RF Pump Controller

[alb]

Hi,

I have recently found out about Picaxe and am definitely in the "hype phase" (they seem to be able to do anything!). So some advice please.

I am looking to build / buy an RF pump controller to save me walking up the hill in the rain to check how full the water tank is (it usually happens in the rain because that’s when I remember I haven’t pumped for while). Ideally the transmitter should wake up every few minutes and check the state of a few micro switches (even one would do at a pinch) and transmit that to the receiver, which should turn on a relay based on the received information (as long as a transmission has been received in the last few minutes). The transmission side needs to be battery powered so low power use is important.

The 433MHz radio modules look ideal but I am looking for something similar to base the encoding /decoding functions and I am not very electronically competent. Can anyone point me at a book / article / web site that would get me started, or please tell me I am over hyping the technology and should try something else.

Thanks in advance

Al Breingan

 

[premelec]

I use a float switch on/in my tank & pump
:-0

Some reason that's no good? Or simply measure the head pressure if you are pumping up hill -if you want a sharp pressure increase when the tank is full put a float shutoff valve on the inlet and the pumping pressure will increase radically when the valve shuts! Or have fun with a wireless link - which can be powered from a small in line water turbine....

 

[moxhamj]

Yes, picaxe can solve your problem!

1) Measuring the level. Lots of options. I use pressure transducers as they give the actual level. A microswitch will work as well. So will conducting wires, floats with magnets, reed switches, mercury tilt switches etc.

2) Collecting the data. For a pressure transducer you need one 324 op amp and a few resistors. A microswitch will need a pull up resistor.

3) Transmitting the data. Use serout and a 433 module. Send at least 3 85's (U)to wake up and then something for serin to match - eg 'Data' viz UUUDataF for full and UUUDataE for empty.

4) Use a 433 receiver into a picaxe and serin to decode the data and light a led or talk to a PC or sound a buzzer or whatever.

5) Power. I use solar panels producing about 50mA and another power management picaxe that detects voltages and turns the charging on and off, but a very simple solution is charge 3 NiMH batteries at 50mA and throw away the overcharge as heat and just live with the fact the batteries need to be replaced every few years.

The only catch, and it is a big catch, is getting ranges of more than a few hundred metres or non direct line of sight. There are huge numbers of RF modules in 50-100 metre range but not much at the 1-2Km range. Manuka and I are will be posting the results of some experiments soon with higher power modules.

 

[manuka]

We will ? Ah yes- Adelaide Wellington 433MHz data link.

I've been contacted via email as well about this pump/level need,& have had to express some hesitation about offering advice with the sort of grunty mains powered pumps that are used. Simple blunders (such as forgetting RCDs) may be lethal...

The link here is only ~150m, so I'm wondering if low voltage wired signalling couldn't be used. As is well known, Picaxe serial is so robust that twin core would handle it easily. I'd say SWER data( single wire earth return) may even be viable over galvanised #8 fencing wire too.

Don't forget the old "broom handle on a float" level trick. With binoculars this can allow optical level verification from miles away,with no flat batteries etc to annoy. Stan



Edited by - manuka on 03/04/2007 09:16:38

 

[moxhamj]

Ah, does that mean Alb knows where you live?!

Anyway, I just paid the chap in China for a couple of those high powered modules, so hopefully they will arrive in the next week or two.

Meanwhile, there is the solution of using fence wires. Or SWER. Or the best option - cheap Cat5 or telephone cable buried in a conduit in the ground.

 

[manuka]

Righto-I'll keep my ears pricked for Trans-Tasman 2400bps 433.920MHz signals. Shades of Marconi trans Atlantic ? I'm still pondering that ½W 470MHz UHF CB sets could legally be used for such telemetry & control. Aus/NZ regs. allow Ch. 21(?) "3 seconds an hour" which looks pretty well suited to occasional water level checking & switching etc. Fancy taking this further? SiChip keep asking me about this in fact...

IMHO, given the mission critical nature of such pumping projects,&amp; assuming everything is on the owners property, Alb's mere 150 metres really best suits a wired link (&amp; budget?). Even with the crazy copper prices <A href='http://www.eetimes.com/news/latest/showArticle.jhtml?articleID=198701947 ' Target=_Blank>External Web Link</a>,there's usually plenty of cheap/surplus sturdy signal level wire around, &amp; this doesn't even need to be copper! I've just put up a simple home clothes line extension using some extremely flexible but STRONG &amp; dirt cheap plastic covered iron/steel wire that's almost wasted in such a boring application. Unless of course coated,the end connections may be rust prone, but you can rest easy with kangaroo nibbling. Check uploaded pix=&gt; <A href='http://www.picaxe.orcon.net.nz/ironwire.jpg' Target=_Blank>External Web Link</a>

EXTRA: There's nothing new about <b>iron/steel </b> wires for signaling of course. I well recall un-ravelling miles of the (insulated) stuff when in the army-it'd usually survive even a tank running over it,&amp; could later be abandoned in place due to cheapness.Although devised 80 years back (NZ-Mandeno 1925), SWER &amp; iron conductors still have numerous benefits even for power work =&gt;<A href='http://en.wikipedia.org/wiki/Single_wire_earth_return' Target=_Blank>External Web Link</a>, &amp; additionally overhead iron/steel wires should be less prone to theft (ah-&quot;steeling&quot; ?). In fact Lloyd Mandeno <A href='http://www.dnzb.govt.nz/dnzb/default.asp?Find_Quick.asp?PersonEssay=5M31 ' Target=_Blank>External Web Link</a>makes me proud to be a lateral thinking &quot;roll up our sleeves&quot; Kiwi, &amp; he'd no doubt be delighted that global copper prices are prompting renewed SWER appraisal !
Stan


Edited by - manuka on 04/04/2007 01:33:40

 

[BrendanP]

Try radiometrixs tx/rx units. I got a around 400 meters 1/4 wave helical to 1/4 wave helical. 151.300 megs.

 

 

[moxhamj]

Manuka - nothing wrong with plastic coated steel wire - I'm controlling 7Kw of pumps with 500 metres of it. Used for clothes lines and ended up being the cheapest. And the wildlife has left it alone and I think it is because the coating is black. Koalas have eaten blue wires (cat5) and blue rope so I think they like blue. The tank level data system uses fence wires. Can solder copper wire to galvanised steel fairly easily using a 60W iron and soldering flux. Paint the joint afterwards to stop rust.

There may be a cunning circuit using just one wire where you send 12V for a bit, charge up a cap, and then a picaxe at the tank end sends back the data. Avoids the solar cell and battery.

As for the UHF radios at 3 seconds an hour, I have yet to find a brand that lets you tune to the data channels. If such a brand exists, then picaxe would be perfect for sending a quick chirp of data at 1200 or 2400 baud.

However, as discussed previously on this forum, my entire pump system gets zapped about once a year with lightning, and so I'm in the process of adding on a wireless system as well.

Brendan - can you describe your helical antennas a bit more? How many mW are the RF modules?

Edited by - Dr_Acula on 04/04/2007 14:55:24

 

[BrendanP]

The radiomterix units are 100 milliwatts.

I bought 'rubber duck' helical antennas from a radio shop and trimmed them to lenght. They come with a chart that gives the formula.

Alternately make a couple of quarter wave whips with artificial ground planes.

If you really want long range buy two off the shelf TV antennas from kmart or wherever. Point them at each other and then youll get range. 151.300 mg is in the TV band

Radiometrix cite ranges in the KM's with yagis either end. Only problem is the impedance matching between 75ohms-50ohms. Stan is this a big issue? Is the mismatch that bad? Will it greatly affect the power transfer?
Speak O RF sage learned one!

I've just orderd a current copy of the ARRL antenna book from amazon. These and other mysterys will soon be revealed.

And if you get the radiomterix units get the 1.2kbps units bcause the slower the data rate the longer the rf range. They come in 10kbps and 1.2kbps. Radiometrixs agent is in Tasmania. He's a helpfull sort of guy. Have a look a the UK web site. data sheets etc. The units are around 60$ each.

Units have enable pin, power it down low power consumption etc. Also RSSI pin, read it with ADC and then serout to LCD signal strenght handy when setting up.

If you want to experiment Dr.Aacula I'll loan you a pair. You can slot them into bread board. Standard 100' spacing.


Edited by - BrendanP on 05/04/2007 01:13:56

Edited by - BrendanP on 05/04/2007 01:17:30

 

[lbenson]

Dr_Acula--amidst all the (justified) hoopla over the release of the X1 parts, I appreciate your recent remark to the effect that many things which cannot be done with one 08M may be done with multiple 08Ms. In that light, in this thread you mention using a &quot;power management picaxe that detects voltages and turns [battery] charging on and off&quot;. This sounds like a very useful feature--can you provide the circuit and code?

 

[moxhamj]

lbenson - power management is a matter of sensing battery voltages accurately, and that means the voltage supply must be stable which means they have to run on regulated 5V. I have two circuits that I use a lot:

1) 7 NiMH rechargeables charged up with solar. Usual volts are 9-10V. Run the picaxe off 78L05. Sense the battery voltage with a divide by 3 network (20k/10k). Charge via solar panel and suitable dropping resistor and via a normally closed relay. When charged (1.5V per cell), energise the relay via 2k7/547 to switch off charging. Energising the relay flattens the battery. Turn off the relay when cells fall to 1.4V/cell. Components - picaxe 08M, 33uF/0.1uf, 78L05, relay, 2k7k/547, dropping resistor, solar panel (15-20V open circuit, approx 50mA), 20k/10k divider.

2) Circuit 2 uses an old car battery (always available free from friends). Charge using an 8V solar panel (16 cells) and a stepup converter consisting of a 9V-12V transformer configured as an inductor. Use the secondary only (put some insulation on the primary wires). Pulse the inductor using a BUK55 mosfet driven directly from a picaxe. Standard boost circuit using inductor and 4001 diode. 7ms pulses every 20ms or so. Advantage of using transformers as inductors is you don't have to wind any coils, the frequency is slow and within range of the picaxe and the diode doesn't have to be high speed. Sample battery volts using divide by 3 resistor network. Also sample a red led voltage reference just to make sure the picaxe has a solid 5V. Run the 78L05 and picaxe off the solar panel rather than the battery so it doesn't waste power at night.

Hmm - this is getting way too complicated to describe. I think I need to draw up the schematic...

Edited by - Dr_Acula on 05/04/2007 03:37:37

 

[moxhamj]

Re Brendan - thanks for the info. When you say helical, do you mean a 1/4 wave wound into a tight spiral, or do you mean a much bigger helix like the ones used for indoor TV antennas?

The latter send out a polarised directional signal and may be a bit easier to build than long multi element yagis (eg &gt;10 elements). Having got lost myself in the world of impedence matching and the like, it may well be easier just to throw more power at the problem and use 100mW 1/4 wave non directional antennas, than 10mW modules with fancy antennas.

I've ordered some high power xbeepro units (cheaper to order directly from maxstream in the US than from any retailer in Australia). I've also ordered some really cheap high power modules from China. Once they arrive the experiments can begin.

The challenge - get &gt;1Km non line of sight for total RF module cost &lt;$20.

 

[BrendanP]

Helicals like rubber ducks on hand helds.
I was going to suggest GSM modules but the 20$ budget rules that one out!

I'll be intrested to hear Stans thoughts re. impeadance matching.

Where do you live Dr?

 

 

[moxhamj]

Those antennas are ? 1/4wave with the spiral to make the antenna shorter. Omnidirectional.

I'm in Adelaide (Hawthorndene).

I wonder where the original poster of this thread has gone and what they think of this discussion?

 

[BrendanP]

Correct. You could just as well use a straight piece of wire cut to the correct lenght. Have a look at the Radiometrix site as they have notes in the data sheets re. antenna design.

You could make a yagi along the lines of the ones that stan made for his 433 modules but at 151.33. Stan wrote a sil chip article about making them.

I'm in Melbourne. I thought you might be around here. If you want to try these radiometrix modules give me a call on 03 95282537 and Ill send them over to you. Youve helped me out on problems Ive had so Im more than happy to return the favour. There just sitting in the parts box here.
&#160;

Edited by - BrendanP on 05/04/2007 08:22:11

Edited by - BrendanP on 05/04/2007 08:27:10

 

[domwild]

Dr-Acula,

Am intrigued by the steel wire idea for pump control as I have a fence to the pump site. Unfortunately, it will be elctrified soon so I will have to add a second insulated wire unless, as you said, there may be a clever single wire-earth return scheme possible.

How did you send the ON/OFF signals to the pump? Pls give me the gruesome details as I do not have a black belt in electrickery!

 

 

[Summertown]

Dr_Acula;
In the above you mention you use pressure tranducers to measure water levels in tanks. What pressure sensor do you use? I currently use ultrason ic detectors to do this.
Thanks

 

[moxhamj]

Pressure transducer - Farnell 24PC05SMT $23.04. Search for &quot;honeywell pressure&quot; and it comes up at the bottom of the page. 5psi. This is a surface mount device and 4mm barbed irrigation fittings fit over the port.

Alternatively, SensorTechnics part 24PC0250D6A which is the same part but physically bigger and the flexible 4mm irrigation pipe fits over the port. Shipping is a bit more from Germany.

I can post a schematic for a x33 differential amplifier that uses a 324 op amp. The tank module is thus a 324 op amp and an 08M doing dual duty as a battery monitor and a sensor/transmitter plus the RF module if you are going RF.

I actually went for a wire solution and used the picaxe to encode a binary stream which is essentially RS232 protocol but at about 1 baud, switched using a relay at the top and driving a relay at the bottom. Takes 9 seconds to send a byte. This is slow enough to debug using a led or bulb or logic probe. I needed a very definite signal as I have some high tension power lines nearby and any exposed wire has low current AC on it.

I took a 40mm blank end screwed into a T piece after the tank valve, and glued a 4mm barbed fitting into the middle of the blank end. Then a piece of clear 4mm pipe glued onto a piece of wood up the side of the tank. This is a visual indication of the level. Then a 4mm T piece at the bottom to the pressure transducer.

Re controlling the pumps I just send 24VAC down the single wire and turn a 24VAC relay on at the bottom of the hill. DC would work fine too. The relay turns on a 3 phase contactor which turns on the pump. Every pump shed already has a good earth for the mains, so only one wire is needed. Two pumps - use the positive and negative phase of the AC signal and rectify at the other end and drive two relays. Three or more pumps - use picaxes to encode/decode pulse width data or binary signals or whatever.

If people are wanting to build similar systems maybe I need to draw up complete schematics and take a few photos...

 

[Summertown]

Dr_Acula,
Thanks for the info! I also live in the Hills (the sign on should be a clew) and appear have similar water management problems.

I have hard wired most of the place up using cat 5. I send 24 and/or 240 volts around the place to control about 30 solenoid valves.
The next stage is to automate, and monitor what happens with the supply side which comes from a Bore as well as about 150,000 ltrs of stored water (bushfires!)
This will be done by a bunch of 18x &amp; 28x chips, hence the interest in pressure transducers.
How long have you been using them?
Have you ever dabbled in flow indicators?

 

[manuka]

All this collective brain storming (mmm- RAIN storming?)is looking VERY productive, &amp; glad Dr_A has run with steel clothes line to good effect.

But back at the 470MHz UHF CB level however. Ch. 22 &amp; 23 are the legal data channels &amp; a quick glance over my assorted cheapie sets here (sourced via Jaycar &amp; DSE etc 2002-2004) shows all cover the full 40 channels. However Jaycar's latest catalog. indicates a ch.22/23 lock out on all their new sets. Grr- no wonder you are annoyed! I'll look at new sets here in NZ to see what offers a full 40ch. No doubt a simple internal 40ch. enable is possible of course, perhaps with a diode matrix etc. Stan

 

[moxhamj]

Re flow indicators I use one that has a paddle made by Kelco. I'm pumping out of a river and as such need to backflush the filter every hour or so. Same would apply to a dam, and possibly a bore but less often. So at the pump shed is a picaxe circuit that takes the control input signal and the flow switch signal and controls the pump and the backflush solenoid. This circuit also has a software timeout so if for some reason the control signal was going on and off every second, the pump would only turn on and off every minute.

Back to UHF radios, the channel lockout is a real pain. There are other problem I came across before I gave up on these radios. One is massive harmonic distortion at audio frequencies. You can hear it in the distorted speech and there is no technical reason why speech coming out of a CB has to sound distorted. Put a 500Hz square wave through the CB and it comes out as a sine wave of about 1Khz. A square pulse comes out as a sort of biphasic sine wave which makes level detection and zero crossing impractical. I think it is a low pass filter doing this. The ear, of course, is largely immune to phase so this biphasic wave sounds similar to a square wave.

Bottom line is I couldn't get any serout signals to go through them - at any baud rate.

The squelch has a very narrow frequency capture range (audio frequencies around middle C) and there needs to be a wakeup tone. If a tone disappears for a short time the squelch drops out and you have to wake it up again.

So I have got experiments to a point where sending data consists of a series of tones and they have to be just the right length. Decoder is a low pass filter of about 10hz and a level detector. Trouble is, I can't get more than a few bytes through in 3 seconds.

I can't see how a PLL using FSK with different frequencies for high and low would work as I think it would lock into harmonics. Unless maybe use a narrow bandpass filter, or try to reverse engineer the phase change of the low pass filter and cancel it out.

And you have to hack into the handheld CBs to get into the microphone circuit. And be careful not to zap the input stage which I did to one. They are horrible to work on as several brands use weird torx screws. And you need a relay on the &quot;press to talk button&quot;. I ended up avoiding opening the case too much by feeding a picaxe tone into my own speaker taped to the CB radio. But that introduced even more distortion.

It was at this point I gave up and found the high power RF modules on ebay!

So, maybe it is possible but all the above was a full 2 days worth of experiments. By comparison, 433 modules are a pleasure to work with. I'm just hoping the higher powered ones from China work just as well. If not, I'll try the xbeepro's and if that doesn't work I'll take up BrendanP's kind offer of a lend of the radiometrix units.

Edited by - Dr_Acula on 06/04/2007 00:54:22

 

[manuka]

A few years back [ref SiChip May 2005) I used both Jaycar &amp; DSE CB cheapies to run Hellschreiber with absolutely no hassles. However Hell. is a human decoded &quot;fuzzy mode&quot; , not suitable for device switching. I'd also considered DTMF tones (which are of course legendary in their robustness), but the many specialised ICs that used to abound for this are now harder to pin down. Perhaps hacking a cheapie phone for them may be the answer? Stan

Edited by - manuka on 06/04/2007 01:28:21

 

[BrendanP]

NO! DTMF ic's live on! I have used the Holtek DTMF generation and reception IC's with picaxe. They're dirt cheap, 50 cents each, DIP package have four inputs or outputs which you high or low in combo according to what tone you want or what tone is comming in.

I have about 40 of each if you guys want some or you can buy them from Futurlec in Thailand.

They have a few caps and R's and a crystal you have to use as well.

Btw, radiometrix has some sort of encode/ decode ic that goes on each end of the rf link. The IC has outputs that high low according to what the inputs on the other end are driven to. I just glanced at it, the guy in Tassie has them a couple of bucks each. I think.&#160;

Edited by - BrendanP on 06/04/2007 02:01:43

 

[manuka]

Brendan- great news! You can put Andrew &amp; I down for a bunch straight away, as we've just been discussing ch.22/23 UHF CB Picaxe controlled data comms &amp; I pointed out the near bullet proof nature of slow but reliable DTMF compared with FSK. Are these Holtek's the older HT9170 or the very recent (mid 2006-) HT9172? I'd considered DTMF before but suitable ICs such as the MT8880 etc were too elusive (or costly) back in 2005. It was even tempting to use a programmed PIC micro! Decoding is not too big a hassle, as PC sound card based programs abound. Stan

 

[BrendanP]

HT9170B,have Andrew email his address and I'll send them this comming week. Decoding is done by the Holtek IC. According to the recieved tone it switches its four outputs high or low. It has a valid data pin which you connecto to a interupt, when it goes high it goes to a routine which looks at the input state of the pins and then does what ever its supoposed to do.

&#160;

Edited by - BrendanP on 06/04/2007 13:15:15

 

[stjosephsopunake]

A &quot;paddle system&quot; piezo crystal suitably epoxy'd in a pipe deflected by water movement could generate a nice voltage = water flow. Digital filtering could remove noise and pulses. We get the blank piezo discs in bulk as sounders, strain gauges and they work a treat as short term pressure transducers. To measure pressure pulses I put one in a pipe &quot;Tee&quot; with 'O' rings and a 1uF capacitor in parallel with it to detune it and tame the huge voltage swing you can get from it. With a suitable capacitor it produces a great sample and hold voltage for up to several minutes ADC conversion. This could pick up pressure pulses (= flow)

The same technique makes a great simple strain gauge / weight scale.

Communications wise I have heard of people trying the plastic pipe out to the tank as a conductor for data! Stan and I have found damp string works as well as lines of kids holding hands (if you are allowed to do that still where you teach!) One kid locally decoded interrupted pulses sequences going down an electric fence line to trigger closure of a farm gate.

Another idea with just CB UHF, I am pursuing an on board 08M ASK audio burst super-slow serial modulation technique I cracked for serial modulation of 38kHz to drive an Ir LED Tx Rx setup. I would go for heavy de-clocking of the 08M to produce a slow stream of serial data beep&#8217;s. I have got down to 1 to 2 baud which is a great method of demonstrating serial data in slow (human rate) data transfer in class. This would probably trigger the VOX of a budget UHF CB (4 for NZ$80) hand held as transmitter. This would send ~ 50 Baud or less serial data as audio bursts direct from pin2 which should be able to be Rx&#8217;d by another UHF CB, decoded with another suitably de-clocked 08M and a simple Diode + RC network and serin + ID command.

- Andrew

 

[moxhamj]

Challenge to self - build a RF level indicator from scratch in under 24 hours. I'll be back...

 

[moxhamj]

Right, done it in 7 hours. Have built the circuits, tested them, drawn up schematic, taken some photos and created a website. Link is <A href='http://www.geocities.com/drvernacula/' Target=_Blank>External Web Link</a> Plus as a bonus, as requested by Manuka, pictures of the koala fight that happened last week. Hope this is helpful to the original poster.


Edited by - Dr_Acula on 07/04/2007 07:24:29

 

[BrendanP]

They've sucked your yahoo download limit dry Dr.

Damn, I wanted to see that bear punch on.

&#160;

Edited by - BrendanP on 07/04/2007 08:38:05

 

[manuka]

Dr-A: Great work! I especially chuckled at the sphygmo. being pushed into more imaginative use. If you can rustle all this up in 7 hours during the Aussie heat then methinks you should give up your day job! Hope you don't transfer yourself away from the setup as well... Stan

 

[moxhamj]

Try again Brendan. I didn't read the fine print but it seems I am only allowed 4 site views per hour (4Mb and the pics total 1Mb - apologies for that but compressing them further loses resolution). Fine print always gets you - is it true about that ?urban myth that when you click &quot;I agree&quot; for Microsoft's products you are actually signing up to the US Army?

I'll post an update when the high power modules arrive. Happy Easter.

 

[BrendanP]

Very impressive, both the project and the bear stoush.
Dr. one question, Why did you choose to place the transducer on the pcb rather than mount it at the tank base and run wire from there?
I was thinking it would be better at the base so in case of any water leak you wouldnt have water on the electronics. I might be missing something however.


&#160;

Edited by - BrendanP on 07/04/2007 17:19:44

 

[moxhamj]

No reason why the sensor can't be seperate and I have actually done that on other ones. The aim though is to have no leaks at all and epoxy can easily achieve that. The other big issue is rain, and all electronics needs to be in boxes with the wires coming up from underneath the box. Plus the little screws that come with many project boxes rust after a few years and you can't undo the box so the box needs to be out of the weather as well. A cheap solution is a bit of vertical 90mm downpipe with a cap on top, and mount that on a star dropper.

Makes a nice home for big hairy spiders.

Edited by - Dr_Acula on 08/04/2007 02:00:07

 

[BrendanP]

Now you've really confused me, how could Helen Clark fit into a 90mm tube?

 

 

[alb]

Wow, I go away for Easter, and Dr_Acula knocks up several different ways to fix the problem. Thank you very much all, and especially Dr. Now all I have to do is go away and understand this (expect a bunch of silly questions around Christmas). Seriously though thanks a lot, I will go get some equipment and start fiddling.

I will (eventually) post a note to let you know how it went.

Thanks again

Alastair Breingan

 

[Ray Fitzsimon]

Andrew Bright : I'm interested in using a piezo as a pressure transducer. Can you help with a circuit? I hope to marry it to the Picaxe datalogger.

Re: water level sensors - See silcon chip magazine which had an article about using washing machine water level sensors - They are cheap, sensitive, robust , high current . What more could you want !
More recently Apr 2007 there was a clever circuit idea by Phillip Foote for an optical switch using an immersed LED and photodiode. The above are ok for level triggers at set points but may not be entirely suitable for measuring various tank levels.

Oh ! another idea ive heard of on a remote cattle water trough is a simple water level switch controling the power to a spare mobile phone. If you call the phone and hear it ringing then the water level is ok. If you call and find the phone is turned off - the water level is too low (or there is a problem with the battery/phone which requires a site visit anyway).

cheers . Ray

 

[moxhamj]

High power 433 RF modules have just arrived and have spent an evening doing experiments. One catch - the 100mW of RF gets in all over the place - even two picaxes directly wired to each other can't talk if the RF is on. But I think I am close to solving that.

Have been playing with sending the data directly at 1200 baud, and also a much slower system that sends tones at 2Khz and 200Hz and uses the HP and LP of a state variable filter with a centre frequency of 1Khz, passed through two rectifiers and thence into the + and - of an op amp. That seems to work too.

I'll post some info once things are stable. Hopefully can then get some range data.

 

[moxhamj]

Using the 100mw modules have just got 500 metres through trees and non direct line of sight using 1/4 wave antennas. Is just on the limit - sometimes data gets there and sometimes not. Next experiment = dipole then yagi.

Am driving the modules directly from picaxe and making sure all the 1's and 0's are in equal number by sending a byte then the next byte is 255 minus that byte.

The transmission is ultra reliable at 300 metres.

 

[moxhamj]

Further experiments with a yagi have shown that a simple picaxe led flasher cannot function sitting in the middle of a yagi's RF field with these 100mW modules. The RF gets everywhere, and even analogue circuits don't work properly. Have tried seperating the picaxe and the RF module by 3 metres but that didn't work with a common 5V supply.

What has worked is a 9V battery running a picaxe at the base of the antenna pole via a 78L05, and the RF module up on the antenna running off its own 78L05. The three wires going to the module (Gnd, 9V and data) go through a couple of turns round a toroid to filter RF.

One could run the RF module on the ground and run coax up to the antenna, but there are problems with impedence matching and baluns. Having the module at the dipole keeps losses to a minimum.

I'm going to keep persevering with this, partly because I need something similar, and also to help the original poster who needed a reliable connection over some distance. Will post some range data soon with the 7 element yagi's.

I'll put all this on a website once it is working.

Edited by - Dr_Acula on 22/04/2007 04:09:13

 

[stjosephsopunake]

This is NOT good for long term static pressure sensing situations but it does give some interesting results... It would be especially good for picking up pressure fluctuations (= water flow)

The piezo element is basically the guts out of a standard piezo sounder. You can get them separate. Have a look at some of Rev Ed projects. They have some interesting characteristics... like putting out enough energy when flicked to drive a U Bright LED.

I have used them as strain gauges, movement sensors in kids projects. If you blu-tac one firmly onto a plastic ruler it makes a crude beam balance and you can measure changes less than a gram.

They make a very good dynamic accelerometer putting out VOLTS of signal proportional to change in pressure, strain, bending movement. The charge will eventually leaks away. Note you will end up with a &#8211;ve or +ve voltage swing depending on which way pressure / bending moment is applied

You will need to reset the voltage across the piezo with a &#8220;low&#8221; command to re zero before applying pressure and reading. I have made a live zero / floating setup by pre loading the piezo with a burst of pwm via a suitable resistor to set the charge across the piezo to say a mid point voltage and then monitored the voltage change over time from there. Dv/Dt

To &quot;de-tune&quot; them and make the output signal hang around for seconds or even minutes you just put a capacitor across them. Obviously this will reduce the OP volts but you will end up with a sample and hold effect that lasts easily long enough to measure with readadc.

Here is the guts of a quick sample and hold pressure transducer we made up: We put up to a Bar across the disc we had. Note the &#8211;ve return via the metalwork.

http://picaxe08.orcon.net.nz/Photos/Transducers/Piezo_Pressure_1.jpg
http://picaxe08.orcon.net.nz/Photos/Transducers/Piezo_Pressure_2.jpg
http://picaxe08.orcon.net.nz/Photos/Transducers/Piezo_Pressure_3.jpg

And this proto setup could be made into a dynamic accelerometer bu just adding a weight to the edge of the piezo disc.
http://picaxe08.orcon.net.nz/Photos/Transducers/Force_Meter.jpg

- Andrew