequency 800mm apart and key them all at the same time, you will get not so much FM limiter capture effects, but a God almighty doubling that might even confuse the tags, never mind the receivers in the readers.

I appreciate your RFID 101 class, and am learning a lot fairly fast, but have to smile from time to time at the degree of spoon feeding. We are writing to a wider audience, so it's no bad thing. I think my sense of humour is probably about as dry as your own, so we at least are on the same wavelength – part of the time!

You mentioned synchronising the readers and you also mentioned that RS485 or SerialPower communication interference would knock about the signals if it was passing data during RFID interrogations. That got me thinking quite a bit and I have to agree with you that it would be a bit of a liability.

You then provided me with a lot of background reading regarding the environment and the details of the Priority1 website as well as the TI chip in use.

SABorn came along and made a lot of sense and I have to agree with him, but hope that he has misread the criteria slightly.

Now, with the apparent communications embargo during breakfast and dinner, and a quick look at Priority1's offerings, I hope that you are considering the RFIDLOG device as a choice of reader/logger. It simplifies matters a lot, but does perhaps introduce a few minor niggles. The niggles being that each unit is it's own timekeeper and so timing errors are a possibility, but I notice that the time can be set from the comms interface so if they were all set at the same time on a regular basis there should be no problem. It is noted that despite being able to make 7 readings each second the log has a granularity of only one second, so cows won't be recorded as arriving by less than a hair's breadth.

Not unlike SABorn, I was thinking of a non-PICAXE solution for my a logging receiver, for the same reliability reasons that he stated. At modest speed there is no problem, but I think of serial comms in terms of communication links that are working near saturation at 'machine speed' and processors just handling everything else as you do on a PC for example, but it seems that short burst message passing is the norm in some micro controllers due to limited processing headroom.

SABorne was also concerned that you were trying to mate a PICAXE chip with a raw RFID chip, which was as he states, well beyond it's capabilities.

Can you verify the RFID situation as to exactly what device you are looking at utilising? That way we will be nearer to nailing down the scope.

Also, when you speak of synchronising the readers, can you clarify the method you are considering to do this? Were the SRD and SRA commands what you were thinking of for this? If you had interleaved groups of three, four or five readers, you would be able to enable and disable groups so that one third, quarter or fifth of were active at any one moment, in three or four second slots?

Set Reader De-active Command.
This command is used to turn off the RF field emitted by the RFIDLOG. Tags can not be scanned while in this mode. This mode is used to prevent the RF field of the reader from interfering with other readers in the vicinity.
The command protocol for which is shown here, along with the available responses.
Protocol: SRD<crn> ,where <crn> is $0D carriage return.

RFIDLOG This is good and looks manageable.
http://www.priority1design.com.au/rfidlog_rfid_data_logger.pdf

I notice that it is flagged as sold out. It certainly seems good value if you need to keep the comms relatively quiet during tag reading periods and so need on-reader logging facilities. The PICAXE side of it is then essentially one of reader control and a suitable 'data concentrator' per Home-Pen, in the first instance.

What do you use at the moment to glean this sort of information if, you do?

Do all the beasts have the relevant tag on the same ear?

What do you anticipate would be a good initial proof of concept pilot, four or five readers at final spacing on a bus, with a 'data concentrator'?

When you said that you had 17,000 readers, did you mean you HAVE them as in, you can kick them? Not just notionally?

The expected footfalls here are a lot lighter than I would have expected. The clever guys are keeping their heads down. Hopefully only until the scope and viability are seen to be entirely in range.

/Ian
 
#9
Hello Saborn, thanks for speaking with me...I see you are from Adelaide, you will know the Mount Gambia sale-yard ???...lots of RFID there...also the Bendigo sale yard where up to 40,000 sheep will be sold on any one day....

Sheep RFID reading is "harder" than cattle as you will know....sorry I take that back a little, for sheep in single file is an easy thing (such as on a farm); while sheep 3-5 abreast running down a saleyard alleyway is a different matter.... this later instance is where I dabble....a whole new set of rules apply.

We use various TI readers RI-RFM-007 + RI-CTL-MB6B + RI-STU-MRD1 for example....you will understand they are well suited for single file detection...in addition nearly everyone uses the point to point RS232 data route back to the remote host....there are tricks one can do with RS232 to get say six-eight readers to communicate ove the RS232 cables...it is messy and poor....yet can be done.

While RS485 is available in the control module (RI-CTL-MB6B)....it is rarely done...in one past case we ran six RS232 cables back to the host and used a commercial 6-8 RS232 input card....it works well....but in huge numbers even conventional RS485 will not easily "handle" more than say 60-100 remote stations....yet this could be relieved as it is an issue of line loading...an interface board would need be designed even optical fiber may need be considered?....later...later

It is rare to meet someone who has used the TMS3705, if memory serves me right I think it is used in the RI-STU-MRD1 hybrid????.

I agree with you that the pickaxe would be pushing it uphill to decode and CRC test each telegram and do the communication back to the host....

There a a couple of streams of thought on how to proceed...(remember we are prepared to start from scratch and design a "turn-key" solution) one more interesting possibility is to AD convert each telegram, hold it in buffer, and at midnight send "basically a mimic" of the raw unprocessed data back to the host and let say VisualBasic sort out the mess.... this may be where pickaxe (slowly) fits in ????

I like the picaxe concept, as most people can eventually learn it or at least the basics...I don't believe it will go the way of the "do-do" bird...this forum is proof of a vigorous knowledge base that will stretch into the future.....this is a good thing.

So

To be truthful, the thing that caught my eye about "SerialPower" was the intention for it to a become adaptable...almost like "plug and play"...an ability for the system to have guests introduce themselves down the bus and be allocate a "find-able time-slot" has great merit...consider the case of replacing reader station number 799 and not having to code into it duplicate a fixed address name...just rip it out and put any virginal new one in and let the system introduce themselves to each other....great idea....good thinking..very practical.

So SAborn, the pickaxe may fail the speed test ... I was not aware of timing issues or calibration adjustments....hmmm, hmmm....I understand nightmares even the business of "openCom and closeCom leaves me cold....and yes I have experienced nightmares with RS232-to-USB converters, which in some mad fit, grab any com port they want and leave you wondering (where did you go???).....

SAbourn, my heart just skipped a beat...were you referring to "no two chips are the same" ?????....surely...surely this cannot be...are you telling me some are left-hand drive and some right-hand drive...is the variance so great that one must fine tune every circuit to achieve conformity....please tell me it is't so....

Any how, let us not stop just yet.... there is a long way to go...will you to walk with me away?







While I am a little familiar with visual basic and Lisp and fourth? I do not want to learn any new language (getting too long in the tooth...its a young mans game) ...at least with picaxe basic I can understand the questions to ask and hopefully recognize the software solutions (with your and other forum members help)

Hello Ian...thanks for the relocation...talk to me more

Regards
Ray
 

SAborn

Senior Member
#10
there is a long way to go...will you to walk with me away?
As long as you dont expect me to hold your hand as we walk.:mad:

I have found and others have tested this too with the same results is some picaxe chips are not 100% in synchronization with serial and require the internal clock frequency to be adjusted or garbage is sent along with the data corrupting the data package, this is ok for 1 or 2 chips in a project as we can take the time to calibrate them but when doing large repedative copies of the chips it is a real pain in the ass.
I have banged on about this in the past and so have others here, but there is always an excuse and not a active solution to fix the problem, for this reason i would take care designing something around some of the picaxe chips with large numbers of chips doing serial transmits of data.

The picaxe is a great chip for development work and projects where speed and data reliability are not an issue, but if it was me and with the numbers you are quoting i would develop it with picaxe and then translate the program to Pic, and use a pic chip in the final product, as once designed all the chips would be the same, and 17000 of them at a far cheaper price and higher reliabilty would convince me to use the pic chip.

Yes i know Mount Gambier well as i grew up not far from there, although its been some time since i have been back of late.

The problem you are going to have with data dumps once a day is the need for each reader to store the tag numbers, and you would need eeprom to do that (guessing its what you are calling a buffer), then it gets limited to the eeprom storage ability, as each tag read needs to be checked against tags stored in eeprom and not repeated or you would fill the eeprom with 1 tag number, as the number of tags increase in eeprom the process time slows down due to the number of checks needed to be done.

It would be better to write the tag reads to a PC where you have processing power and speed, then the tags can be stored with a time stamp, so it can allow some filtering of repeat tag numbers within a time frame.

RS485 is not somethink i have had a lot to do with so i have no real comment there.
 
#11
Ian .....................................................................................................................
What I had picked up on from the a couple of posts ago when you first mentioned data collisions, was that the conversation related to multiple tags and a single reader. In my researches this afternoon I noticed that the standard was somewhere between 500mm and 800mm for a beast at the trough, and so on a bad day there was going to be a lot of transmitter overlap and if you take just half a dozen FM transmitters on the same frequency 800mm apart and key them all at the same time, you will get not so much FM limiter capture effects, but a God almighty doubling that might even confuse the tags, never mind the receivers in the readers.

Ray..........................................................................................

Yes Ian, I was relating to a single reader and multiple tags

Yes your research is correct about the spacing between the animals heads being about 500-800mm though I am using 300-600 as more representative...this mimicking a 2:1 body mass ratio as the cows grow...another reason will arise shortly and this will relate to &#8220;nudging or forcing each cow to adopt a preferred eating position just above each reader loop&#8221; ...we had to find a way to force each cow to sit in the best place.

Thank you for going to the trouble of searching this out....I appreciate this.

It is highly desirable that the cow &#8220;sit on the chair provided....not in the space between the chairs&#8221;.... sorry Ian my puns are poor and it is not my intention to be 101...however its late and humour helps keep me sane.

Concerning your 800mm apart cattle and causing every transmitter to &#8220;fire in synchronism&#8221;....yes this is as I meant it to be...no good having one of many shouting when all the rest are intently listening...the chorus must transmit in unison.

Now, I have a suspicion that you misunderstand my reference to FM Capture effect&#8221;
You speak of &#8220;limiting&#8221; in this case you see hard limiting (clipping the tops off...removing amplitude variation...yes/no .....

Well I speak of another thing and this relates to vector analysis...it is a matter beyond subtle and yet you see it everywhere in your little pocket FM radio....have you ever half heard one station and half of another....ok I understand inter-station muting....this is not what I mean...your radio background may have shown you this phenomena, the stronger vector (Radio signal) takes control of the demodulation process....snap, its either all one or all the other....well close anyhow.

If two adjacent FM carriers be they on the same frequency or slightly different are together then the one that is about 3-6db stronger will dominate the demodulation...the weaker one is virtually &#8220;masked out&#8221;

Now in the world of RFID and the world of magnetic field intensity falling off as a recripical of the cube of distance away, this all means that if one tag is closer to the Fence-Reader than the other or others, then this one tag is effectively the only one seen....weird...beyond 101.....

In contrast two AM radio stations side by side will &#8220;always&#8221; incur a heterodyning or beat note which is superimposed on top of the two sets of information....at night time you hear this waxing and waning (sort-of)...this was another reason why amplitude modulation and its variants were/are poor seconds to FM modulating tags.

This FM capture effect was a key ingredient in Dual Frequency Instrument Landing systems for aircraft....I use this as a side reference in case it rings bells with yourself or other forum viewers...sorry, I&#8217;m slipping off the pickaxe track again....

The receiver in the tags is very insensitive it need hundreds of millivolts up to volts AC. Remember, when receiving the tag is only trying to suck up enough power to provide power for its electronics to spit back a simple telegram in the 20milliseconds of storage life it has...and then it cuts its own throat and dies a graceful death. (sorry again).

Communication or magnetic coupling between adjacent tags is no real worry to us, it is weak and at the wrong time. Should all this above be priori and well known to you...say so, and we can quicker cut to the chase.

So, Ian....when you speak of bedlam, when many cows are in the trough it is not violently so... and the management of the three dimensional shape of the detection volume is the tool that mitigates the problem you see... Remember this whole business is another thing...it is a three dimensional problem.

Ian, I am reflecting on your earlier words...they are good...and I digest slowly.

Kind Regards
Ray
 
#12
SABorn...........................................................................................................
The problem you are going to have with data dumps once a day is the need for each reader to store the tag numbers, and you would need eeprom to do that (guessing its what you are calling a buffer), then it gets limited to the eeprom storage ability, as each tag read needs to be checked against tags stored in eeprom and not repeated or you would fill the eeprom with 1 tag number, as the number of tags increase in eeprom the process time slows down due to the number of checks needed to be done.

Ray.................................................................................................................

Thank you SABorn, even crystal control of n-number micros will fail a timing test...worse still simple RC oscillators or a little bit better are ceramic resonators.....on top of this I feel you allude also to rise and fall times variations between chips and variation in thresholds due to temperature etc etc.

To get and keep all in “lock-step” is no easy thing as the numbers rise....you thoughts a well taken.

A remote master clock seems to have merit together with variable phase delays introduced for more distant processes, ie to allow for cable propagation delay etc.

There is a time –keeper in the sky...called GPS...perhaps it is a source of synchronization....just a though...just a thought.

You make reference to comparing the last reading with the present reading and only making a new recording if they are different....as you will know this is one mode already used by TI. ....where cattle sheep are in single file this is a good system as it is pointless to use up valuable reader-to-host time telling it over and over again the same thing “this is cow 999, this is cow 999, this is cow999 ........this is cow 999” once is enough in this case. I see you pre-empted me here ...thank you.

A more interesting case is in the Abattoir, where the approaching carcass is not/never going to be overtaken by the one behind it... once is enough, I have seen this cow and know it exists ...stop repeating.

However the application I came to you with is different; sure I want to know who the cow is, I equally want to know how long he has had his head inside the trough (hopefully eating).

You see I need the start time and end time, that is the differential of the Durant of his stay.... the above recursive concept is not what I want yet I don’t want to know every 1/14 of a second who he is....a balance is needed....do you see the difference.

SABorn, yes I am still “BlueSky Dreaming” and no we don’t have 17000 readers nailed to a trough (yet) though we will shortly have 50 at our experimental site....issues of interference and hardware interconnection are being resolved....plus my immediate activity is the shaping and confining of the flux field at each reader station...it seems highly likely the coils will be etched onto large PCB sheets for repeatability....repeatability with pickaxe now concerns me however let us keep going, I value you thoughts

Regards
Ray
 
#13
SABorn...........................................................................................................
What antenna system will you be using, as you quoted a 1m read range, this will be a reasonable size coil i would expect, and require around 1000 volts within the coil to have a 1m read range, its a lot of energy over that many coils, and what is the regulations on transmitting that amount of energy.
I know the Australian standard (and NZ but not all countries) will allow up to 1000v within the coil but not sure on mass use of coils in a small area.

Ray............................................................................................................................................................

Hello SABorn, you are correct in that the AC voltage across the reader coil can go to somewhere near 1000v peak.... this being due to “Resonance”... allow me to explain this for others ... it may have perplexed them to hear of such dangerous levels.

The typical industrial HDX reader requires a Q-Factor of up to 200 so as to achieve high circulating current....this being so because the need for “lots of” emitted or magnetic flux density.

When striving for high Q inductors at these frequencies, litz wire or its variants are commonly used...this concerns the issue of “skin-effect”.

As well as being dependant on coil area, flux density is directly proportional to the magnitude of the circulating current... and we want this current to be as large as possible...remember this whole system is analogous to a loosely coupled transformer.

Indeed the circulating current can get up to tens of amps at 134.2khz (if you let it) when this happens the peak AC voltage across the series resonating capacitor will easily reach to +1000V

Yes the systems use series resonance, therefore the driver is approaching z a ero ohms source impedance...sometimes a push pull output stage uses a step down ferrite transformer with 1-2 windings on the secondary...from a 24v DC power-pack huge circulating currents around the fence mounted reader coil can be had..... and yes, sparks can be seen when things go wrong....

To confirm your concerns SABorn, I recount an instance we sprayed a coil with carbon paint, to provide electrostatic Faraday shielding...the shielding worked well and still does now, however on one unit, the insulation of the winding wire was “nicked” before the carbon spray was applied.....yes you guessed it, after sealing and closing up that coil....it...it ...caught fire on its first use.....your 1000V found a home... and my face went red......

Safety is one of the “balls in the air” that need be minded....there are also regulations concerning EIRP or allowable emitted power...perhaps more on that later.

But before I leave this topic on safety, let me give you a more frightening instance...this concerns “heart pacemakers” and also under the skin “electronic shock pain blockers”...under no circumstance stand behind or against a large RFID reader if you have one of these sort of implants... we warn people of this, and you will understand why.

Thank you SABorn...I am curious as to why this topic even entered your mind...will you expand on an experience....

Ian, yes all cows have the tag in their right ear....however if a dog or “mad kangaroo has bitten off his right ear then there is nothing for it than to put another in his remaining left ear ...(sorry, I lied about the kangaroo....I meant red-back spider.....it’s getting late , I must go)

Thanks to both of you
Regards
Ray
 

SAborn

Senior Member
#14
For anyone who would like a little light reading of the animal tag protocol in use here than it can be downloaded from here.

http://ozelecforum.com/viewtopic.php?f=31&t=463

I posted it on the forum some time back after all the trouble i had with obtaining a copy for myself. (im sure Ray has his own copy)

@Ray.

Yes i understand about your need for multible reads of a tag but its only the first time stamp and tag number needing to be stored, along with the last time stamp, so hence its just the time stamp that is upgraded after the first tag read, for within a given period.

Example..

tag 999 is read and saved as 999.....6:00am - 6:12am

It would just be the 6:12am value that would be updated while active tag reads of animal 999 was recorded, saving huge data storage space and far less to sort through in data files.

If there was a break in 999 readings for say 15 seconds (a given time period) then a new reading log for animal 999 would be started when detected again (maybe it left for a drink or a pee and returned to the feed trough)

Now cows sitting on chairs provided..........i would like to see that...........be more like stools in a bar room brawl i would think.
 

Paix

Senior Member
#15
Hello Ray; SABorn,

The FM capture effect: I believe it is a phenomena more to do with swamping limiters rather than the discriminator circuitry. A quick check suggests that it can also occur in the discriminator in limiter-less designs, but it is indeed one of FM's great selectivity features. I regularly hear the picket fence doubling effect in the demodulator when FM amateur stations of similar strength accidentally key up together; rather than enjoying the benefits of the capture effect. I thought it worth mentioning, but our individual experience varies. The wide shift 10kHz and AFC in the receiver helps to mitigate any small frequency inaccuracies brought about by the environment (one hopes!).
- - -
We appear to have different ideas regarding reader syncing.
You appear to favour all transmitters illuminating the tags at one time and then everyone listens for local Tag returns while the transmitters are off. From what you say about line delays along the length of a cable commanding such behaviour, it is not totally straight forward. It almost begs to have one large transmitter illuminate the whole feed line at once, with it's attendant prohibitive technical and licensing issues. Simultaneous operation also infers cycle by cycle control or lock-step syncing, which you have already said is not a walk in the park.
I am dubious about the wisdom and practical issues of simultaneously keying 100 transmitters in close proximity. The sky makes a wonderful mixer if nothing else!

If instead you operate 1 of 5 readers in each two or three second burst of operation the spacial separation between emitters would give you a lot less in the way of intermodulation products. How much potential loss of data would there really be?

14 to 21 interrogation cycles, pause for 8 to 12 seconds, rinse and repeat. The readers only read following their illumination phase, so adjacent receivers are unlikely to be confused. In that 8 to 12 second gap the other four groups have had their time slice. Given that you are looking for a granularity of feeding in minutes, then sampling every 15 seconds should suffice to build a feeding picture I would think. This would also have the effect of lowering the average EIRP, as only 1/5 of readers would be active in any single period.
- - -
SABorn makes good sense when he mentions that beyond prototyping and establishing a proof of concept, that a compiled, rather than interpreted, solution should be sought.
- - -
RS485 versus RS232. I accept that you have paralleled RS232 devices and got away with it.
After my life as a telegraphist I moved into computer operations and then programming with DEC Basic and thence onto Ingres database applications in the quarrying industry (1988 to 200), then had four years with EDI on Unix systems. I had sites where the business analyst looking after the quarrying sharp end had serial ticket printers connected on &#8220;Y&#8221; cables, which frankly gave me palpitations, but I have to say that they worked.

RS485 on the other hand has push-pull transmitters and with differential receivers therefore it goes so much further.
The Max487 and Max1487 devices specified earlier (there are many others) are enhanced spec in that they are designed to allow 128 nodes on the bus (instead of the standard 32 nodes), which would typically be used to drive CAT5 UTP cable, but given that you have what might be called an electrically noisy agricultural environment (who would credit that, except for that kangaroo on smack that bites off bovine ears :). I would incline towards a Cat5 STP (UTP=Unshielded / STP=Shielded). For those who may say why not Cat6? The lower the Category number the thicker the copper conductors.
Linear Technology makes some very interesting RS485 transceivers
- - -
The interface output from the RFID readers hasn't been established and where the boundary actually lies.
I perceive that on reader data storage is required and essentially all that could be done with it once it is written to file, it is to make the dumped data available, as is, for squirting to a computer platform after the feeding period that has sufficient power to append data to records etc.
The point really is that a micro controller is not a computing platform.
Your scope is very wide and has to do with the whole RFID package. It is useful to know how it all fits together but strictly speaking other than being a sounding board, and I have no problem with that, my main concern is from where the PICAXE realm and line interface communication begins. What is the scope and formatted content of the data there and what controls and responses need to be generated and swallowed in return?

Is this link from Priority1 Design http://www.priority1design.com.au/rf...ata_logger.pdf significant at all. I somehow think that you are building custom readers?

I understand what you are saying about the duration of feeding and throwing away duplicates. How you actually accomplish that is of interest. Inside the reader, a single reading could be taken as representing a potential '15 seconds head down activity', with duplicates discarded until fifteen seconds later when another valid reading could recorded. Even better would be if a counter could be incremented on the original contiguous recorded reading and only the one record kept until a different tag is successfully read or no tag is read for a suitable time out period.

I'm expressing this badly, but it would minimise the number of actual records stored.
If I was making such a reader, then I would probably use two processing chips, one to control the emission, acquisition and CRC validation. The second to take the tag data, give it an initial time HHMMSS stamp and do the subsequent discarding of contiguous valid duplicates, incrementing the 15-second-sample counter and writing of final records to storage. Unless it was felt that a problem might prevent data being moved from the tags, then the date could be omitted or a day number used for brevity: XXX-tag-no-XXX HHMMSS 003\nl. So an RFID task processor and an admin processor. On being commanded to download this information from memory, the reader ID and a date as header and footer records. AR99999 YYYYMMDD and XR99999 YYYYMMDD
This is very bare bones and probably way off the mark. I don't presume to know what makes a good RFID reader I would pad record lengths to total 32 bytes to simplify memory storage. It would probably be written in C or assembler.

Forgive me for rambling, My brain has been going in circles, I will read and digest SABorn's RFID tap digest and doubtless come out of it either brain dead or a lot wiser, Don't put money on the outcome though . . . :-
)
 
#16
Ian..........................................................................................................

I am dubious about the wisdom and practical issues of simultaneously keying 100 transmitters in close proximity. The sky makes a wonderful mixer if nothing else!

If instead you operate 1 of 5 readers in each two or three second burst of operation the spacial separation between emitters would give you a lot less in the way of intermodulation products.

Ray..........................................................................................................

Thanks Ian , I believe you and SABorn are saying the same thing....regarding the interleaving in sets of say 5...that is readers 1-5 fire while readers 6-10 are disabled and readers 11-15 fire concurrent with readers 1-5....that is a gap between them to &#8220;speed up&#8221; the scanning in each pen

I agree with what you both imply...an extreme case of what you speak is to multiplex the whole 100 Readers in each pen and scan through then one at a time....

Let me &#8220;throw another log on the fire&#8221; now, and this concerns mutual coupling between adjacent feed-trough readers.

If you have say a middle coil with two &#8220;close&#8221; adjacent identical coils (one either side) these being series resonant circuits at 134.2khz....then the mutual magnetic coupling between them is certain to bizarrely detune all of them...yes/no?

So

It now speaks to say that when the one in the middle is being excited, the two adjacent coils should be &#8220;open-circuited&#8221; to therefore deny the action of mutual coupling....what I say to you is this

The pattern of interleaving will need to "miss" every second reader as those ones will be open circuited and therefore not available for performing work!!!! 9the design of the reader will need include this feature....

our discussions here are preliminary to the design brief...nothing substantial exits just yet...I am on the same thin ice you are here...interesting?

Sorry for throwing you from wall to wall so quickly....take a breather catch your breath if you can keep this up we have just begun

Ian...I sense a man called &#8220;Scroggy&#8221; in your word craft ...do you know of him?

Regard to both
Ray
ps... I will be heading back to Melbourne tomorrow, however I will watch out for you conversations
 
#17
Ian.....................................................................................................................

.... rather than enjoying the benefits of the capture effect. I thought it worth mentioning, but our individual experience varies. The wide shift 10kHz and AFC in the receiver helps to mitigate any small frequency inaccuracies brought about by the environment (one hopes)

Ray....................................................................................................................

Thanks Ian, my pocket radio metaphor was poor...

This business about &#8220;FM Capture Effect&#8221; is real and not explained by clipping or channel spacing or AFC....come to think on it, you and many others have probably have never experienced it...

I don&#8217;t mean to be precocious here, and will come back to talking more on it later...

sufficient to say, that if you hold two tags about 15cm apart, so they both present equal &#8220;amplitude&#8221; FSK return telegrams, then a &#8220;collision&#8221; occurs and &#8220;neither&#8221; tag is CRC validated... it is a &#8220;form of invisibility&#8221;

To all intents, neither tag exists (to the reader)... the intertwining of the carrier-vectors has a &#8216;phase modulation nature&#8221; this where you look to find the explanation.

Now

If you push one RFID tag about 10-15cm closer to the reader than the other, then....&#8221;bingo&#8221; only that closer tag is correctly crc validated and is deemed to exist..... (snap hello..snap goodbye)

The behaviour of two tags alternately exchanging &#8220;positions of closeness&#8221; is like two sprites which come and go...it is a beautiful dance to see.

For you to mimic this behaviour with a couple of flea-power FM transmitters will require you to disable any AFC or PLL processes and listen &#8220;in the raw&#8221; (as it were...)

Sorry, I can&#8217;t get off this subject just yet...in the world of &#8220;HDX FSK eartags&#8221; it is just so important... in our sphere of wireless/microprocessor there are &#8220;weird things like this&#8221;. I recount to you the tunnel-diode, it was for some time a solution looking for a question...in all my time I have only seen one once!! (Whoops, sorry).

Let me pre-empt something else here concerning spatial x,y,z orientations... between the Fence-mounted reader&#8217;s coil and the coil in the RFID...go back to the two side by side tags referred to above...now (in your mind) move one closer till it is revealed (detected), now rotate it through 90-180-270-360 degrees.

Can you see it? ...First you (the reader) see the closer tag then as it rotates it disappears and the one behind it suddenly appears... as you continue rotating the first tag now suddenly reappears again as the spinning goes on and on.

The reader now sees tagA then tabB then tagA....etc... the tags &#8220;occult&#8221;, that is one hides behind the other....I mention all this as it is part of the tool kit I use for mitigating collisions...it relies on chaotic behaviour

you will be forgiven for saying I am trying to make a silk purse out of a cows ear....and you would be right....existing readers were simply not meant for this application...nowhere in the world does such a ting yet exist (best I know anyhow)...your support is gratefully received...even if along the way the picaxe falls to a minor role...we shall see....we shall see

Regard to both
Ray
 

Paix

Senior Member
#18
Hello Ray,
I experience the capture FM effect probably every week. A distant net works on the same channel and the our net controller is not keen to QSY in case others can't find the net. Between breaks in traffic most of the members of the distant net are heard relatively clearly and of course when one of our net members is transmitting, it is he who is heard at full quieting. The situation is not quite as dramatic as you describe I have to agree, but don't discount my limited experience.

Otherwise I will have to get MG Scroggie to hit you with his famous text. The Foundations of Wireless. I bought that book in 1965 whilst still in telegraphist training. The book was sent to me even though the price had gone up in price from the old advert that I had used. I sent away the extra 2/- or 5/- promptly. I don't suppose that would happen often in this day and age. I can never claim to have read and understood the book and it went missing several years later. Every couple of years I to this very day I wonder just where it went and would dearly love to have it back. Although I wouldn't buy another copy now, even if it was available. My very first radio book.

I just had a quick Google search and although the book had been recommended to me by a Chftech instructor at the time, I have never before been aware of Scroggie's connection with either the RAF, his pseudonym of &#8220;CathodeRay&#8221; in Wireless World. (I used to read it and skip over all the formulae because I didn't understand any of it. Later some of began to make more sense) or Creed & Co. Creed having produced most if not all teleprinters that I have ever had any experience of. Bit I digress.

I appreciate, now, that the capture effect is particularly demonstrative in animal RFID reading operations. As for flea powered transmitters, Hmmmmm.

I also hadn't really given any thought to varying tag orientation and signal polarisation. With it's theoretical 30dB loss between antenna of opposite polarisations; horizontal and vertical.
I converse several times a week on a 2m VHF net with friends, but we use a colinear antennas which are of course vertically polarised. Often we are joined by more distant stations that have come across us and are using yagis of one sort or another. Often they are concerned about the lack of performance and so I now get in early and state that we are vertically polarised, which tends to explain a few mysteries. SSB, by convention, is normally operated with horizontal polarised antenna.

Coupling effects of adjacent coils/antennas B1.jpg I hadn't considered it, despite it being one of the major features of RFID Tag reading. I have obviously spent too much time in supermarkets recently. An investigation regarding the exceptionally poor propagation performance of a large RAF HF transmitter revealed a wireless technician who lived very close by with a large coil in his 'free electrically heated' green house. Had he been a civilian he might have avoided repercussions!

The grouping was actually 1, 5, 10, 15, followed by 2, 6, 11, 16 etc, not 1+2+3+4+5 then 6+7+8+9+10 etc. strictly interleaving rather than grouping I suppose. My groups weren't contiguous! In the same way that I had not considered de tuning through the coupling/loading with adjacent loop antennas, would there not be a potential for some hot spots if all the readers were transmitting at the same time?

I have heard about tunnel diodes of course, but have no great understanding, and only knew two people that ever used them. One was a wireless tech and his mate was a radar tech back in the day. Even then they didn't use them for anything desperate.

I still have a lot of SABorn's RFID documentation to wade through, and there may be a few answers there, but given a single reader and your chaos scenario, what would you expect the reader to see natively? And what would you typically expect to see? I am assuming that what is seen is either a CRC pass, CRC fail or no response (three states). This is before the question of post acquisition processing and data duplicate reduction takes place. This is course an Aunt Sally for you to knock down and modify so that I get a better picture of a typical short period of time (not quite 14 reads a second unless you have Ferrari beef cattle) I appreciate the need for 3D spacial coordinates due to the hopping characteristics of the kangaroo tagged by one of your mates after a refreshment break on a Friday night.
A,A,A,x,A,B,x,A,B.B,B,_,_,_,_,B,x,A,B,B,B,x,_,_,x,A,A,A,x,C,C,C,C,C,_,_,_
Where _ is no detection and x is a failed CRC because of occlusion, intermodulation products etc.

Ian
 
Last edited:
#19
Ian.........................................................................................................................
The grouping was actually 1, 5, 10, 15, followed by 2, 6, 11, 16 etc, not 1+2+3+4+5 then 6+7+8+9+10 etc. strictly interleaving rather than grouping I suppose. My groups weren't contiguous! In the same way that I had not considered de tuning through the coupling/loading with adjacent loop antennas, would there not be a potential for some hot spots if all the readers were transmitting at the same time?

Ray................................................................................................................................................

Ian, thanks for your reply.....I am now in Melbourne (backwards and forwards)

Like a beacon in the mist, I spotted a "transcendal thought" that shows you are a good parry..."hot spots".

First though, I now see yourgrouping 1,5,10,15 followed by 2,6,11,16 is a good start...good, there is implied physical seperation 1,(2,3,4)5 =4 "=three physical slots" between "active-readers"...

but it still does not remove the issue of adjacent coils absorbsion ie mutual coupling into the non-active adjacent loops....the extra thing we need is a system of "tri-stating"....that is we need to mitigate (not eliminate) the three tuned circuits coupling to each other, in the "local-triplicate-cell formation"...will come back to that.

Sweet...Sweet...I want to get back to your "hot-spots"....but first there is a practical matter concerning the grouping of the readers into lots...you choose "sets" of five...good start, but now I tell you there are vertical poles 3metre apart (they hold the "head-rail")imbedded int the feed trough.

Therefore in a trough length of 3metre we can mount 10 reader-loops (with attached transmitter/receiver)...etc...etc

It is now revealed that readers will be in sets of ten (10)....and your and SABorn basic interleaving strategy somewhat adopted...more later

SABorn, if you are there will you point me to where I can view your earlier work on RFID and Picaxe.....you obviously put a lot of work into it as you have insight.

Now back to "Hot-Spots"....

.in the case where all of a set of say 5-10 readers transmit at the same time (and this is not yet discounted...but...but..later) then some where in the physical space between them, there will be "HOT_SPOTS"...that is places where the vector addition of 5 or 10 or whatever carriers come together in phase momentary......like a standing wave...like a sunami..like a crescendo and because the 5-10 carriers (134.2khz) are not frequency locked then the hot-spot MOVES continusiously.

Ian ...well done, you have seen far ahead and seen another story "if the cow won't move then move the field....this is another tool I use and by design or accident you have seen it.....

This waxing and waning of the beat between multiple unsynchronized carriers is another "beautiful" thing to see, watch and use....should any new chum be watching this thread these glimpses may later benefit you....

I talk too much ...let me pause and digest this (your) insight...well done...well done....the sound of footprints may be muffled or absent, no matter,no matter... let me for a while wax lyrical.

Regard to both
Ray
 

SAborn

Senior Member
#20
My involvement with tag readers comes from building a reader from scratch with a fellow electronic enthusiast, as he has a small farm with cattle, and it was a challenge to have a go at designing a reader for about $20-30 compared to the $500-1000 the commercial readers sold for at the time.

We first did it with discrete common components, then i acquired some of the TI TMS3705 chips and redesigned the reader around this chip, less components and simpler operation.

My partner in crime here wrote the pic program to decode the telegram, all tag data is stored to eeprom and / or transmitted via serial as well.
With the option to download the eeprom when back home at a later time for processing/ filing.

The problem we have is the read distance, at present is limited to 300-400mm from a hand wound antenna coil, we have always intended to redesign the antenna to increase the power in the antenna and the read range, but toooo many other projects to do and it has not happened as yet. (as a hand held wand this read distance is exceptable)

The original data on this project was removed from the forum for a tidy up and a better presentation of the project, but that is still pending also, so i can not point you to any listing of our work in this case.

The hardest part was accessing the tag protocol data to figure out the decoding of the telegram and is why i posted that public for others who follow in our footsteps.

I dont see the problem with multiple antennas as you guys do, as i had read about using more than 1 antenna to increase field strength, but can not find the data sheet at present, from memory the main point was to take care with the polarization of each antenna (coil) so the fields had an adding effect rather than a canceling effect.
I also dont see why a non energized coil will have an effect by the near energized coil, as the chips driving the antenna coil place a short across the coil when not transmitting.

As i had said earlier the picaxe was not suitable for this/our project as its like using a push bike in a drag race......... just tooooooooooooooooooooo slow!! but if all the process control is being done via onboard chips (micro) and its just the serial data or binary data that needs to be transmitted via the picaxe then the picaxe should handle that task ok.

At this point i dont fully understand what it is you want / expect the picaxe to do, or why the picaxe is actually needed to start with.
 
#21
It is now revealed that readers will be in sets of ten (10)...
-----------
Interesting reading and since I have no grounding in the subject, it presents a perfect opportunity for me to make an ass of myself, but I'm sure it will not be the first/last time for this. So, my thinking on this is that as more information is gathered on specifics, that a software model be created to handle simulations. Excel is capable or select a good Basic for the PC. In this way, you can identify transmitters and deal with the RFID in a matrix across which you can define any gradient distance. Excel is a 2-dimensional matrix by default and PC Basic and easily allow the creation of X,Y matrix. Then, signal propagation, receiver power-up and transmit time, transmitter interaction, and such can be manipulated.

The other interesting theme is that the data-collection receivers seem to be of a HUB topology... all feeding back to a central data collection site. Yet, the classic cow-at-feeding trough problem with 10 energizers and receivers (reader-loops) across a distance of 3 meters seems a classic engineering L-R / R-L scan and aggregation; that is, energize and scan at 20+ mS intervals across the trough one direction and then scan back the other way. If all 10 readers were serviced by a collector, then that unit could simply do the time/date and unique ID log. An embedded "DOS" or Linux (other embedded <OS>) controller running a compiled program is all that is necessary to manage a grouping of scanned receivers... the pattern and specific locations being mapped to the Excel model.

Just from an engineering prospective, the modeling makes sense to help document issues, interactions, and core rules.

- Ray B.

Some fantastic Excel modeling showing product capabilities: http://excelunusual.com/
 
Last edited:
#22
Hello mburnette (Ray M.)...thank you for talking with me.

SABorn.............................................................................................................................................
I dont see the problem with multiple antennas as you guys do, as i had read about using more than 1 antenna to increase field strength, but can not find the data sheet at present, from memory the main point was to take care with the polarization of each antsteering" the enna (coil) so the fields had an adding effect rather than a canceling effect.
I also dont see why a non energized coil will have an effect by the near energized coil, as the chips driving the antenna coil place a short across the coil when not transmitting.

Ray......................................................................................................................................................
Thanks for your reply SABorn...your reference to using "more than one" antenna(coil/loop?) to create a crescendo or multiplicatiom of magnetic field intensity at some favoured or prefered point...is a good idea....and driving the coils in or anti-phase (or any phase angle between 0-180 degrees) is part of a kit of "rules" ... doubtless you have in mind the concept of forming and positioning "Hot-Spots"... a matter also raised by Ian just recently....let us return to this later or perhaps Ian can expand for you as required.


mrburnette {Ray.M}.......................................................................................................................
Just from an engineering prospective, the modeling makes sense to help document issues, interactions, and core rules.

Ray..............................................................................................................................................
The creation of "Rules" which mrburnette(Ray) has wisely flagged is pivotal....yes Ray, but not just yet....
There is a process called huerostics (moving from the known to the unknown) which I find calmer than "top down processing" which is more conventional in defining a problem....any how that is just me.

Ray M....There are rules for animal behavior, rules for data recovery, rules for data conveyance, and rules for making rules....it is not necessary for us to be so formal...life and death are not hinged on our words here....so Ray I invite you to brave the world of "what if"..."what if?"... is that not the starting point of "Lotus 123"?...... and now the land of "Excell"? ....

It was an awakening for me when I heard someone say that AutoCAD was nothing more than a glorified spreadsheet. Lines between starts and ends, xyz ordinates....now it's all an art...the simple times are gone...the point I wish to make to you concerning this thread is that beyond clever is the basics and they must be mastered or revealed first..... (I am trying hard to not be off with the fairies.....well ok, its ok some times)....this is the word of "blueSky Dreaming" a place where the cat says to Alice when she asks where do I start?...and he says at the begining my dear!.....Ray, that "ass" is not your burdon but your friend....

Ray M.... I went to the site you pointed to....how do these guys do it?....beyond clever...I caught the word "sprite" as I scanned....Browning motion and animal behavior share a common (or so I think) theme...chaos... The rules for handling chaos, such as to add more chaos to quieten chaos and
to confound chaos behavior use its enemy...order.....set a bomb off in a rowdy "beer-hall" and for a moment the chaos will be managed into silence....all a long way from pickaxe?...but I think not.

SABorn.......................................................
At this point i dont fully understand what it is you want / expect the picaxe to do, or why the picaxe is actually needed to start with.

Ray......................................................................................................................................................
SABorn....the reality was I was exploring the possibility for finding a cheap solution...I agree that perhaps pickaxe may not do the real clever bits but there may be a role for it in the pedestrian periferal things...don't know...don't 'know...perhaps all I can take away rom this forum is some key thoughts and considered opinions...with that I would leave happy.


Ray.M...............................................................................................................................................
The other interesting theme is that the data-collection receivers seem to be of a HUB topology... all feeding back to a central data collection site.

Ray................................................................................................................................................
Yes Ray.. I am seeing a Hub too....my secret fear is we may be contemplating reinventing the internet...well sort of

Regards all
Ray
 
#23
Ray M............................................................................................................................................
*Excel is a 2-dimensional matrix by default and PC Basic and easily allow the creation of X,Y matrix. Then, signal propagation, receiver power-up and transmit time, transmitter interaction, and such can be manipulated.

Ray..........................................................................................................................
Sorry Ray....I reread your post....I am a slow learner, but now recognize the "gem" in your words and this is "manipulated"

I see another word behind this, and it is "adapting"....to change one's behavior or use another stimulation to achieve a more favourable outcome.....

Let me instance an example related to RFID reading and collision avoidance.

Take the case of just one "fence mounted RFID Reader"... forget all the others for a while....
The round button thing in the cows ear is called a "transponder".

Consider this sequence....

The reader first fires off a short 15 millisecond duration electromagnetic burst,

A cow at the extremity of the detection range (say 1.5 metre away) will not have aquired sufficient energy to fully charge up the "crystal-set-storage-capacitor"inside the button tag.....BUT...BUT...

BUT...it will have aquired a little power, perhaps just enough for this faint "little Tag" to echo back only the first few words of its complete telegram.....this would be the "header" or even just a fragment of the header reply.....

The Fence mounted receiver, (with a pickup coil perhaps 1 metre in diameter...these are big ears for listening (goldilocks)"...will hear this fragment...sure it cannot make head nor tail of it , but what the heck the reader knows "something exists...yes/no?"...it needs to adapt...adapt.

Well what would you have the reader do? Sit there and wait for the faint cow to come closer or should/could it not send off a "stronger interrogating pulse???? to more fully charge the approaching ear tag.....and if even this new pulse is inadequate then up the power of the next after next emitted pulse.....hopefully before the cows brain is fryed (microwaved?) the little tag will have answered with a full legal telegram.

In practice I have used a three stage power sequence...a 15mS wide burst then a 30mS burst the a 50mS burst....you could go on forever like this however beyond 50mS every thing starts to slow down.....wht if a 60 second pulse were used...yes you are fight...all the cows would have run past (sneaked by) without any of them being detected...not a good thing

With a little thought you will see further inferences with this concept of "adapting" the pulse width...one can seperate cows into "close and far"....well sort of any way....more perhaps later

Suffice to say Ray...a software routine that incorperate adaption is screaming out to be used here

By way as a reinforcing example on "adaption" perhaps you recall that ear=tags use a free running L/C oscillator and can therefore be listening and transmitting anywhere between say 132-136 khz..
T manage this reality, I would propose three seperate pulses one below 134.2khz, the next at 134.2khz and the next at 136khz.....this is spectrum jumping and if the tag is found and successfully read at 136 it not necessary to keep frequency jumping...move on break ot of the loop and wait for the next cow...sort-of.

Regards Ray
 

Paix

Senior Member
#24
@SABorn, I'm very glad that you were able to advise that an inactive reader effectively shorts it's antenna and so takes it totally out of the equation. A refreshingly positive point given the alternative prospect of complications from passive coupling to inactive readers.
My reading at the moment is a little like some of the early software manuals, where the technique for survival was to effectively identify the few pages in the manual that were needed. So many devices that are not relevant, but not obviously so. I'm expecting to hear a 'phut; and my head to explode like an egg in a microwave oven any time now. :) Your knowledge and experience is invaluable. I guess that RFID forums tend to be focused on issues that are a lot less dynamic than this and have more constrained thinking on the matter.

I did hear of a Dutchman who's small company produced state of the art SDR receivers which used a neat algorithm for sorting out multiple interleaved NMEA beacon signals. A vessel at sea listens and hearing no beacon transmission in progress, sends it's NMEA identifier, with little possibility of a data collision with other vessels that are within range, but NASA with their elevated viewpoint have a totally different signal horizon and their satellite born NMEA receivers typically receive multiple NMEA identifiers. This guy and his company supply state of the art SDR receivers using this algorithm. Occasionally NASA get back to him because one of these receivers had become broken. They can't send it back, can't tell him what they had been modifying, how it had failed, or anything other than the fact that it was broken and they needed another one under warranty. As he pointed out this technology was so very cutting edge that it fell under ITAR rules and NASA's interpretation of them made life 'very interesting'. This was reported at the UK AMSAT colloquium in 2010.

@Ray, I feel that we have moved back considerably from the original proposition which was perceived as the handling of data from 170 x 100 readers and the role of PICAXE. We now seem to be discussing the readers and the strategy involved in their operation. Patently PIC chips with assembler or compiled code supporting the RFID front end chips from the semiconductor houses is where the current solution lies, as SABorn has pointed out.

I feel that it's only a matter of time before the question of 'dual diversity reception' rears is raised. The signal processing is way beyond my capabilities, as are most things actually, but I have no problem with being a sounding board, provided that no one believes that I have any great technical answers to pull out of the hat. The tortoise and the hare appears to have nothing on this!

That said, my initial perception was of multiple tags and readers in a fairly well controlled volume providing a fairly easily reading. I originally believed that a pretty standard reader was to be used, but I now feel that there is a move to enhance the performance of readers and their function. We appear to be moving into signal processing territory. Whereas my view was originally one of sequentially scanning to keep interaction down, modified to scanning with a number of concurrent shorter sequential scans, giving a greater number of potential reads; I can see now, with talk of steerable hot spots, that that concept appears to be changing.
G2640.jpg

Between the header bars 3m apart. I see only three cows feeding in your pictures and from what you were saying earlier. So with more than three readers per “bay” you appear to have moved from one reader per seat to five, or more, per three seats? I can only assume that the reader density is increased, but not the number of radiators. It seems that control of receivers and antenna is devolved somewhat, with one transmitter controlling and sharing that control of more than one antenna/receiver. T1 R1 R2, T3 R1 R3, T5 R4 R5. The Transmitters and associated Receivers and their receive antennas operating sequentially as a moving group, perhaps over two bays or six dining seats in a sequence. Is this what you are now thinking about. It would need a lot of correlation to sort out who was actually dining.
- - -
@RayM, To be honest, the Excel blows me away. I'm a simple soul and more suited to the Linux ADP collection operation than to any form of real modelling. It's an interesting idea and one that, if understood properly, can save a lot of cutting and trying and it's attendant costs.
- - -
What I do glean from all this, is that my concept of an RFID reader was the same as everyone else. Where the transmitter antenna and receiver were interlocked and controlled by the transmitter logic. I see now that we appear to be moving to a transmitter controlling it's antenna, receiver and other receivers and their antennas. Furthermore that the devices controlled are not exclusive, but can be shared between transmitters. This makes my mind boggle ever so slightly.
I like the idea of the data port at the other end of the box . . . :)

\Ian
 

Attachments

SAborn

Senior Member
#25
So far this discussion is basic brain storming with no real fixed direction.
I feel it time to stop peeing in the trough and start to look at what would be used for a test setup for proof of concept.

What reader modules will you use (data sheets for them) or will you fabricate you own and what RFID chips would you then use (data sheets for them) what antennas will be used (data sheets for them) at least then we have some tools and direction to work with then.

So far many discussions here have been open ended and off with the fairies without nailing down some form of hardware as a starting point.

Its pointless talking about excel without even knowing how the data can be delivered to the PC, and also pointless talking about serial comms without knowing where and in what form the data is coming from.

Some foundations are needed here so the building blocks can be reviewed to fit the application, as i asked how and where do the picaxe fit in here, to this point it has no role until we know what it will interface with.

Im sure there is several people sitting on the sidelines here waiting for some tools to work with before they jump into this discussion.

Lets look at just one bay for a proof of concept setup for starters, with some information of the length of a single bay, how the antennas will be mounted, what height, structural interference, cable condutes, Etc, etc...........

Some real and hard facts to work with, or we will fill the trough with just peeing in it.
 
#26
...
The Fence mounted receiver, (with a pickup coil perhaps 1 metre in diameter...these are big ears for listening (goldilocks)"...will hear this fragment...sure it cannot make head nor tail of it , but what the heck the reader knows "something exists...yes/no?"...it needs to adapt...adapt.
...
In practice I have used a three stage power sequence...a 15mS wide burst then a 30mS burst the a 50mS burst....you could go on forever like this however beyond 50mS every thing starts to slow down.....wht if a 60 second pulse were used...yes you are fight...all the cows would have run past (sneaked by) without any of them being detected...not a good thing

With a little thought you will see further inferences with this concept of "adapting" the pulse width...one can seperate cows into "close and far"....well sort of any way....more perhaps later

Suffice to say Ray...a software routine that incorperate adaption is screaming out to be used here

By way as a reinforcing example on "adaption" perhaps you recall that ear=tags use a free running L/C oscillator and can therefore be listening and transmitting anywhere between say 132-136 khz..
T manage this reality, I would propose three seperate pulses one below 134.2khz, the next at 134.2khz and the next at 136khz.....this is spectrum jumping and if the tag is found and successfully read at 136 it not necessary to keep frequency jumping...move on break ot of the loop and wait for the next cow...sort-of.

Regards Ray
The hi-Q of the receiver works against us here. In addition to the issues associated with distance from transmit antenna to receive antenna, D, we also have an angle concern as well as a non-exact receiver tuning concern. So increasing a 15mS burst to 30mS may work (or not) simply from by how the RFID tag is hanging... that is, whether the supply capicator becomes charged to sufficiently permit a full transmit sequence. Also, there is a possibility of a 4kHz deviation in tag-tuning... for receive & transmit!

Not to disagree, but pulsing transmitter power from 15mS to 50mS and also pulsing the frequency in three separate power bands seems to be far to complex. Rather, I would consider 3 separate transmitters/receivers to cover the 3 frequencies- low, middle, and high. The pulse width should be the same... 15mS-20mS... or whatever is needed to fully charge the RFID even if the tag is not optimumly hanging. The antennas should be placed a staggered pattern such that a moving cow would cover the full range of each antenna's "optimum reach" in approximately "n" second(s). If we feed these 3 transceivers' received signals into an aggregator that logs tag# and real time and antenna number, we should be able to uniquely identify 100% of the RFID tags. By judicious 'cancellation' of duplicate tag #'s per antenna groupings in the aggregator, we now have unique tags with time stamps that represent a unique tag ID event. As an example, if tag 00876098 was received by aggregator "G6" antenna 01 then the record in the log may look like "876098, G6, 01, tttttttttt" where tttttttttt is a system time. If we had multiple 876098 records in the log, we could discard the multiples that existed within a time bracket of nn seconds, say maybe 10 seconds keeping dupes that exist longer than 10 seconds apart... point is, the algorithm is tunable to suit our needs.

So, lots of words and stuff but really is a statistical problem- a population of devices with dissimilar characteristics and random placement transversing a defined area covered by multiple pulse-chargers and receivers. Our job is to figure out how to get our confidence level very high and our error rate very low.

- Ray B.
 
#27
Its pointless talking about excel without even knowing how the data can be delivered to the PC
_-------------------

My intention was to suggest that Excel could be utilized as a modeling tool... that is, grid Excel such that each block represented the physical confines of a cow... so many square feet. By building out a feeding area, the antenna placements, the RFID (cows) and associated power being delivered to the tags can be calculated such that one can approximate which tags will provide a complete datagram...

This would then be tuned with some field tests to see if theory and reality align.

- Ray B.
 
#28
mrburnette (Ray M.)

The hi-Q of the receiver works against us here. In addition to the issues associated with distance from transmit antenna to receive antenna, D, we also have an angle concern as well as a non-exact receiver tuning concern. So increasing a 15mS burst to 30mS may work (or not) simply from by how the RFID tag is hanging... that is, whether the supply capicator becomes charged to sufficiently permit a full transmit sequence. Also, there is a possibility of a 4kHz deviation in tag-tuning... for receive & transmit!

Ray..............................................................................................................................

First, let me apologize to all for appearing to be so vague....I came here not knowing your backgrounds or capabilities...I barely know my own....so I had to find out how to talk with you....

Should I say to you, I want this, this and this and one of these...etc.... I see some of you want a clear cut problem to solve...no bad thing... yet if I need to know how to turn on a light or properly use a picaxe function/directive then surely, as you would, I will go away and read the book.

Surely; if I gave you a list complete project document, with a "bill of materials", you would bombard me (or whoever it was) with a stream of requests for clarifications....and why do you want to do it this or that way?.....and worse still, ask why did'nt you go this way or that? ......

I come to you soliciting participation, a little bit of ownership in a mind game that WILL become a reality...Faries or not, is it no bad thing to put the cart before the horse, patience is called for as we correct each others preconceptions of the problem......(and basics)

Like yourselves, multiple topics tumble over in my mind and fragments come out in the wrong sequence.... I see now I have confused some by begining with cows standing in a feedlot with counting cows running off a truck.....these are both things I do ...

Ray M. Already you see this later case of my fuzzy thinking, and are talking to it....you already see the trilogy of (fast-medium-slow) and (short-medium-high) and (aboveTune-onTune- belowTune) and (X-Y-Z) ...thank you.

Sorry guys, I'll try for a better mix....I value all your thoughts.

So, RayM. On the matter of High Q, yes and no....the designers of the TI readers, in fact after the 50mS power burst, they switch in parallel another capacitor to cause the now quiet transmit coils to resonate not at 134.2 or 124.2khz but at 129khz ie midway between the high shift/low shift of the incomming FSK tag telegram.....PLUS....

PLUS...a damaping resistor is also switched in parallel with the now listening receiver...... both these actions are better conditions to detect the FSK signal......when the next power pulse (70ms Later) is required the damping resistor and mid-band capacitor are removed (electronically).....

mind you ,not all reader brands do this and as a consequence have less detection range....hand readers are a typical instance, they use a tuned circuit Q of about 15 (compared to 200) ....they are a compromise but achieve their desired ends just the same....SABorne will verify what I say from his work with the RFID chip.

Also RayM. Yes you are right, about it not mattering how much power or "off-tune carrier-shift" one might use when the tag is "edge-on" to the flux enimating from the large "fence-reader" or "trough-reader" if you are in a hole you are in a hole if the tree fall and noone is there to hear it well ...well what....as Ian said earlier, "if you are drowning the extra bucket of water makes no difference..."

RayM. Staggering the height of progressive readers and playing games with pulse duration and pule frequency etc...etc goes straight to the heart of ALL these RFID challanges (as you say) and that is "you may only get one bite at this cherry,

Do everything you can to get 100% of the tags read.....no one is going to round up those thousand cows or sheep and bring them back for another run-through, just because you were lax in not doing all possible to read them all.....

Therefore let me exhaust all matters before going to Axium/Protel or Pickaxe....

Oh, the rail above the heads of the trough eating cows is a device to keep their "heads down" and stop them day dreaming when they should be eating....what can I say?

ps...how do I do that trick to put someone's previous quotes inside the grey box...as you do RayM. ?

Regards
Ray
 

SAborn

Senior Member
#29
ps...how do I do that trick to put someone's previous quotes inside the grey box...as you do RayM. ?
Copy the quote, then in your post click the quote tag up the top (second from the right) then paste the quote between the 2 tags on the screen .........{quote} message here {/quote}....... although it will be square brackets around the tag, or you can just type the quote tags in, just use square brackets.
 

SAborn

Senior Member
#30
if you dont give us something as a base of the tag reader to work from how can we tell you our thoughts if it is suitable for the application at hand. (yes it may open a can of worms)

I understand your view of theories being advised against but without something to start with as hardware then this is going no where.

Its like asking here "how can i drive a motor from a picaxe" well we all might give a different method and advise the use of 101 different motor driver chips, etc, etc.
But if you asked "how can i drive a xxx stepper motor using chip XXX " then we have some tools to work with to narrow the field and offer more direct advice for the task at hand.

I am no expert here just some fool who has dabbled in this area and thought i might be able to offer some advice from knowledge gained, but to be honest im loosing interest fast reading page upon page of dribble with no starting point.

Sorry for being blunt, but another ten pages of waffle and still no building blocks can not be a constructive thread.

Early you quoted Priority one devices, are these what you intend to use? as i have never worked with their products but had read the data they supply some time back, and considered buying some of their chips to have a fiddle with but as yet have not done that.

Like all electronics what one device will do often others wont, so knowing what device is intended for use allows a bigger picture to how best to manage the design to work.

I see this a much more proactive direction than pages of "what ifs"
If your first choice of device is not the best suited for the application then lets weed it out and move on.
 
#31
I spent years as a 'technician', years as an 'engineer', and my last 10 professional years was as an 'architect' in IT. Obviously, each roll "plays" to a completely different audience:
- Technician ... make this thing work again ... are you done yet?
- Engineer ... Here are the toys corporate purchased, make them do this and that ... and do it quickly 'cause we need it yesterday!
- Architect ... connect these things on paper, test in the lab, and document what they are useful for... 'cause management is going to buy a zillion!!

So, it is time to start documenting what we know and what we suspect and what we need to learn.
1. The technology is a broad category identified as RFID
2. Is any of the technology components pre-identified. That is, are all ear-tags from the same manufacturer? Are all components commercial? Are any components home-made? Are reference sheets (PDFs) available?
3. Has the problem-to-solve been articulated? That is, is it fully documented? (Beyond counting cows...) In the documentation, every facet that we know should be identified, documents are editable as more is learned.
4. What error (rate) is acceptable... it is difficult to imagine that 100% is anything other than a design goal.
5. Are there check and balances; that is, do we plan to check that a cow comes through the entrance gate to the trough and catch the fact that it does not register feeding?

The document that must be produced is simply a statement of the current environment and the problem that needs to be solved along with any "scope" limitations. Basic project management. Customer needs must be identified as immediate and future so that one can stabilize the design effort to the immediate but provide flexibility for a future phase. There is little reason to consider an expensive solution if it cannot be adapted later unless we know that the customer only has one-shot at the project and no future activity will occur - in this case, we attempt to solve the immediate and add as much value-add in as we design the solution.

- Ray B.
 

Paix

Senior Member
#32
I was under the impression that a tag wouldn't operate until it had acquired a sufficient charge to produce a complete datagram transmission. This being a function of the point at which the charge would trigger the transponder. Receiving a partial datagram is a problem that we have to live with, but transmitting half a datagram should be technically prohibited by the transponder design. [DISBELIEF] Does incomplete transmission of datagrams actually happen? [/DISBELIEF]

These reader/receiver loops are getting bigger too. Up to 1 metre diameter. I was thinking that a reader antenna would be about 200mm diameter or thereabouts and originally imagined this to be pretty much a fixed feast. Apparently not so.

Correct me if I'm wrong, but to the best of my belief, if a cow hasn't put it's head past that head bar, then it's not feeding and we aren't interested in reading the tag &#8211; at all?

We seem to be making excursions into reading tags in deep space, but now I am beginning to see emerge a picture where, the size of the antenna could actually fill the seat. &#8220;Just pop your head through here to register please Daisy.&#8221;

Once a cow has her head under that bar, what speed can she make in feet per second, other than with a shake of the head of course? Too low a frequency to consider Doppler shift.

With a data shift frequency of 10kHz, then a 4khz tuning error should still be able to read it. It isn't beyond the wit of man to read RTTY for example when one or other of the tones fades and because the detector is probably an FM discriminator it should be able to accommodate I would have thought. No problem with interrogating with weak, strong, strongest power settings. A weaker setting bringing in a clean read should perhaps preclude the necessity for a higher power reading. Frequency agility shouldn't be needed if the reader bandwidth opens up slightly after a failed read.

Two images in my mind. The first is from the film &#8220;The Dish&#8221;, the story of Parkes observatory NSW, in support of the 1969 Apollo moon landing. When they lost the spacecraft signal and were looking to re-position the antenna. Just point it at that big bright thing in the sky, it should help narrow down the search . . .

The other image is a cartoon one, with a cow sat with it;s head under a large parabolic dish. No foil helmets allowed.

SABorn and RayM are right about needing something definitive to work with though.

We started with what appeared to be a communications problem from RFID readers and have devolved to considerations of cutting edge RFID reader design. Far, far away from PICAXE territory.

/Ian
 
#33
Wow! ....just back this morning, and find my three new pickaxe friends are baying for my blood.....

Thank you for at least talking with me .... as a coutesy I will respond to each of you before I go.

Sorry SABorn, you will have to "wade through dribble" for the next several paragraphs or switch off now...your choice.

Ian, suspend your dis-belief, the tags DO send partial telegrams,,,,SAborn's reasearch will confirm that for you...or if not there, find the patent for the HDX tag and see it first hand, as I had to do.

A header to the telegram for synchronnization, preceedes the full message, and even that is
not as you expect...the animal number is transmitted backward, that is least significient number first....why this is done will be a matter for your own discovery. Should I have introduced this earlier?...perhaps.

To the best of my knowledge, even the more modern tags do this too....should I have told you earlier?...perhaps...should it have been a appendix to a weighty tome or project document you all want me to lay before you?....

RayM is correct, when he implies, but does not openly (yet) say, position your self within a list, to know all there is to know about what it is you seek.....if memory serves me correct it was some ancient chineese warlord who said first know your enemy... (how a tag works)....

Ian, sorry for bringing in a 1 metre loop...it was by way of a supporting example for another point....you conception of 200mm diameter loops mounted along the animal-side lip of the trough is correct (at least that is how I am proceeding)..was it not you who blue-skyed a single loop around the pen?...no laughing matter there, entirely possible yes/no?

Ian, you see a clearly delineated detection volume just inside the feed trough "created and suspended in space like some plasma ball"...great if we could achieve it, but the taming of the curl and three dimensional nature of electromagnetis fields has driven better men mad....
As the cow approaches, it is probable and fully expected that pre-detection will occur....
How to handle this, and the kindred issue of passing cows (sprite like) being registered as eating, when they are not, is another topic we could have discussed.

Ian, it is not beyond the wit of man to invoke all the demodulation techniques; the faint whisper from your "moon-ship" could easily be dug out of a sea of noise, 30-40 signal to noise ratio or not.
You misunderstood the significience of the role of resonance here...
My implication was that if the tag was off-tune when taking in the "50mS blast of power" from the "stationary reader" then if it did not aquire or store sufficient energy to charge it's internal capacitor-battery then it is like say, your 08 pickaxe running itself down on say a a 1uF capacitor (acting as the better 3v battery you would sensibly use).... You see by point and counter point, we could have proceeded....its a learning process and works both ways.

Also, Ian....in this world of RFID work...infinity is only 1-2 metre away...not the other side of the moon...what I say to you is, if wou hold the tag in the "null-zone" of the directivity pattern of either the reader coil OR the figure eight null zone of the solonid coil structure of the tag itself...then you will see nothing...nothing .... even a million watts of emitted power will not activate a HDX RFID tag even it is ony 1 metre away and held the "wrong way"....

You see Ian, and others....infinity is only the "twitch of a cow's ear away" and this can happen in less than 1/14 of a second.....should I have given you all this hard won information on my first visit to you?......I think not....you see I hope, that "blue-sky- dreaming" is not to be confused with secret men's business, it has a role long before the compilation of lists and schematics and PDF and bills of materials

Even this comment by "Tazan" did not stop me from visiting you

Tazan..............................................................
Firstly a gripe about image and perception. The image is the one of children toys each time I download a Picaxe .bas file to a microcontroller. The perception is that I&#8217;m playing with a toy. How can I be taken seriously if anyone is looking over my shoulder as this even takes place? My wish is not to display this image just a progress bar will do.

Enough of this now we all have other things to do.

Until another time
Thank you all
Ray
 
#34
Wow! ....just back this morning, and find my three new pickaxe friends are baying for my blood.....

Thank you for at least talking with me .... as a coutesy I will respond to each of you before I go.

Sorry SABorn, you will have to "wade through dribble" for the next several paragraphs or switch off now...your choice.

Ian, suspend your dis-belief, the tags DO send partial telegrams,,,,SAborn's reasearch will confirm that for you...or if not there, find the patent for the HDX tag and see it first hand, as I had to do.

A header to the telegram for synchronnization, preceedes the full message, and even that is
not as you expect...the animal number is transmitted backward, that is least significient number first....why this is done will be a matter for your own discovery. Should I have introduced this earlier?...perhaps.

To the best of my knowledge, even the more modern tags do this too....should I have told you earlier?...perhaps...should it have been a appendix to a weighty tome or project document you all want me to lay before you?....

RayM is correct, when he implies, but does not openly (yet) say, position your self within a list, to know all there is to know about what it is you seek.....if memory serves me correct it was some ancient chineese warlord who said first know your enemy... (how a tag works)....

Ian, sorry for bringing in a 1 metre loop...it was by way of a supporting example for another point....you conception of 200mm diameter loops mounted along the animal-side lip of the trough is correct (at least that is how I am proceeding)..was it not you who blue-skyed a single loop around the pen?...no laughing matter there, entirely possible yes/no?

Ian, you see a clearly delineated detection volume just inside the feed trough "created and suspended in space like some plasma ball"...great if we could achieve it, but the taming of the curl and three dimensional nature of electromagnetis fields has driven better men mad....
As the cow approaches, it is probable and fully expected that pre-detection will occur....
How to handle this, and the kindred issue of passing cows (sprite like) being registered as eating, when they are not, is another topic we could have discussed.

Ian, it is not beyond the wit of man to invoke all the demodulation techniques; the faint whisper from your "moon-ship" could easily be dug out of a sea of noise, 30-40 signal to noise ratio or not.
You misunderstood the significience of the role of resonance here...
My implication was that if the tag was off-tune when taking in the "50mS blast of power" from the "stationary reader" then if it did not aquire or store sufficient energy to charge it's internal capacitor-battery then it is like say, your 08 pickaxe running itself down on say a a 1uF capacitor (acting as the better 3v battery you would sensibly use).... You see by point and counter point, we could have proceeded....its a learning process and works both ways.

Also, Ian....in this world of RFID work...infinity is only 1-2 metre away...not the other side of the moon...what I say to you is, if wou hold the tag in the "null-zone" of the directivity pattern of either the reader coil OR the figure eight null zone of the solonid coil structure of the tag itself...then you will see nothing...nothing .... even a million watts of emitted power will not activate a HDX RFID tag even it is ony 1 metre away and held the "wrong way"....

You see Ian, and others....infinity is only the "twitch of a cow's ear away" and this can happen in less than 1/14 of a second.....should I have given you all this hard won information on my first visit to you?......I think not....you see I hope, that "blue-sky- dreaming" is not to be confused with secret men's business, it has a role long before the compilation of lists and schematics and PDF and bills of materials

Even the inference of picaxe device immaturity and timing sloppyness as voiced by SABorn and all the wishes in the thread called "Pickaxe Editor Wishlist" did not deter me
Even this comment by "Tazan" did not stop me from visiting you

Tazan..............................................................
Firstly a gripe about image and perception. The image is the one of children toys each time I download a Picaxe .bas file to a microcontroller. The perception is that I’m playing with a toy. How can I be taken seriously if anyone is looking over my shoulder as this even takes place? My wish is not to display this image just a progress bar will do.

Enough of this now we all have other things to do.

Until another time
Thank you all
Ray
 

lbenson

Senior Member
#35
Ray,

I would be sorry if the impatience of some regarding your fuller explanation of this problem--about which you know far more than anyone here--would keep you away. This is a very interesting, large, real-world problem, and I have enjoyed the blue-skying.

Unfortunately, I have no knowledge to impart, and I imagine that few here do. I'm not at all certain that "picaxe device immaturity and timing sloppyness" would be an issue for you. Timing issues regarding serial i/o have been reported and documented, especially for the 20X2 chip, but for most of the chips there are some very reliable speeds. But Dippy often says that if he wants reliable serial i/o, he would use an external frequency source.

A bigger concern for me with regard to the use of the picaxe on a large scale would be price. While the 08M2 is not expensive, especially for demonstrating a design, if you scale up to 17,000 or 60,000 units, you could well be talking double the price of some competing parts which might need more non-recurring engineering development costs.

In any case, if no one else steps up here, good luck.
 
#36
...
should I have given you all this hard won information on my first visit to you?......I think not....you see I hope, that "blue-sky- dreaming" is not to be confused with secret men's business, it has a role long before the compilation of lists and schematics and PDF and bills of materials
What is happening here?

lbenson and I both share the same sentiment, " I would be sorry if the impatience of some regarding your fuller explanation of this problem--about which you know far more than anyone here--would keep you away. This is a very interesting, large, real-world problem... ". SABorn and Ian have entered the discussions in an honest and open manner, as I have. But I doubt that any one of us has time to waste when critical knowledge is withheld knowingly. For myself, I can independently read and research and even prototype in my lab, but while I am interested in RFID, I am starting with no practical, hands on experience - this is why a statement of objectives and existing knowledge is critical. I do not mind being on a team effort, but we all come with different backgrounds and areas of expertise- it is critical that everyone know what is trying to be accomplished.

Ray, you appear to be the expert in this subject so it is upon you to ask the appropriate questions and set the correct tone to guide discussion. You will not find a more dedicated group of helpful folks than those here on the PICAXE forum. Your participation and knowledge would be a great asset , too. But your participation will require a commitment to not view knowledge as proprietary or secret... everyone here shares gracefully source code and experiences without a second thought. When one examines code from others, it is not unusual to see many comments throughout giving credit to another forum member for the code snippet or the inspiration. Yes, some of this code finds its way into the for-profit sector.

I wish you well in your quest... it is a most interesting project.
mrburnette (Ray B.)
 
Last edited:

SAborn

Senior Member
#37
It would appear that i have rattled your cage some, and you have decided to take your bat and ball and leave, just because you dont like the score.
Not very sportsman like is it?

One needs to ask the question if you have ever intended to disclose what the hardware intended for use would be anyway, to fill the gap of "blue sky dreaming" and actual hard facts.

We are all here to help and to learn from each other and if your only intention was to gleam what you could from our collective knowledge base and not give back in return, then to go off and make a million dollars from the information give to you freely, i would think then you have made the right choice.

On the other hand if you are prepared to give back and work with the collective minds here the results can be startling of what you can extract from the knowledge base.

your comment of............

should I have given you all this hard won information on my first visit to you?......I think not....you see I hope, that "blue-sky- dreaming" is not to be confused with secret men's business, it has a role long before the compilation of lists and schematics and PDF and bills of materials
Would imply we will never get to know about the hardware and its use, you also need to respect you are asking for information that others here have gained from "hard won information" from their prior developments and expect them to share freely with you as we all do here, so is it so unfair to ask you to give some insight into the hardware intended for use in your application as i had requested.

I would like to see the thread continue, as would many here who has followed silently in the back ground, but remember its a two way street of sharing information and helping each other within doing so.

The ball is back in your court.
 

Paix

Senior Member
#38
@Ray, I don't know about baying for your blood, but seeking a firm base to stand on most and direction to travel, most definitely.

I have no prior RFID knowledge, but have long been intrigued. It just never quite bubbled to the top of my interest list. Somehow I think that

Disbelief suspended. I feel that in permitting the transmission of partial data someone missed a trick. My thinking is too simplistic to want to start with anything that is less than complete. After all, no one would countenance the reader emitting sprogs willy nilly, but but having said that I can appreciate that there is information to be had from partial reads, but discounted it as having any practical use because without a CRC check I assumed that it was naturally discarded. This after all is a &#8220;tag reading&#8221; and not an &#8220;intelligence gathering&#8221; exercise.

The acceptability of a partial read is reinforced by the fact that the tag number is transmitted backwards. Suggesting that it is seen as more than just a static binary pattern. It does seem to be predicated on the fact that in any large group of tags the most important information is likely to be in the tail of the number. So, by transmitting the serial number backwards, you have more of a chance of identifying an individual animal than the batches of tags in use. Don't knock that either as the batches are likely to show the different sources of the cattle, well the tag pooling/buying habits anyway. If the tags in a particular Home-Pen are known, then four or five bytes of data are probably going to be sufficient to match one beast. So, the reversed transmission of the tag number indicates that probabilities are almost as important as CRC-certainties and it seems we are in the intelligence gathering business after all - Comint Central. Should you have mentioned the tag number reversal earlier. Not really, but it does give a bit of context to the wider operational aspect of what's important. I assumed only good reads were viable, unless of course they are being used within the reader to support a technical strategy.

I don't think that the laying out of the complete documentation is warranted. I have seen people come along and ask for advice and then play the &#8220;This is commercially sensitive card&#8221; and become miffed that people don't want to play. From the direction that we have taken it eventually became apparent that your agenda seems more to do with aggressive reader development in a very challenging environment, than it did with the original brief of how to get the data from 170,000 tags to the spreadsheet/database of the decision makers in the back office.
Come along and we'll make a bag of sweets Sunny Jim. Only to find that the object of the exercise is to make the bag and not the sweets. Don'tcha just love these analogies.
I think that you are floating a lot of interesting ideas and a fair bit of noise too. I did wonder, briefly, why you weren't asking your questions on an RFID forum, as there seem to be no great shortage, but plainly, from a quick glance, there seem to be not a lot of blue-sky-thinking there and probably more than a few people that might be more than happy to hijack you as soon as look at you. The RFID Establishment? I think here there are a lot of open minds, but the eventual aim is to be able to arrive at and advise a solution or at least a direction. We haven't quite got to a solid problem yet; it's still out there spinning in the dust with all that beef :)

Sun Tzu &#8211; The Art of War. Sun Tzu translated by Lionel Giles (2005) [Translation first published 1910]. The Art of War by Sun Tzu &#8211; Special Edition. El Paso Norte Press. ISBN 0-9760726-9-6. Available from Project Gutenberg. From the frequent amount of translation it has enjoyed, this sounds like a text everyone should read at some time in their lives. The earlier the better to benefit from the advice and insights given.

No, you imagined the single loop around the pen. I believe that I said single antenna and was actually thinking of a dipole, but given the frequency and the wavelength, then I suppose that around the pen might have been a realistic assumption. I was thinking of around the immediate feeding area. It's a shame that the loop can't be round the top of a bucket from which the cows eat. The trough needs to be kept clear throughout it's length for filling and cleaning operations which obviously restrict your options a little.

Have a read of this short item:
http://k9jy.com/blog/2008/03/04/cw-skimmer-a-monster-or-killer-tool/
Then if it has anything of interest Google &#8220;CW Skimmer&#8221;. Just the number of categories alone will tell you that it's not something casual. I would be interested in hearing what you think you can glean from it. Certainly it indicates what can be done once the tag signals are in the ether!

You have said before about mounting the antennae on the animal side lip of the trough; is it concrete? I have thought this before, but not followed up due to a bit of confusion as to what is actually meant. Two points: The animal has to reach over the lip and surely once there the lip would tend to occlude the signal? Not only that, but being on the near side, it would be more aft to pick up passing and arriving animals. It also suggests that the plane of the antenna would be flush with the flat front of the trough, at around breast height? This seems like a lot of cross polarization before eating takes place &#8211; and then? Too much information I know, but I don't flick my ears too much while I'm at the trough! I'll observe the OH more closely later. On the back lip of the trough, would there be less potential physical interference (potential rubbing) from cows but more prospect of damage from cleaning operations?

I think that I have asked before, but what is the established practice in this area, or isn't there any prior art to speak of?

I definitely like the idea of a plasma ball energy/detection volume. It would certainly solve a few problems. Nicola Tessla would be a great adviser there, but you might find that there is no smoke without cooked beef . . .

I stand corrected about the off tune tag. The problem isn't with the data signal, but sub-optimal powering of the tag. That makes good sense to me.

PIXACE Programming Editor Wish List &#8211; Well, there is a bit of blue-sky-thinking. I firmly believe that we should be careful about what we wish for . . . ha ha. Every time I install a new Linux OS, I spend fettling time trying to get it as close to the experience that I had previously. So call me conservative.

Should you decide to disengage, then it is perhaps understandable, but a bit of a shame. I have learned much and hope that my efforts have not been in vain. I saw my original mission here to engage with you and elicit enough firm information so that it was available for the clever guys that could be of more use to you than I. I don't know the answers, but sometimes know the questions to ask.

I will continue to read SABorn's reference works on the subject for my own edification and do a bit of project work for my bees. Alas, my electronics experience has all been hobby based, the communication and programming at a bread and butter professional level.

Not a lot of direct contributors perhaps Ray, but 830+ views must mean that there is a fair bit of interest out there. I know that you and I haven't had the time to pump up the numbers. The subject has a lot of interest but I suspect that without a firm base a lot of potential contributors will be holding back rather than risk being drawn into the void. As can happen when infinity is but an ear twitch away.

Regards,
Ian

PS Take a deep breath and KICK THE BALL.
 
#39
Sorry guys...a bad day by me yesterday....now, where were we?...back to cows.
First, I am sending you two pictures, these should anchor the scene for our discussions....the goal is to document the eating activity
RFID and things Radio will get the data....a special micro will probably recover the data...picaxe may send the data, Excell/VB may assemble the data and do "the eye-candy things"

I will try now for brevity...but first
I came to this forum because of two things....I sensed a mix of mature age and "young-gun-up-and-commers"..... I was one once, and relished the thoughts and words of those who went before me.
Sure this project may result in rich bounty....I new pickaxe was/is an educational tool, so it stood to reason that fresh young minds would be not far away from here.
 
Top