maximising the output from solar panels - maybe OT!

[stocky]

Just looking for some *considered* feedback from the community...maybe be OT - maybe not if we can incorporate a PIC(axe) into it!

I have a product running from solar panels and commercial solar reg charging 12v SLA batteries (long distance modular telemetry systems)

I'm redesigning lots of the bits (from the commercial stuff to our own design) to improve power consumption and ease of *end user* install-ability!

Where I am at the moment is the solar reg - we use a *nice* STECA SSR6.6c regulator that incorporates LVD and PWM charging

http://www.steca-solar.com/en/art/uid_kategorien/0000581/id_matchcode/up_solarladeregler/id_artikel/0000037/bop/0/chksum/3a5b0dcd2606edfbe57e3a4c11fa5bdd/beetools.html

These units work great - but at a rating of 6A and a reasonably large physical size I'm looking to see if there is a better way.

Max panel size we use is 20W or an SC current of approx 1.3 amps

I wish to retain the LVD (low volts disconnect) feature to maintain reliable "reboot" of the attached radio transceiver as well as some *basic* charge level indication (like the SSR6)

Basically - is it worth going to the trouble designing out own charge controller to screw a few more mA out of (or in to!) our batteries?

Do I go the "discreet" component way, off the shelf "switcher", boost-buck controller or microcontroller PWM?

Looking around there seems to be a lot of ideas for "micropower" stuff running from "garden light" panels and stuff for 20A+ but nothing around the 1A mark which is where we are aiming for this idea.

With the size panels we are using is it even a real benefit to have all the fancy PWN charging etc - as I trying to go too far? Am i better off looking at a simple analog charge cct with low quiescent and just be happy with that (i wouldnt be happy though - not tech enough lol!)

Overall the unit will be integrated into a DIN rail mounted module that gangs side-by-side with the rest of the I/O modules (some with micros, some without - depends on module function)

One not-so-neat option is to "gut" the SSR-6 and mount it on another PCB - the actual board is HEAPS smaller than the huge box they put it in!

Cheers in advance for any ideas/feedback!

Stocky

 

[Dippy]

Qu'elle coincidence.

I'm just designing a solar charger right now. It switches between various charge schemes depending on the conditions to try and squeeze the pips out of the panel.
And it is ending up much more complex than I first envisaged.
I am using a PIC controlled Buck for the controlled-current, constant voltage and pulse modes. And with the charge current / load current / panel load V / bttey V calcs etc.etc. it's quite convoluted code to get it right. I'm even looking at charge history/ability to get some idea of battery condition too.
So far so good and I've only spent 4 weeks on it

I'd suggest you buy one or gut the old one.

 

[manuka]

I'm an enormous solar PV fan, BUT whenever such applications arise I tend to ask leading Q's such as => Reliability? Budget? Battery types? Seasonal issues? Time frame? Servicing?

Is this intended for the mW level 433MHz link? If so then at first glance a 20W PV should be MORE than enough if feeding a 12 V 7Ah SLA via just a simple regulator. You could run bidge pumps with 20W! Given that you'll inevitably spend time tweaking, I'd say start SIMPLY & learn what your setup needs. As Dippy's 4 weeks initial design tends to testify, you may be well served conducting outdoor trials in the sunshine, especially since multiple or extra PVs (as needed) may be a better approach than a single large one.

At the KISS level,I've quite a soft spot for the 1-2 Watt 12V car solar trickle chargers, which retail here in NZ for ~US$15. Aside from just a simple Si blocking diode, these can be connected directly to a 12 V 7 Ah SLA & rarely threaten overcharge. (A typical days sunshine may bring in 8 Wh, which is ~ 10% of the SLAs 12 x 7 = 84 Wh.) I made one up in this style for a boatie mate to drive a few LEDs etc. AFAIK it's still thriving after ~10 years.

 

[stocky]

Manuka,

I'm an enormous solar PV fan, BUT whenever such applications arise I tend to ask leading Q's such as => Reliability? Budget? Battery types? Seasonal issues? Time frame? Servicing?

100% reliable, sub $50 budget in qty build, SLA batteries, generally outback australia (lots of sun!), timeframe = when i get it done, end user swap out if servicing needed

Is this intended for the mW level 433MHz link? If so then at first glance a 20W PV should be MORE than enough if feeding a 12 V 7Ah SLA via just a simple regulator. You could run bidge pumps with 20W!

5W commerical band UHF - distances in excess of 50km with omni antennas
I'm running my mW units from garden light panels & 2AA nicads

I don't want to change panels or put in a bigger one - rather find a better way to get more from what I have- maybe time to reverse engineer the commerical one (to a point) and work back to the control mech from there

 

[stocky]

I'd suggest you buy one or gut the old one.

Cheers Dippy - i'd buy one if I could find one that will integrate nicely with the rest of the stuff

off to hunt Alibaba.com for a chinese manufacture that we can buy bare boards from maybe.....

 

[westaust55]

Battery charging from solar panels

Stocky,

I have a Stecca 6.6 (6Amp) solar regulator/charger in a camper trailer and a Morningstar (10Amp) solar regulator/charger in my 4WD . Both with LVD feature. Different dimensions but very roughly the same volume.

For something different (ie build your own), have a look at a recent (April 08) Silicon Chip mag. From scale of components still no smaller however.
http://www.siliconchip.com.au/cms/A_110348/article.html

 

[stocky]

Stocky,

I have a Steca 6.6 (6Amp) solar regulator/charger in a camper trailer and a Morningstar (10Amp) solar regulator/charger in my 4WD . Both with LVD feature. Different dimensions but very roughly the same volume.

For something different (ie build your own), have a look at the latest SIlicon Chip mag. From scale of components still no smaller however.
http://www.siliconchip.com.au/cms/A_110348/article.html

I use both the STECA 6.6c & MS Sunsaver regs now - both very good and the Steca is amazing for the $$ - they just dont fit what i'm trying to do.

Thanks for the tip on the SC article - been so busy this month havent had a chance to read mine yet :-(

 

[moxhamj]

I'm using 5W solar panels for telemetry coupled with a 7Ah SLA. I have looked a lot at all the power tweaking and concluded it isn't worth it. Just using fixed voltage at present (13.6V). Not perfect by any means but other factors are far more critical like:
*weathering of all materials - wood in particular
*decomposition of all forms of glue in sunlight
*leakage of water into any part of the system
*leakage of air into any part of the system - mist is essentially water
*condensation in sealed parts of the system and how it collects
*wind damage - I just lost some solar panels after 3 months after UV decomposed the "liquid nails" glue and panels fell off
*UV degradation of most forms of wire insulation
*animals eating and destroying the device, especially cows and koalas
*insects of all sorts colonising the units, particularly bees and ants
*lightning damage
*humans throwing rocks at modules (the bees above turned out to be helpful!)
*ability to reprogram the unit easily
*ability to make modules cheaply and easily (painted wood, galv steel and aluminium)
*waterproof external connectors (not many exist)
*ability to swap in and out various components including the panels and batteries
*not using too big a solar module - too tempting to thieves
*keeping birds from messing solar panels
*getting radio antennas high enough off the ground

After working on all those, the problem of getting another 10% of power out of a solar panel seemed rather trivial. It was cheaper to spend a bit of time hunting for panels that were 10% cheaper.

 

[stocky]

I'm using 5W solar panels for telemetry coupled with a 7Ah SLA. I have looked a lot at all the power tweaking and concluded it isn't worth it. Just using fixed voltage at present (13.6V).

Thats the direction i'm looking now - simple voltage sense & a mosfet
typically the systems see only a low standby drain for 99.9% of the time, so its really a "float" charge scenario. I'm come up with a *basic* circuit to try and will run one side-by-side with a Steca and see how it goes!

*weathering of all materials - wood in particular
*decomposition of all forms of glue in sunlight
*leakage of water into any part of the system
*leakage of air into any part of the system - mist is essentially water
*condensation in sealed parts of the system and how it collects
*wind damage - I just lost some solar panels after 3 months after UV decomposed the "liquid nails" glue and panels fell off
*UV degradation of most forms of wire insulation
*animals eating and destroying the device, especially cows and koalas
*insects of all sorts colonising the units, particularly bees and ants
*lightning damage
*humans throwing rocks at modules (the bees above turned out to be helpful!)
*ability to reprogram the unit easily
*ability to make modules cheaply and easily (painted wood, galv steel and aluminium)
*waterproof external connectors (not many exist)
*ability to swap in and out various components including the panels and batteries
*not using too big a solar module - too tempting to thieves
*keeping birds from messing solar panels
*getting radio antennas high enough off the ground

Much of that we have already seen/experienced/fixed/solved etc etc!
*IP66 rated enclosures - ants/water CANT get it
*100% of all cables conduited (much of it METAL)
*6-8m steel masts for antennas
*over the air configuration and Firmware upgrades
*Corellas & Dingos are my biggest enemies.....not humans! (amazing the lenghts you have to go to keep the corellas at bay!)
*bird mess - amazing how well the things keep going with 50% coverage on the panel - a factor of how much sunlight I have to work with i think!

After working on all those, the problem of getting another 10% of power out of a solar panel seemed rather trivial. It was cheaper to spend a bit of time hunting for panels that were 10% cheaper.

IMHO its not worth buying cheaper panels - the environment is to harsh and they fail - we use only BP or Kyocera now

 

[manuka]

Dr_A: Very well put, & VERY close to my PV experiences.

Stocky: Ahem- your bill of materials is looking quite costly & IMHO professional design looks justified = $$$$($). 5W UHF - is this on 470 MHz PRS CB data channels? Considered a repeater ? Wireless links terrestrial only ? Considered antenna with squashed omnidirectional radiation hence? This alone may greatly passively improve the setup. What are you monitoring!? How far away is the nearest service person? SOMEONE WILL HAVE TO VISIT! As a reliability example I'm reminded of the 1943 German WW2 "Kurt" radio met. station that a U-boat sneaked ashore onto Labrador. It was never detected by the allies, but in spite of an industrial strength design it soon failed. See => http://en.wikipedia.org/wiki/Weather_Station_Kurt

For all the beauty of the Oz. outbacks, it's fair to say some of those who frequent it (both human & animal) have quite a mongrel mentality! Aside from rocks you also need to consider air rifles- when live game are scarce & the hunter's are frustrated, PVs make lovely substitute targets (especially for more powerful firearms). As a single large PV may be ruined by a "bullseye", this alone may justify multiple smaller panels instead.

Bird poo,autumn leaves and lichen growths are well known PV panel maskers here in NZ (~lat 35-45 Sth), & as your your tropical latitudes = more horizontal mounting, such panel deposits may be a significant issue.

EXTRA: For Europeans, those *Corellas * are not Toyota models, but huge flocks of small parrots!
www.pro.corbis.com/images/42-18346149.jpg?size=572&uid={C80877B6-5E00-47AA-94A6-312D8A249732}

 

[stocky]

Ahem- your bill of materials is looking quite costly & IMHO professional design looks justified = $$$$($).

With price i can buy components at should be do-able in my $50 per unit budget - I buy STECA 6.6c for less than HALF of that currently (i buy LOTS)

5W UHF - is this on 470 MHz PRS CB data channels? Considered a repeater ? Wireless links terrestrial only ? Considered antenna with squashed omnidirectional radiation hence?

450-520 commercial band. each site works as a "store & forward" repeater now if required. I try and avoid yagis where possible as it limits further expansion possibilities (i have access to 25w UHF & VHF transceivers if distance becomes an issue - we currently can stretch our 5w *reliably* to nearly 70km with omnis (-117 to -119dbm)).

What are you monitoring!? How far away is the nearest service person? SOMEONE WILL HAVE TO VISIT!

Systems are deployed on large cattle station properties to monitor water and various other things. Usually sites are visited once a week by station staff at a minimum and sites are generally remote from ANY public access. Servicing is possible by a person qualified to drive a screwdriver....I designed ii that way due to the generally available skill levels present on many of the properties!

 

[manuka]

??? Did I read that $50 per unit correctly? Even boring things like guy wires & bolts could be more than that! Real world telemetry setups are usually orders of magnitude more costly...

UHF antenna are pretty compact rugged "plumbers specials". You'll be well advised to review the numerous designs,with of course Yagi types being at the forefront, as they'll greatly ease your overall system power needs. Why do you need omnidirectional if monitoring fixed sites?

 

[moxhamj]

I am very impressed by what you are doing Stocky. A number of people here are working on similar problems and you have solved a number of them. Are you working on something commercial? If so, how much are you willing to share on this forum? My interest is to make my hobby farm work better so it isn't commercial, but maybe if enough people got together it could end up being an interesting joint project.

How are you getting the cost under $50? A 20W module even at wholesale prices is not much under $7/watt.

Re charging, the units I am using are drawing around 8mA and at those currents the voltage on a SLA ends up being a very close approximation to the remaining capacity in the cell. Hence constant voltage charge is a pretty good compromise. But it isn't perfect - for instance there is no temperature compensation. The UC3906 chip has all the algorithms and it could be possible to replicate it in picaxe software but may not end up any cheaper (especially since Altronics sell the complete UC3906 kit). The disadvantage of the UC3906 is it has a quiescent current that is too high for running off a string of solar garden lights that would struggle to get over 20mA in full sun. But for a 20W (even a 5W) panel this would not be an issue.

Buck/Boost is definitely not worthwhile. You only get a tiny bit extra energy at sunrise and sunset. But you find that if a panel has 10V on it and the sun is just peeking over the horizon, as soon as you try to draw some current the volts collapse. The current and volt curves for panels explain this in a more scientific way.

Low dropout disconnect is great (for the battery) but not so good if the critical data is not getting through. Especially if that data is saying when a tank is full. I've gone for grossly oversizing the panel and battery so that dropouts never occur (enough power for a month of no sun).

Can I ask what sort of UHF units you are using and how you are getting them under $50?

 

[stocky]

??? Did I read that $50 per unit correctly? Even boring things like guy wires & bolts could be more than that! Real world telemetry setups are usually orders of magnitude more costly...

UHF antenna are pretty compact rugged "plumbers specials". You'll be well advised to review the numerous designs,with of course Yagi types being at the forefront, as they'll greatly ease your overall system power needs. Why do you need omnidirectional if monitoring fixed sites?

The $50 is for the REGULATOR ONLY - sorry if I have confused you!

We use a range of high quality commercial grade antennas - this is a commercial application after all!

We use omni's as the system uses some sites for "store & forward" repeaters and hence need to be able to RX from any direction - also the systems are generally installed and expanded over a few years - using a yagi on a site effectively puts that site out of action as a potential future repeater site

 

[moxhamj]

I'd agree on the logic of omnis for wireless mesh. The topology is more expandable and more robust. Just out of interest how much is the total cost of one of your repeater stations?

 

[stocky]

I am very impressed by what you are doing Stocky. A number of people here are working on similar problems and you have solved a number of them. Are you working on something commercial? If so, how much are you willing to share on this forum? My interest is to make my hobby farm work better so it isn't commercial, but maybe if enough people got together it could end up being an interesting joint project.

We are using commerical radios for the main transceivers but we are developing a low power version for expanding sensors around a site and to keep cabling to a minimum (hence playing with the HM-TR's!)

How are you getting the cost under $50? A 20W module even at wholesale prices is not much under $7/watt.

See my comment to Manuka - sorry again for any confusion!

Buck/Boost is definitely not worthwhile. You only get a tiny bit extra energy at sunrise and sunset. But you find that if a panel has 10V on it and the sun is just peeking over the horizon, as soon as you try to draw some current the volts collapse. The current and volt curves for panels explain this in a more scientific way.

Thats what i was trying to acheive - basically suck a few more mA of charge early & late in the day.

Low dropout disconnect is great (for the battery) but not so good if the critical data is not getting through. Especially if that data is saying when a tank is full. I've gone for grossly oversizing the panel and battery so that dropouts never occur (enough power for a month of no sun).

Its more a backup *just in case* we get a LV brownout to the radios - just ensures a clean reboot of the radio thats all

Can I ask what sort of UHF units you are using and how you are getting them under $50?

full commercial grade data radios - closer to $500! The $50 was ref to the regulator only

Repeater can be part of ANY site - repeater on its own is about $1500 - but thats complete with IP rated cabinets etc etc etc!

Stocky

 

[moxhamj]

Great - that answers quite a few things. I'm working with repeaters with a range of about 1km and a budget of $150. Each module has one radio and 3 RS232 connections and so the system is expandable. Yours goes a lot further and no doubt is a lot more robust! From a practical perspective radio round a rural property will have the long distance links to stock troughs and the like, and also a need for a smaller lower power mesh, eg round the house and sheds. These shorter links would have smaller panels/batteries etc.

What sort of computer/microcontroller is sending data through the data radios, or do the data radios do all that internally? And how do you interface, say, a pressure sensor into the system?

 

[stocky]

The radios themselves do a lot of the work themselves and thats the reason the links are so robust even down to -117dmb

Support for Digital I/O and analogue inputs is std - we can flash the radios and make them work like *big* HM-TR's as well!

Analog inputs are 0-5v but scalable for 0-10 & 4-20mA without much hassle - software at the base diplays the readings

We do a lot of "custom" stuff to interface real world to the radio - and the PICAXE is THE perfect tool for the job as each project may be a one off!


Typical property size may be in excess of 200km from the homestead to furthest site - bigger in many cases!

 

[moxhamj]

Sounds like the sort of thing people here are trying to do for a lower budget! Yes there would be lots of synergies with picaxe - and lots of opportunities to expand your system. Would it be impolite to ask who you work for?

 

[stocky]

IM sent.....

 

[premelec]

A simple way to come close to maximum power transfer is to regulate the PV panel _output_ voltage to about .43 volts/cell by modulating the load with PWM - this can be refined a bit by sensing the panel temperature as well...

 

[moxhamj]

PM replied - thanks!

Re the original question, the problem of solar power regulation is really two problems - the problem of getting maximum power out of a panel, and the problem of managing lead acid batteries. MPPT trackers etc are all commercial devices and they certainly work, though they are usually meant for big installations eg 1Kw. Battery management is also a complex art - there are commercial solar controllers that even cycle (wet cell) batteries through overcharge periodically to cause a bit of electrolysis and stir up the electrolyte. They quote 7 years lifetime. I don't know how long a battery lasts in a home alarm system but I've had ones last a lot longer than that just sitting on a float. I think the lifetime depends a bit on how much you discharge a battery. So that tends to lead to a solution where the battery is really big and hardly is discharged.

So I came up with 8mA fitted with a 5W panel and a 7Ah battery. Way more than you would expect!

So a very relevant question is; how much current does a typical data radio use when it is sleeping?

 

[stocky]

our commerical ones draw about 60mA at rest - but these still have a usable audio stage that is about to be removed to get the draw lower. Also on the horizon is the ability to sleep the majority of the RX stage to get power consumption even lower

So standby usages is 1.5Ah a day (roughly)
TX bursts to 2Amp but total airtime for a day may be less than 10 seconds

 

[Dippy]

And I've put cheap lead-acids in UPS and they only last 15 months. I have one sitting at my feet right now. Not a patch on the original/expensive UASA supplied which lasted 3 years. My other UPSs have also gone floppy a bit quick because of the cheapness/unsuitability of the battery. BTW: these cheap batteries were claimed to be suitable for UPSss.... so many variables - including 'You Get what you pay for'. Now I wonder where these cheap batteries were made?. And where was the good, reliable, expensive UASA made?

Seriously, Drac, yes, those are some of the considerations that I'm trying to address in my charger design.
The main problem I can forsee is the battery degradation becuase it never gets properly/fully charged. I'm trying to get the most out of the panel, but if the panel/battery ratio is wrong then (assuming 100% sun) the battery may never get fully charged - not good for longevity - or battery gets charged quickly and a lot is wasted. And then, of course, there's the weather.
There is NO WAY you can get it 100%... too many imponderables.

But I've found by switching between different modes of regulation you can get a reasonable efficiency.. so far. It's certainly better/faster than simple pulse charging... so far. My prototype should be able to handle a charge of up to 10 amps.
But I haven't tried it long term. It may explode or conk out.
Assuming it works then it will be more expensive than a simple/cheap pulse charger but I am hoping to improve the lifetime of the (expensive) battery.
Fewer service intervals/fewer battery replacments... how much does that save?
(PS. I'm talking fat batteries not little ones).

Do people care about battery costs if it lasts over a year? I dunno? I would. It somewhat defeats the object of saving money with (cough,cough) 'free' power if you have to spend £60+ per year on a new battery.
I've had this garden spade for 40 years, cost ten shillings ... changed the handle 5 times and the blade three times.

 

[moxhamj]

Good points Dippy. There are so many chargers around and so many batteries. I wonder if there are some objective tests out there - eg we tested these 10 brands of battery with this one charger and they each lasted this long. And, we tested these 10 brands of charger with this one brand of battery and they lasted this long.

My longest experiment is 5 years with an average current draw of 10mA, a simple "disconnect for 10 minutes when over 14V" algorithm and an old car battery. Still going for 5 years after it came out of the car and it was 4 years old then. And it cost nothing. But you couldn't recommend that for a commercial application. Though car batteries might end up cost effective given they are mass produced. 15 months for a UPS battery does seem very short though.

 

[manuka]

Dippy - it's not UASA, but YUASA => www.yuasabatteries.com/

 

[manuka]

Dippy: Ahem- no wonder they're expensive! It's not UASA, but YUASA => www.yuasabatteries.com/

 

[Dippy]

Sorry about bad spelling. I thought it looked funny as I typed it. Early morning dickslecksia I guess. Got muddled up with NASA. I promise never to moan about other people's poor spelling (for a week or two).
Brain needs a recharge with coffee.

Yes, UPS battery very poor. Though obv it's not just the cost, it's 'will it work when the mains fails?'. If it's duff after 15 months then it would have been poor after 12months and my 2 hours of CADding would have been lost.
(Exaggerated for dramatic effect but you know what I mean).

Yes, it's getting complicated to get 'ideal' charging. I have to consider loads of up to 30 amps and very variable charge from Solar as we don't all live in sunny climates, so over-specified panels have to be considered too. Oh my word! What hot FETs I've got.
... I wonder how many kWh of electric £500 will buy?

 

[stocky]

just doing some reverse engineering on a STECA 6.6c regulator

• Custom ASIC "Atonic v3.3" => 28 pin device!!!!

• basic SHUNT type reg
  • IFRZ34N across the panel terminals
  • Varistor/Transorb for "protection"
  • "SUN" LED across input terminals
• TO220 Schottky in +VE rail between shunt & battery
• battery connected with fuse in NEG rail
  • 1N5402 for reverse battery connection protection
• LVD controlled by MOSFET
  • another IFRZ34N
  • 1N5402 across load terminals as protection against inductive loads?
• 3 terminal dual colour LED driven by two "mcu" pins to indicate state of charge
• voltage sense by resistive divider with variable resistor in ground leg of divider.
• Thermistor for temp compensation

This is looking like a possible picaxe project.....and maybe a PIC later on when i get time to learn....

• power picaxe from battery
• run slow clk to save mA
• monitor volts
  • if above say 14.1v turn on shunt, pause and check again
  • if below say 11.0v turn off load till returns to above say 11.8
• turn on a "charged" led

Goals:

• achieve a physical design that fits in with remainder of system => most important bit!
• cut self power consumption (ie get rid of LEDS!)
• attempt to keep as *similar* charge performance as possible

watching regulator in action gleams the following info:

• when battery reached full charge reg pulses shunt @ approx 2hz
  • based on pulsing of LED across panel terminals
  • assume this also limits thermal dissipation in the mosfet
• charge level led is obviously driven by some form of PWM
  • LED flickers
  • colour gradually changes from RED, thru yellow to GREEN

further info to follow once I have drawn out the rest of the cct - and redrawn it to make sense!

 

[hax]

1.5 Ah a day at 12V = 18 Watt hours

18 x 60 minutes x 60 seconds =

= 64800 joules storage required per day.


Ultracapacitors may be worth a mention here.

Maxwell have a 2.7V 3 kilofarad capacitor available for AUD $100. It holds 10935 joules.

So 6 ultracapacitors would do the trick.

Yes, $600 is a big outlay, but it will pretty much work forever without any maintenance.

 

[stocky]

found some info on temp compensation - just have to work out best way to implement without (too much) calibration.

Looking at LM50 or LM75 temp sensors - nice *predicable* output to work with.

 

[BCJKiwi]

Issue with UltraCaps is that they don't behave like Batteries in that the voltage drops more or less linearly with drop in "charge" so to maintain a usable voltage you either need a lot more storage, or, a voltage boosting circuit to compensate for the dropping voltage in the Ultracap.

 

[stocky]

Ultra caps are NOT an option for this app

$600 of ultra caps vs $50 for a 12Ah SLA.........

 

[stocky]

08M based shunt regulator Ver 1.0 code is 70 bytes

Now to add temp compensation...........

 

[Dippy]

Blinking heck!
I'd better check this out as mine is 11K bytes and I haven't finished...
I hope to optimise it down to 18 bytes.

 

[stocky]

Ver 1.0 is pretty simplistic........

• START
• read battery volts
  • if above "LVD_ON_value" turn on LOAD
• LOOP
  • read battery volts
  • if below "LVD_OFF_value" turn off load
  • if above "LVD_ON_VALUE" turn on load
  • if above "SHUNT_VALUE" turn on SHUNT & Charged LED
  • if below "CHARGED_value" turn off led
  • LOOP

Next to build in temp compensation calc
bit of maths with an ADC input from a LM50 temp sensor (or NTC)

it gives me a point to work from and is based on reverse engineered commerical reg cct - minus some "pretty" bits (that waste mA!)

 

[stocky]

BTW - its up to 84 bytes now with the extra code for CHARGED led

 

[Dippy]

I take it you're only using a baby solar panel?

 

[stocky]

10w-20w max (1.5amp max)

shunt transistor will be IRFZ34N

Ver 1.0 is to see how it works in comparison to commercial reg - fine tune from there

 

[moxhamj]

Just for interest have a read of the UC3906 charger - http://www.a-aengineering.com/UC3906.htm and also http://focus.ti.com/docs/prod/folders/print/uc3906.html (click on the pdf and wait a bit as it is 3Mb). There are some good descriptions of the complexities of charging. Nothing that can't be replicated with a picaxe.

Another design option - if you have a 20W panel, consider using the panel to power up the charger - whether it is the picaxe or UC3906. The usual way is to power the charger all the time but that is wasting power at night. So in the morning, the charger wakes up when the sun comes up, measures the battery, goes into charge mode (as the battery will have flattened a bit overnight), then when it is charged it goes into float mode. It doesn't matter if lots of leds are on as they are only on when the sun is shining. Downside might be how it fails in partial sun but I suspect the UC3906 fails gracefully.