Hi all, I have had a few instances of high voltage alarm with respect to my pylontech US2000Cs. I have had one or two alarms over the last 4 years but recently they are more frequent but no discernible pattern with regard to charging or discharging behaviour. I have read a few threads and maybe understand that an update of the Venus OS might help, I’m currently on 3.41. I would rather not update as these things often bring their own complexities and the system is otherwise working as it should so would rather leave well alone, but if an update would save me from real battery failures then of course I will do it.
Any advice ? Also, i have noticed for a couple of years that one or two batteries’ leds, the charge level ones, were often one led ahead of the other batteries. Now it seems to be only one of the batteries that does that so I am guessing that is the rogue unit.
You are way behind, there are several updates that can help assuming your batteries send min/max cell data, if not it won’t work.
Reality is you probably have some cells on the way out, that generation has plenty of failures.
If you are allowing to balance for a decent amount of time then chances are you should open a pylon ticket and have them investigate for potential rma.
You can artificially drop the CVL via dvcc but that is sticking a bandaid on a problem.
Just go to the current release, there have been 3 revisions to the algo. If effectively adjusts the CVL based on the cell differential. It can help for balancing issues but it won’t fix cells that are on the way out.
Remember, if you have issues you can just boot the backup OS version.
Updating the Venus OS can help reduce the errors updating the firmware of the batteries can also help to improve the behavior, but both will not solve the root cause of the errors.
I also think that your batteries have one or more bad cells.
Just a few weeks ago I had a customer with 8x defective US2000C from early 2022.
As I said though, i have had the odd high voltage alarm even a couple of years ago. And it’s certainly not every 100% charge that sets off the alarm (at least not yet).
I will double check that tomorrow (weather permitting). Cell voltage difference is usually fine, whether heavyish load or charge but when SoC gets to 89% approx. SoC stops going up even with a good charge current then half-an-hour later it shoots to 95ish then sometimes* between 95 and 100% i get these alarms and the single battery blocking charge for a few minutes while the other creep up from 3.45 and this one creeps down from 3.65, then it begins to charge again for a few minutes then the same happens again. Once it gets to 100% all seems fine again.
Sadly, that’s typical behavior for Pylontech. They don’t reduce the charge voltage limit if a cell voltage shoots up. The charger happily tries to raise the pack voltage with whatever current is available, and the BMS can’t cope with the current. Reducing the charge current far enough, (or reducing the charge voltage, which has the same effect) will give the balancer a chance to keep the overshooting cell in check while the rest of them catches up. Newer versions of Venus OS have improved algorithms for this. The 89% wait time is normal, at least for that generation of Pylontech firmware.
Nevertheless, if all cells were of similar health, this problem wouldn’t occur. Like, at all. I’ve got one US 5000 (out of 8) that exhibits this behavior, while the rest of the batteries has never done so.
This afternoon, 97% SoC, min and max are 3.43 and 3.52 at the moment.
Edit: 99%, min is 3.47, max is 3.52.
Edit: 100% still 3.47 and 3.52.
So thankfully not anywhere near 3.65v which I saw the other day. Is it the new algo or is it just a case of random bad behaviour that I have experienced.
Putting a 1kw load on batteries (19kwh) the volts drop fairly quickly to 3.41 and 3.44
Yes, definitely update. It should be at least 3.75, but avoid 3.80. That version has a different issue with Pylontech batteries, which won’t be fixed until 3.82. I don’t know about the versions in between.
I’m also eagerly waiting for the new version, as the charging algorithm has been improved again.
As mentioned, if you run into problems, you can always roll back using the backup function. I’ve done that too, when the new charging algorithms were introduced and my battery took days to finally reach 100% SOC again. You may need to do the balancing for several days after the update as well.
But it’s worth it, and it should be safe enough that there won’t be any more overvoltage warnings.
A cell voltage of 3.65 V is extremely high. It can cause the cells to age very quickly. Updating is cheaper than having to replace the battery, even in the best-case scenario.
Have a read of my post about this. I’m experiencing exactly the same problem, and in my case it’s the internal resistance of individual cells. Also, the SoC apparently isn’t calculated from the charge, but from the cell voltage. So the SoC doesn’t really tell you much when the battery is nearly fully charged. My post: Ursachenanalyse wiederkehrender Ladeabbrüche bei Pylontech-US2000C-Batteriemodulen
I have updated to 3.80 although it has given me several headaches: no more easy-to-see remote console colour display, node-red flows broken, and the remote console crashed today - I don’t think I’ve ever seen that in 4 years on old firmware.
Very interesting and I will be looking forward to reading about your repairs and balancer construction.
I have generally been very kind (I think) to my batteries. I have node-red automations to switch to grid power when SoC is lower than certain levels to avoid cycling them for nothing, and to switch to grid when a certain discharge current level is exceeded at certain times, plus to make sure they hit 100% at least once per 7 days. I assume these things would make life a bit easier on the batteries.
I have had a few high voltage warnings (once in a while) since the beginning (2022) and I’m hoping my recent warnings are the same “rogue” battery although I didn’t pay attention to which battery it was back in the beginning. Either way, it has still kept going and been more or less stable for 4 years despite the warnings.