Temperature compensation at discharge

[Simen]

I'm constructing a temperature compensated dumpload controller (PicAxe. ) that compensates for temperature at 'load on' voltage. I'm wondering if i should use the same temp. comp. values at 'load off' voltage? I'm guessing yes?

A side question for the same topic;
When 'measuring' SoC voltage during load/unloaded at low temperatures; should one temp. compensate with the same values as at charge? (if at all?) (I know; SG are the best way to determine SoC, but it's much more comfortable to take a quick look at my voltmeter in my livingroom when it's -10C outside... )

[SparWeb]

Hopefully you aren't using the temp sensor included with the PICAXE kit?  They don't work well.  Except between 19C and 20C.  After calibration they are dead-on accurate in that range.

State of Charge is not very daunting for Lead-acid batteries, but it's not quite as simple as a spot reading of voltage.  You need to know the history before making that determination.  There are different levels of "truth" as well.

For instance, take a full battery at rest and you just hooked up your datalogger, and the measured voltage is 25.0V or so (using the 24 volt system reference).  If voltage is your logger's only source of information, then it hasn't a clue.  If your logger also knows the charging current has been zero for several hours, then there's more confidence that the state of charge is close to 100%.  If recently the datalogger measured a charging current of (roughly) C/80 that raised the voltage up to 28V, then this is a good sign that the batteries are fully charged, because only a float currrent (very small) was needed to maintain float voltage.  This requires certain knowledge of the battery bank, its capacity, required charging set points (gell-cells use 28.2 while flooded batteries should see more than 29), current and voltage, AND a record of the progress over time.  Hope you're not trying to cram all that info into a 08M micro chip because it doesn't have enough pins.

Even better than that is a constant measure of AH in and AH out of the battery, with periodic status checks to correct the projected SOC no matter how much load and charge are happening.  I can keep making this better and better, but at some point it is good enough.  The more "correct" you want the meter to be, the more work and cost it requires.

This is a funny coincidence Simen.  I've been working on an enhanced datalogger / controller for my solar/wind system for some time now.  I'm using a PICAXE 28x2 and I'm using up pins fast.  I could clean up the drawing and show you a copy without much trouble, if you are interested.

[Simen]

First, i'm not trying to make the 08M to tell me the SoC; it's only there to control the dumpload, and i want it to do it right according to temperature, since i live in a country where the temperature varies from -30C to +30C +/-.

Fyi; i'm using the Dallas 18B20 1-wire digital temperature sensor in 8/12bit mode... (8bit for temp. comp. and 12bit out for logging...)

I am aware of most of the variables it takes to determine SoC relatively precise. What i'm not sure of, are if temperature plays a role when measuring voltage during Discharge? I know that during Charge @ - 4C, one should add a whole 1.0V to the top charge voltage (12V bank), but would similar adjustment apply during discharge at same temperatures?

I Do know my bank rather good; a 1yr old 2x Rolls S-530 bank (2x 6V/530Ah)

The dumpload controller i'm putting together are simple enough; a 08M with a DS18B20 temp sense; 10bit voltage monitoring, and some irfz48 fets at the output driving 2x 16A ceramic heaters (so far.); log output are enabled and goes via usb to a laptop. Later, i'm going to add an X-Bee.

I Do have a 20M2 ready for duty that i'm planning to use for a more advanced controller... (Using the examples over at TheBackShed as basis... )

If it's not too much trouble; please do post your controller. It's always interesting to see.

[SparWeb]

You're way out ahead of me, then.  I can only use the PICAXE in a "cookbook" fashion, but slowly learning how to make stuff work on my own.  I have to admit that I have learned the most from the Backshed projects, too.  Thanks for the temp sensor p/n - it looks much more useful than the picaxe part.

Battery state of charge measurements and the ratings that correspond to them are usually only given at 25C.  Other temperatures require you to infer and calculate.  It can depend on the battery type, too, but at -15C, there are standardized tests for starting batteries (not deep cycle) that require them to have at least 50% of the capacity at 25C.  From the point of view of specific gravity, that too must be adjusted for the climate.  I have another document I can send you, on that subject.  Rather old but there is a lot of information about batteries, particularly about their state of charge and specific gravity explained.  I think it may even be thorough enough to help you make the calculations you want to carry out.  It's a scanned PDF, so I'll get it uploaded somewhere and send you a link shortly.

Yeah, and it was -39C here (Alberta) for 3 consecutive nights, just last month.

[Rover]

SparWeb,

" I'm using a PICAXE 28x2 and I'm using up pins fast." ... this is why I generally use multichannel ADCs etc (SPI, one wire, whatever) , 8 chanel ADC gives you a bit more breathing room. You could also use an I2C bus with the proper ICs (bit slower though but usually not enough to worry about it).

I've moved to 32 bit though so pins are no longer an issue on my micros.

 

[Simen]

Seems i can answer my own question from this document:
http://www.buchanan1.net/lead_acid.shtml

At lower temperatures, total capacity are lower and voltage and SG follows down...

[SparWeb]

Simen,
Excellent finding there (with references to his source, too, wow)

Here is what I have:

http://www.rigel.ca/AABM_Battery_Manual_1964.pdf  (10 MB)

and

http://www.rigel.ca/Battery_Charging_HomePowerMag.pdf (82 kB)

[Simen]

They do mention low temp discharging voltage in the HomePowerMag article...

At 32°F (0°C), the effect of temperature becomes pronounced enough to distinctly change not only the battery voltage vs. SOC profile, but also its useful Ampere-hour capacity. The discharge voltage curves may be depressed by as much as 0.5 VDC from those shown on the graph.

Edit:
That was the information i needed to give me some peace of mind. Looking at the Voltage; 12.2V @ -6C with only a few lights as load DID worry me a bit, since the battery was fully charged the day before..

Here's another good article: http://www.jgdarden.com/batteryfaq/carfaq4.htm#ocv_soc (only SG and Open Voltage under no load though...)

[Simen]

For those interested, this is the table i'm basing the temperature compensation on for the 'dumpload On' voltage...

Charging @12V   Spesific Gravity / SoC
F.    C.   Voltage   100%   75%   50%   25%   0%
-----------------------------------------------------
120 48.9°   -0.672   1.249   1.209   1.174   1.139   1.104
110 43.3°   -0.504   1.253   1.213   1.178   1.143   1.108
100 37.8°   -0.336   1.257   1.217   1.182   1.147   1.112
90  32.2°   -0.168   1.261   1.221   1.186   1.151   1.116
80  26.7°    0   1.265   1.225   1.190   1.155   1.120
70  21.1°   +0.168   1.269   1.229   1.194   1.159   1.124
60  15.6°   +0.336   1.273   1.233   1.198   1.163   1.128
50  10°   +0.504   1.277   1.237   1.202   1.167   1.132
40  4.4°   +0.672   1.281   1.241   1.206   1.171   1.136
30  -1.1°   +0.840   1.285   1.245   1.210   1.175   1.140
20  -6.7°   +1.008   1.289   1.249   1.214   1.179   1.144
10  -12.2°   +1.176   1.293   1.253   1.218   1.183   1.148
0   -17.8°   +1.344   1.297   1.257   1.222   1.187   1.152

I do not remember where i found it, but i think it's generic enough to work for most FLA's...

[SparWeb]

Interesting that site has two different tables, both seemingly for the VRLA batteries. 

This one:   http://www.jgdarden.com/batteryfaq/carfaq4.htm#mf_soc

but the spreadsheet has higher numbers:  http://www.jgdarden.com/batteryfaq/SoC.xls

This could be from different plate types or manufacturer sources, that will vary somewhat.  The data sources on this site aren't clear (one could be Exide again).

[SparWeb]

I was confused by your table, Simen, because the voltages go down with rising temperature... then I realized it's for dump load control, not State-of-charge.  The data on BatteryFAQ (and everything else I've seen) says that battery voltage at rest will rise when temperature rises.

From the data you've dug up, you can use this formula to calculate the 100% SOC voltage at ay given temperature:

=0.0000002328*T^3-0.00005095*T^2+0.0033678*T+12.742

where T is temperature in Celsius and 12.742 is the typical rest voltage at 0C.  Adjust for lower states of charge by subtracting 1V*(1-%) and interpolate linearly between (ie SOC=75% is about 0.25v less than fully charged voltage and SOC=25% is about 0.75v less).  Not exact but close enough I hope!

[SparWeb]

SparWeb,
" I'm using a PICAXE 28x2 and I'm using up pins fast." ... this is why I generally use multichannel ADCs etc (SPI, one wire, whatever) , 8 chanel ADC gives you a bit more breathing room. You could also use an I2C bus with the proper ICs (bit slower though but usually not enough to worry about it).
I've moved to 32 bit though so pins are no longer an issue on my micros.

Hi Rover,
You're so far ahead I didn't really understand you.  The picaxe can be run at 16MHz, so I figure I'll be collecting data at a plenty fast rate (provided I get it to work).  I don't see any problems with that.

What would you recommend for reading (and displaying) the data coming in the serial port?  VB like everyone else?  I've programmed lots of things, but never a serial interface.  I'd like to stop using the interface that comes with the Picaxe because it doesn't write to files automatically.

[Simen]

For my purposes (as of yet), you're complicating things.

Yes, my table are for when to turn on the DumpLoad (when to stop charging according to temp. )

Regarding SoC/Voltage, i just wanted to get an idea of the SoC by glancing at the voltage; knowing it's -6c and knowing i'm using 8-12A and reading 12.2V, i don't want to panic, running out in the snow and hooking up my 12V, 50A grid-tied charger...

Bit nevertheless; this is interesting stuff, and the deeper the understanding, the better; it might be useful in future projects.

I'm close to finish programming the little 08M based Dumpload controller; just missing some more If-Then sentences for the temp.comp. I'll post it here if interest exists.

[Simen]

Sparweb:
I'd probably recommend VB if you're familiar with it...

I'm one of the weird ones, using an obscure language called Clarion... (Had to prototype comm calls from windows dll manually... )

[BrianSmith]

Everything I've ever seen says lead acids should be charged at higher voltages when its colder, lower voltages as it gets hotter.  That's what temperature compensated chargers do with SLA / FLA's I believe unless my dyslexia is acting up again... 

[SparWeb]

No you're not dyslexic.  The rest voltage goes down when it's cold, and to charge the battery you need to maintain a higher float voltage. 
The colder it gets, the farther apart these numbers get, that's all.  I don't know enough chemistry to say why, I just do what I'm told in this domain. 

[Madscientist267]

Simen -

How well does that chart translate to SLA?

Steve

[Simen]

I think you can use that chart for SLA too, since it's the same chemistry as FLA...

[SparWeb]

How well does that chart translate to SLA?

Not perfectly.  Sealed batteries cannot be charged under such high voltage as the flooded types.  And the resting voltage can be a few % higher.

If you go back to the batteryfaq.org website you can compare the FLA and SLA tables to get some idea of this.
Also you can search for charge controller settings and in most books (Xantrex, Outback, Tristar, etc) you can see that the bulk and float charge voltage levels are different depending on battery type.

[Simen]

Sparweb: Do not mix recommended base bulk/float voltages with the temp. comp. chart that i posted.

The +/- voltages in the chart can just as well be used for SLA, as long as one use the recommended bulk and float for the specific battery as a baseline (For example 14.1V @ 25C bulk charge for Gel type SLA...)

It's more important NOT to equalize sla's (much), and most sla manufacturers recommends keeping sla's below 25C; high temp and gassing will evaporate moisture.

I took a quick look at the Morningstar Tristar manual, and in what i could see, they're running the same temp.comp regardless of battery type selected. (for 12V: 0.03V/C (0.017V/F))

As i mentioned earlier; the chemistry are the same, and it behaves similar at different temperatures; it's just different ways the acid are bound and vented.

This is my understanding.

[SparWeb]

Yup.
I went in a different direction (didn't mean to confuse).  I'm only pointing out how there are prickly details with some battery types.  Your batteries sound like the flooded type so my comments about my AGM's shouldn't affect you.  The correction factor is the same, it's the range of temperatures and voltages that you're allowed to use that change - and only in extreme cold and hot do they matter.

[Madscientist267]

So what are you using as you're baseline (80F = 0V) voltage to trip the dump?

14.4?

Steve

[Simen]

Yes, for my Flooded Rolls S530 batteries i'm using 14.4V at 80F (26.7C) as baseline to start the dumping. For cutout i'm using at present 13.2V, which are at the lower end of float voltage. (I'm a bit in doubt when to stop the dumping...)

[Madscientist267]

I like to let it flicker quicker than a christmas tree twinkler...

I've also noticed that when I pulse-discharge at high frequency (ok, 25Hz?) that it has a tendency to 'wake' the battery up, and it seems more active afterward. It may cause the need for more frequent watering, not sure just yet, but basing it on the fact that higher current charging at upper limit voltages causes more gassing. The higher current charge comes from the idea that you 'just discharged' it. And heavily.

Not sure I explained that as well as I could have, but it's late, I'm tired, and tried to just get the concept out...

To accomplish it, trigger threshold @ 14.4, release at ~14.3. All depends on your input power, bank capacity, and dump load.

Steve