question

rh5555 avatar image
rh5555 asked

BMV-712 SOC during charging reads too high

I have been testing various converter setups to charge 2 6V AGM batteries connected in series. Total capacity is 225Ah, Peukert exponent = 1.12. I discharged the batteries to 20% state of charge, then connected the converter and monitored battery voltage, charge current and SOC as reported by the BMV-712. I also calculated the state of charge using the Coulomb counting method, assuming a charging efficiency of 95%. I would have expected my calculated SOC and the SOC reported by the BMV-712 to match, but they don't:

The BMV-712 SOC reads consistently high and reaches 100% while the converter is still in bulk charge mode. My calculated values seem much more reasonable and believable. Interestingly, if I set the charging efficiency in my calculations to 120% (crazy, I know) then my calculation of SOC matches the BMV-712 exactly.

Anyone any idea what is going on here?

SOC
victron.jpg (120.7 KiB)
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

3 Answers
thomassolar avatar image
thomassolar answered ยท

I can not see anywhere you mention the "charge efficiency setting" in the BMV..? (Only in your calculation "outside" the BMV you talk about that) You only mention adjusting the Peukert.

Maybe that efficiency setting is too high?

2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

rh5555 avatar image
rh5555 answered ยท

The charge efficiency setting in the BMV is 95%, which is the default value. If I were to set that to 79%, I could fool the BMV into reporting the SOC that I measured, but that seems to be rather extreme. I'd really like to know how Victron are calculating SOC in the BMV-712.

6 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Gregory Cavanagh avatar image Gregory Cavanagh commented ยท

Why do you think that is extreme and fooling it?

Thats what the setting is there for?

0 Likes 0 ยท
rh5555 avatar image rh5555 Gregory Cavanagh commented ยท

I say "fooling it" because, based on my Coulomb counting, the actual charging efficiency is close to 99% (these are brand new AGM batteries). With cells discharged by 80% (180Ah discharged out of 225Ah), the reported SOC regains 100% after just 150Ah of charge is delivered. So the BMV seems to be working with an effective charging efficiency of 180/150 = 120%. Setting the charging efficiency parameter to ~80% would just be offsetting this apparent internal error. Hence the fooling. I would much prefer to understand how the SOC is being calculated.

0 Likes 0 ยท
JohnC avatar image JohnC โ™ฆ rh5555 commented ยท

Hi RH. You're expecting too much of your batts, new or not. If you were to watch/log them over a long period, you'd be fortunate to see 80% OUT vs IN. I seem to see about 75% on ave with my fla's, right from new.

Disappointing yes, but it is as it is.

Victron aren't too keen on releasing specific details of their algorithms, understandably so for many reasons. But the BMV manual has a few clues, even how to calculate your own Puekert from batt specs. Charge efficiency I'd expect to be applied just on charge, and Puekert both on charge and discharge. That's not gospel, just a guess.

From memory, bmv rte 95% and Puek 1.25 is a recommended start point for pb's. Adjust to suit as you need, take a deep breath and resign yourself to the real world.

Remember too that pb's never really get to true 100% with daily-cyclers, but a 'practical' SOC is used and synced to.

0 Likes 0 ยท
rh5555 avatar image rh5555 JohnC โ™ฆ commented ยท

Hi John, Interesting that you should mention the Puekert correction (which I did calculate from supplied battery specs). The correction should only be applied on discharge, not on charge. If was applicable to charging, then I could invent a perpetual motion machine: Charge the cell at high current, where its capacity is less, then discharge at low current when (magically) its capacity is greater. I've just generated some free charge out of nothing! But this got me thinking, and I re-did my Coulomb counting calculation incorrectly applying the Peukert correction (as well as the charging efficiency):

Wow! Now my calculations match up with Victron's SOC.

This could be a coincidence, I need to do some more experiments to verify. First will be to discharge with correctly set Peukert, then change Peukert to 1 (which disables the correction) before doing my measured re-scharge. Stay tuned.

0 Likes 0 ยท
victron2.jpg (119.5 KiB)
vineyardgray avatar image vineyardgray rh5555 commented ยท

Wow - that is some compelling research and a valid argument regarding Puekert. Interested to see how this plays out. I guess I'm glad that my BMV-712 is not driving my Multiplus charger :)

0 Likes 0 ยท
JohnC avatar image JohnC โ™ฆ rh5555 commented ยท

Yeh, I admire your dedication. There's some well-explained stuff on Peukert in this link (more in the index), and where you might get tripped up with the calcs. http://www.smartgauge.co.uk/peukert3.html

You're right of course about it only being applied on discharge, and I think Victron have said that somewhere too.

And just for fun, the rte divided by Peuk for me gives 95/1.25 = 76%. About what I see, could it be that simple?

Just food for thought. Good luck with your research.



0 Likes 0 ยท
rh5555 avatar image
rh5555 answered ยท

Just a follow-up on this issue. I recently re-ran this test after upgrading the firmware in the BMV-712 to 4.07 and the issue has been resolved. We now agree 100% on the SOC during charging. I had tried to bring this issue to Victron's attention, but never seemed to gain any traction - I'd send information, they'd be quiet. I'd call, they wouldn't be available. Seems like someone was listening though.

1 comment
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

JohnC avatar image JohnC โ™ฆ commented ยท
Hi again @rh5555 Welcome back.

I went for a look through the firmware changelogs and can't see anything reported about algorithm changes. Nor in the Smartshunt's.

What you may be noticing is a seasonal change, like battery temperature and solar charge rates. I've learnt a little from my own system since I last posted here - notably, overall efficiency of pb's improves in cold weather. This might be a combo of chemistry and slower solar charge. Or other things too.

In the cold I have to raise set charge efficiency from about 91% to 94%, then if that's not enough I raise Peukert from my specced 1.16 to 1.20 (my loads are way less than C20).

Temperature is underrated as a factor, but Peukert is known to be subject to (unquantified) temp variations, and all chemical reactions slow in the cold.

So we have to learn to live with it. I was quite surprised recently after a two week bout of cold and foul solar weather, that when I managed to get back to full charge I was within 0.3% of reported SOC. I can live with that, but I expect to need to change settings as summer comes on. Now all I want is for Victron Connect to report SOC to 1 decimal place rather than single digits.

And oh, the poor rte readings I reported earlier weren't really that bad (a VRM thing). The shunt readings vary from ~81% summer to ~83% winter. Still not good, but at least something to reference as they age.

Thanks for reporting back though. I wish you well..



0 Likes 0 ยท

Related Resources

Additional resources still need to be added for this topic