I didn't top balance my packs before installing

[Joeboy]

1500W charger ordered.

[Andy]

Now I've had cell level history graphs for a few days, I have some thoughts. The screen shots below are from one of my earlier packs. There is a misbalance particularly in the mid cells. ie those in the middle of the series. This will have been caused by the extra resistance. The voltage slowly drops as you get further down the packs. I think this has been exacerbated by the balance settings I had. I left the .01V threshold and I think I had 3.35V start balance. The trouble is if you have cells that are lower charge than the rest of the pack they go higher whilst charging. So in normal operations, they get balanced down over a longer period than they get balanced upwards near 100%. (so to speak) . In the second image you can see how the pink yellow cells drop back below the others.

So the settings are very important. On the worst packs, I've now got my balance set to not above .025V ( the green dashed line) to try and bring it back in. I'll tweak it again as they improve. The important bit is to make sure that balance threshold is not exceeded during the charging period. So either have a very high start balance voltage or a high enough delta that exceeds the delta during charge.

There also seems to be a chemistry thing going on around 3.29V where the cells drop off a bit faster. If a cell is lower charge then it dives down at this point much quicker so you can see the cells that need charging. You can also get a rough guess of how much charging is required. I think about 66hours at 2A for this pack. So I'm sat nannying it with my 10A charger. It was an amazon cheapie that has produced lots of magic smoke according to one reviewer so I'm not leaving it on its own This splitting off disappeared after completely balancing one of the better packs.
Hope that makes sense and helps someone.

[Joeboy]

Is there and I mean this with the greatest respect, any chance you are overthinking this? No worries if it's a big thing and what you're into but does it matter that much?

Bear in mind I'm a 10% to 100% soc type and my P-techs are at 87% soh after 8.5 years heavy use. Not making it an abuse competition but that seems fair use for many cycles.

I'm asking where are the gains in worrying? (But nicely).

[Andy]

I hear you. But I think if I'd left it as it was, it was slowly pushing the cells more and more out of balance. Ultimately this would have led to the BMS shutting down the pack well before it should have. The settings are more important than I thought as I was unbalancing my pack rather than balancing.

I've finally set my pack to 10% as well as I can't get through the day otherwise. I'm going to need 4kW of extra charger as well. Good to hear the pylon tech are lasting so well

[Joeboy]

I hear you. But I think if I'd left it as it was, it was slowly pushing the cells more and more out of balance. Ultimately this would have led to the BMS shutting down the pack well before it should have. The settings are more important than I thought as I was unbalancing my pack rather than balancing.

I've finally set my pack to 10% as well as I can't get through the day otherwise. I'm going to need 4kW of extra charger as well.

I'll be honest and say I'm a complete slacker at checking my cell V. Must be 4 to 6 months since I last did it. I'll run across them tomorrow and see where we are. A good reminder to do this, thanks Andy.

[Saladin]

The nature of a battery internal resistance is the outer cells always work harder. There's no way around this...well not with a chemical battery anyway..I suppose fuel cells etc... ...

If you've got active balancing that's the best you can do to compensate.

If you want to go uber-nerd split the battery once a year. Use a CC CV PSU to force every cell individually to ~3.55V at 50mA tail current to erase any memory effect because that'll create an imbalance by nature you'll see the voltage bump down and the current ramp every time you overcome a false trigger . Then parallel the lot and push them to 3.6v as a parallel group. Let them rest for a day in parallel. Then remove the parallel bus connectors and Isolate them for another day. Following this take no load resting voltage readings once they've settled. Pair weakest to strongest and middlest to middlest if you are running parallels. Then put the strongest cells outermost (closest to the feeder cables pos & neg.) progressively working to and weakest innermosts. Law of averages you can unify the herd most evenly this way. Cycle battery to down to 10% & back 100% before returning to service.

It's a right pain in the bottom, get a good book.

It's not possible to accurately measure cell voltage with active cell balancers running because they're effectively buck-boost regulators contaminating the data.

Never turn off the PSU with the battery connected to the output you can damage it without the opposing voltage..you can use a diode to protect it but that'll mess with the meter accuracy. That said you need an highly accurate meter for this, it's unlikely a PSU display meter will have high enough resolution.
If you have multiple PSUs then you can speed up the process by doing several cells at a time. You can also use overvoltage as long as you don't overshoot charge termination tail current {50mA @ 3.55V}.

[Andy]

I hear you. But I think if I'd left it as it was, it was slowly pushing the cells more and more out of balance. Ultimately this would have led to the BMS shutting down the pack well before it should have. The settings are more important than I thought as I was unbalancing my pack rather than balancing.

I've finally set my pack to 10% as well as I can't get through the day otherwise. I'm going to need 4kW of extra charger as well.

I'll be honest and say I'm a complete slacker at checking my cell V. Must be 4 to 6 months since I last did it. I'll run across them tomorrow and see where we are. A good reminder to do this, thanks Andy.

I forgot I had captured my previous pylon tech info as I wasn't graphing it at that time. They were perfectly balanced with almost zero delta after a year of use. I know you have other cells as well. I think whether they get imbalanced or not depends on luck with regards which cells are ahead or behind when you first build the pack. Depending on your settings and usage pattern you can drive them apart or bring them back together.

[Stinsy]

I forgot I had captured my previous pylon tech info as I wasn't graphing it at that time. They were perfectly balanced with almost zero delta after a year of use. I know you have other cells as well. I think whether they get imbalanced or not depends on luck with regards which cells are ahead or behind when you first build the pack. Depending on your settings and usage pattern you can drive them apart or bring them back together.

^This.

Having very well-matched cells is the first and most important factor in keeping balance. Pylontech as the manufacturer of the cells as well as the assembler of the pack is very well-placed to ensure cells are alike. Back in the day you'd measure huge numbers of cells and select those most similar to each other to build a pack. I have no idea if this is still done, maybe batteries are better matched straight off the line.

After that you have balancing. This almost always means passive balancing at the top end where any cell over a set voltage has energy drained off it through a resistor. Pretty much all BEVs use this kind of balancing. Active balancing is much better (where charge is taken from the highest cell(s) and given to the lowest cell(s)). All of these forms of balancing tend to be very slow indeed and can fail to keep up if cells are determined to get out of sync.

[AlBargey]

You can also use overvoltage as long as you don't overshoot charge termination tail current {50mA @ 3.55V}.

I would urge readers not to do this! Unless this post was for people with only 15a/h cells, then a 50mA tail current @ 3.55V would be fine, but definately not for Lf280K or MB31's!

Look at the cell data sheets for manufacturer recommendation, most like EVE, Gotion etc all say terminate charge current when it drops to:

0.05C @ 3.65V which = 14a for a 280a/h cell

which is equivalent to:

0.032C @ 3.55V which = 8.96a for a 280a/h cell

Otherwise the cells will be overcharged!

[Saladin]

Oops!



Reads fine print....{lead-acid brain.. }
The rest of the practice is sound.