Author Topic: Determining state of charge (Read 3477 times)

madlabs · « **on:** April 12, 2009, 03:05:49 PM »

Hi All,

Well, my lawnmower/alternator rig is coming along. Aside from a little carburetor issue, I have hopes to run and test the rig a little today, although it may have to wait for a new carb. Anyhoo...

I was wondering about determining state of charge while charging. I see all the methods discussed involve resting the battery with no load then reading the voltage or specific gravity. What I can't seem to find is how a "smart" charger does it while charging. I am going to be using the 50/80 charge regimine that we discussed in another thread.

I got an alternator with an external reg so I can build my own. For todays tests (if the dang motor runs) I'll be using a 25 ohm power pot to set/limit current draw. I found a schematic for a simple charge controller in an old issue of HomePower that uses a 555 and an LM723. I can build one of those, but I'd rather make an MCU powered version. I would also like to incorporate throttle control, and I have some solenoids that would work for that. My question about that what is the best feedback? Is it simply RPMs? Meaning that I set an rpm that is best, then have the MCU maintain that RPM via throttle control? I have some Hall type sensors I can use, and some small NIB magnets that I could epoxy to the motor pulley. I also have some vacumn sensors, although I'm not sure where I could port them in.

T -30 days to off grid and counting...

bob g · « **Reply #1 on:** April 12, 2009, 09:26:48 AM »

there is no good way of determining state of charge while under charge

that i know of, but

seeing how you want to go with the micro controller this is something you might look

into

take a current sensor and insert it between the battery and the load, convert the amount of current to a digital value, and factored with time will provide a amp/hour

figure that has been taken out

use the inverse to put back, plus may 15%

so basically if you use 100amp/hrs the micro has read and stored that, it then

sets out to track what is put back in and adds another 15% to get the job done.

that 15% maybe be much less using the 50/80 regime, and certainly a factor when you top off the batteries ever week to 10 days.

a simpler way is by emperical evidence, if after a time you find that one hours charge time gets you back to 80% SOC (as measured by temp compensated specific gravity, after the battery has rested for 24hours), then you know what you are doing is pretty close.

then there are of course watt/hr or amp/hr meters that measure what went out of the batteries, and also what is put back, that is very effective but will require some oversite on your part as well.

bob g

madlabs · « **Reply #2 on:** April 12, 2009, 10:02:10 AM »

Bob,

I was having similar thoughts about tracking aH in and out, which would be easy. I found in my junk a nice 250amp/50mV current shunt, dug out a meter and made it all work. So I can manually watch the amp flow. Reading it with an MCU would be a bit of a pain to get decent resolution. I have some hall effect current sensors, but not in the needed range, so I guess I'll need to get one.

The trouble with resting for 24 hours is that I will of course have no electricity during those times. I can time it with going out of town I suppose, but that doesn't happen all that often.

I still wonder what "smart" chargers do.

Flux · « **Reply #3 on:** April 12, 2009, 10:10:00 AM »

I don't know any method of measuring state of charge from volts when charging either but you can read sg, you don't have to wait for the thing to stand off charge as you do with volts.

I wouldn't get too technical with things, if you use a rheostat for the field you will have fairly constant current and when the volts start rising above about 14.4 you are near gassing and nearly charged. Every few weeks you can equalise and lower the current and let things gas until the sg doesn't rise after a further hours charge.

It is not cost effective to do a complete charge every time, the final few % take for ever and have to be at low current.

Flux

bob g · « **Reply #4 on:** April 12, 2009, 10:27:00 AM »

i suspect most of what is called a smart charger is marketing hype

sure it is easy to make a charger smarter than a simple line cord, transformer and rectifier

having a decent regulator that provides a couple steps of control is certainly smarter than the old style non controlled chargers.

also i suppose they could measure internal resistance and come to some sort of determination of state of charge?

who knows, i don't think it is all that necessary

letting the battery sit over night once in a while will likely give you an idea what is happening.

alteratively, if you charge to a rate that freely gasses and mixes the electrolyte

taking a measurement of the specific gravity (temp compensated) will likely get you very close in my opinion.

bob g

Flux · « **Reply #5 on:** April 12, 2009, 10:48:18 AM »

"I still wonder what "smart" chargers do."

They cater for a different world. They all cater for the case where you discharge a battery then charge it under controlled conditions to get it fully charged.

This is a fairly easy and straightforward case but no use to you. When it is being charged you will have no power.

The awkward bit is for the type of duty where you are charging and discharging as and when you can. This is particularly true for wind or solar but even with your engine charging I expect you will be wanting to use power as you are charging and if you have to use no load during a complete charge cycle it will not be very convenient.

The AH meter approach is the only one other than sg readings that will reasonably cope with opportunity charging on any sort of automatic basis. Even so such things need temperature compensation and if the current changes much you need compensation for Pewkert effect. Even these have to make some clever decisions to determine full charge every so often as the thing drifts off course. The only defined point you have is full charge and they normally make a decision on current falling below some set value at a certain charge voltage and then add a time factor and assume that is full charge they then reset the counter to zero. Charge efficiency alters with age and you may have to change that factor if things drift too much too often.

Flux

commanda · « **Reply #6 on:** April 12, 2009, 03:37:57 PM »

If you read the datasheet for the UC3906 it will give a good idea how a smart charger works.

You can amplify the voltage from your shunt using a rail to rail opamp. Similar circuit here.

http://www.fieldlines.com/story/2004/9/4/193922/2839

I've just built a SOC meter using a picaxe. Not published here yet as it's still under test. Put it into service last night. If you read the adc repeatedly at a timed interval (I'm using 1/2 second), keep adding those readings in a bucket until it reaches some suitable amount (mine comes to 36000 counts = 0.5 AHr for discharge) then add one to the charge accumulator and subtract 36000 from the bucket. I have 2 ADC inputs, one for charge, one for discharge. I take the difference between the two, and add that to either the charge or discharge bucket. The discharge bucket rolls over at 36000, and the charge bucket at some higher number to account for efficiency.

Battery capacity minus the charge accumulator will give remaining capacity.

When the voltage reaches 14 volts, I reset the capacity remaining to 100%, and disable the accumulators until the voltage falls below 13.2 volts.

Hope that makes sense.

Amanda

wooferhound · « **Reply #7 on:** April 12, 2009, 05:31:00 PM »

As pointed out above, it is difficult to measure SOC using voltage measurements, but this chart may still help you some...

State of Charge (SOC) Chart:

% of Charge- - - - - - - - Charging- - - - - - - At Rest - - - Discharging

- - - - - - - - - - - - - 14.75 - - - - - - - - 12.70 - - - - - - 12.50
- - - - - - - - - - - - - 13.75 - - - - - - - - 12.58 - - - - - - 12.40
- - - - - - - - - - - - - 13.45 - - - - - - - - 12.46 - - - - - - 12.30
- - - - - - - - - - - - - 13.30 - - - - - - - - 12.36 - - - - - - 12.25
- - - - - - - - - - - - - 13.20 - - - - - - - - 12.28 - - - - - - 12.15
- - - - - - - - - - - - - 13.10 - - - - - - - - 12.20 - - - - - - 12.00
- - - - - - - - - - - - - 12.95 - - - - - - - - 12.12 - - - - - - 11.90
- - - - - - - - - - - - - 12.75 - - - - - - - - 12.02 - - - - - - 11.70
- - - - - - - - - - - - - 12.55 - - - - - - - - 11.88 - - - - - - 11.50
- - - - - - - - - - - - - 12.25 - - - - - - - - 11.72 - - - - - - 11.25

http://solarjohn.blogspot.com/2007/03/measuring-battery-state-of-charge.html

tecker · « **Reply #8 on:** April 13, 2009, 03:41:38 AM »

The state of Charge of a good battery is best measured with a hydrometer. However making this measure results in Holly pants syndrome . You can get a measure by current dropoff The magic number for current drop is aroung 2%of the AH rating and that would be a complete charge if limiting the voltage is 14 t 15 volts . I find a really good check is to use the cap Bridge circuit on house current the discharge rate is held to a fairly even cycle rate and when the current drops off the battery is matching to the Cap .You may be able to use the car alt and cap bridge circuit to listen to the rpm increase of the motor . Just pull the bridge of the alt and switch between the Alt bridge and the Cap bridge circuit . Watch the using the bridge of the alt with hose current as some of the diodes have low PIV.

frackers · « **Reply #9 on:** April 13, 2009, 07:24:40 PM »

I'm using a similar approach using the one-wire DS2438 Battery Monitor chip which does full current accumulation on-chip 30 times a second.

It was all working well until my inverter blew up and now after the repair and 4 days successful operation the inverter seems to be interacting with the current measurement in a strange way - it reads 30% high compared to a panel meter that shares the same shunt and what the load actually draws. I guess I have another fault on the inverter that is putting high current pulses onto the batteries that the panel meter misses but the monitor chip is seeing.

I think tonight I'll be putting the forks onto the tractor and bringing the whole battery bank and inverter assembly to the house (its built on a pallet for just this reason!!) so I can get a 'scope onto it.

By the way - I've got a second WRT54GL router set up with OpenWRT software and hardware converted for the irrigation controller...

scottsAI · « **Reply #10 on:** April 13, 2009, 08:12:38 PM »

Not to hard to identify when above 80% SOC.

Article on identifying SOC while charging at different rates.

"Lead-Acid Battery State of Charge vs. Voltage"

http://www.arttec.net/Solar_Mower/4_Electrical/Battery%20Charging.pdf

Notice how the battery voltage jumps when SOC is above 80%.

Think its possible to incorporate it into your controlled charge?

The shunt is not difficult either. Where do you see a problem?

Have fun,

Scott Beversdorf.

scottsAI · « **Reply #11 on:** April 14, 2009, 09:39:55 AM »

Above article talks about the importance of temperature without showing you.

This article shows values.

http://www.powerstream.com/SLA.htm

10C difference may cause the charging to terminate early or later than programed if all your looking at is voltage/current.

BTW, gassing starts above 80% SOC, conveniently the battery voltage jumps up.

Looking for the voltage jump maybe a better indicator than actual voltage. Temperature measuring becomes optional. Varying charging currents will make determining SOC more challenging, not impossible, I think (not done this).

Counting amp hours is OK below 80% SOC, battery efficiency is 91% (typically), once above 80% SOC battery gassing starts, charging efficiency drops to zero at 100% SOC. Not a straight line function. Thus predicting actual SOC based on AH is trouble some. If the efficiency curves are known then it could work. The controller must reset AH during an equalizing charge. Nice article about this:

http://www.sandia.gov/pv/docs/PDF/batpapsteve.pdf

Testing was done with low charge rates.

Have fun,

Scott Beversdorf.

madlabs · « **Reply #12 on:** April 15, 2009, 09:28:54 AM »

Wow, once again a wealth of practical info!

It's no real problem to use an opamp to use the shunt, it's just another thing to make sure isn't affected by noise. The Hall sensors are really easy to use and are easy to read in and out amps. Once again that can be done with the shunt, but I'm soooo lazy.

I see what is said about determining %80 charge. That will be where I start, along with monitoring aH in and out. I'll watch the system when I want to equalize.

Got the rebuild kit for my carb, so if all goes well I'll be testing my lawnmower alternator rig today. Built the platform for my outhouse yesterday and scrounged up a couple of nice windows too.

Anyone here ever use the Homepower Magazine Mark 8 alternator charge regulator? Without bothering to use the calcualtor, anyone know offhand what freq. it runs at? Any idea why it changes freq instead of duty cycle? Is there an ideal freq to run at? My plan would be to use an MCU to generate the PWM and monitor current.

Thanks again all! Such a great info source.

Madscientist267 · « **Reply #13 on:** April 15, 2009, 10:05:08 AM »

One thing I fail to see mentioned much in all these voltages mentioned here is charge rate. Just keep in mind that the rate radically affects terminal voltage at a given SOC.

Fieldlines.com: The Otherpower discussion board

News: