Different cell capacities after battery ECU short circuit - advice re: fix?

I have a slightly unusual problem with my HV battery on my 2008 Prius;

My HV ECU shorted inside the HV battery
ECU at the voltage sense connector; the pins for banks 1 and 2 vaporised as did the pin for bank 7 (or 8?). The short occured in a way that makes no sense (no path for current but a plasma arc appeared and burnt for a bit, leaving metal vapor deposited inside the ECU case).

Blocks 1 and 2 had lost 1 volt each (out of around 15 volts or so, as each block is 2 slabs).

I plugged in a spare ECU which works fine without fixing anything else; I drove around discharging and force charging, and have managed to get the cells back in line again - blocks 1 and 2 are now much closer to the rest of the cells.

The problem now is that the two blocks that got shorted and would have had ridiculous currents for a few milliseconds are stronger than the rest of the battery! As the battery reaches 80% SOC, the first two blocks are 0.2v under the rest, but by the time it gets to 15% SOC they are 0.25v above the rest; they seem to have a higher capacity now, ?due to the short circuit rejuvenating them.

I have a basic grid charger and can make a light bulb dischargedr; is it likely that doing a few cycles can bring the rest of the battery up to the level of the ‘improved’ shorted cells? I have a digital charger and discharger which I can use to do one cell at a time if that is better (I have enough time and a spare car!)

Thanks for your advice!

 

The latest blown up ECU was actually mine! The sense lines and battery are intact; modules 1 and 2 (hence the first four slabs) lost 1 volt during the plasma arcing episode!

The blown up ECU works except for the vaporised connector lines; I’m going to repair it but want to look at it with a microscope first; I am using an undamaged ECU in the meantime. The white connector on the damaged ECU was fine - the orange voltage sense connector has a little damage to the plastic, the main issue being the vaporised pins inside the ECU.

I ‘rebalanced’ the cells using the car itself as described above; what I have found is that the cells involved in the short have a higher capacity in that, sitting and watching all the battery modules in real-time on my phone (EngineLink app), the previously shorted cells seem to charge more slowly as the pack reaches 80% SOC, specifically the voltage rises more slowly.....

The fact that on discharge, the voltage falls more slowly, so that at 15% SOC, they have a higher voltage than the rest of the pack, seems to suggest they still have some charge left compared to the rest of the pack.

I haven’t got round to downloading techstream; I have some old laptops but will probably use it with a VM; the car isn’t throwing codes anymore so I don’t need it for that.

 

I just sat in the car for two hours and did some cycles from 7.5% to 85% SoC (7.5% was scary but I did have a grid charger in case it wouldn't power up); 8% if where the HV battery relay clicks off - but even then it lets you start it (maybe just one more time!).

I set up my app on an iPad (with Englinelink + wifi OBD2) and had the cell voltages, delta, energy in/out, engine speed, battery temp, and recorded the screen during 2 cycles; when going to the lowest charge level I let it drain at 180 watts max and watched for any signs of cell reversal. The two shorted cells definitely have more capacity, but they don't actually throw any codes, so it is OK for now. I am going to inspect the battery (still haven't opened the main case yet!). I don't think it will be pretty given the high (ish) mileage (270,000km approx), but I want to see how much life I can get out of the original battery, as it hasn't let me down yet (besides the ECU!)

Regarding the battery ECU connector fires thread, I will post there again once my digital microscope arrives; I am pretty sure I have figured out what is happening and why, but need to look at it and get some high-res pics to make sure it is what I think it is!

 

Apologies re: the three threads; the only two threads I will post on will be the HV connector thread and this one - but I don't want to spam the HV connector thread with my experiences as it is too important for that; but will post up photos when my digital microscope arrives, as I think I know the problem - and, I think it is going to happen to all Priuses and likely most other cars with HV batteries, but will take longer in drier climates. But it is just a theory so I am leaving it there for now!

I live in New Zealand; it isn't humid here, but it is small enough that the sea is always within 100km or so, so salt is probably an issue. I live further north, so there is no salting of the roads in winter like in the UK; one thing I like about NZ is the fact cars basically don't rust (in the UK rust was the number 1 killer of cars - which made scrapyards great for getting engine parts)

Regarding my in-car balancing attempt;

Using Enginelink, it didn't want charge above 80% true SoC, but I found that, even if the multi info display on the dash didn't show the battery charging, it was actually charging slowly (300 watts - 1.2kw approx). I managed to get it it 85% SoC but that is where the 'hard' cut off appears to happen. I had the air con running during this time (in fact the air con took enough power from the MG1 (or 2?) and that may have been what allowed the battery to continue charging slowly.

The discharge to 7.5% was done at the risk of having a temporarily dead car; I do however have a grid charger which is why I dared do it in the first place. Basically all I did, was put the air-conditioning on at a medium level, drawing air in from outside to further assist battery cooling, and once it got down to about 3 bars on the car's display, i switched the air-con off, and used the headlights; towards the end I had everything off on the car, so it was draining at 0.18kw. I kept an eye on all the modules to (hopefully) make sure none of the blocks went low enough that a cell reversal was possible.

The car switches the engine on automatically at (I think 30%, maybe higher, SoC, to prevent it going low). All I did, was to leave the parking brake on, leave the car in neutral. It complained, later gave a red triangle and told me to put it out of neutral, and I managed to get both the car with ! mark and a 'main' battery symbol on the touchscreen. At exactly 8% SoC, the relays in the back clicked (my car is stripped down for repairs etc, so the relays are loud), and the HV battery wasn't draining or charging. It dropped to 7.5% and I shut the car off, expecting it to be immovable without my grid charger...

When I started it up immediately after it actually started the engine. Mine is a 2007 so I think that Toyota may have put more 'backup' in their software to prevent people becoming fully stranded! Since it took ages to get down to 8% and I was amazed it still started.

The four modules (blocks 1 and 2) are still 'stronger' but not so out-of-balance that they throw any codes, which is good - the battery ECU seems to use all the other modules which are balanced but weaker to decide SoC, so the battery should be good for a while longer.

PS : As a warning to others reading this, *DO NOT* try the battery drain method I used unless you are prepared to set up a grid charger and cooling fan if your car becomes immobilised! Also, deep discharges should be done more slowly (I did 180 watts discharge but even that is probably too fast if a cell reversal had occurred).

More battery adventures to follow. I am picking up a used battery today which said 'replace HV battery now' and one block down about 1.2 volts from the rest. I will then have two spare batteries to play with. Overkill in a way but I have a few ideas (such as a HV battery for a mini RV, since the only monitoring needed is total voltage, and much better than lead-acid). Maybe also for a small solar setup. I also have a plan to install one in the boot of my Leaf, but that will be a real project - the idea would be to let the NiMH battery take most of the burst loads during acceleration and deceleration, given the propensity of Leaf batteries to wear out quickly - kind of like a supercapacitor. The working (total) voltages are different, but 1 and a bit Prius batteries could get the job done quite nicely!

 

PS - during the entire deep discharge cycle, the battery indicator on the touchscreen went to one bar at (maybe 30%?), but it stayed at the one bar for the entire discharge, even down to 7.5%!

 

I didn’t have the current in my display but it would have been over 10 amps!! It was the slowest the car could do it, using air con and headlights and fans to divert as much power away from charging. The battery didn’t heat up much as the car cabin was very cool due to air con, and I left the car running for 15 mins afterward to let the thermal management bring the battery temp back down.

I wouldn’t recommend doing a fast charge like this unless you can accept loss of the HV battery. Mine has been deemed a 'write-off' although it is still working fine (in my opinion!); so I am experimenting with it as I have two used spares anyway (one from a 90,000km car I got yesterday which I am tempted to fit!)

 

I do have a hobby charger (obtained in last few days) as well as a 'grid charger' (LED driver that has the correct voltage and low current);

With my original battery pack, it works fine (with a second-hand ECU with no corrosion) - even using the corroded connector on the original battery. The two 'stronger' module pairs are still close enough to the rest of the pack that they don't go far enough out of sync (voltage wise) during any normal use of the car (not even when running down to 8% using neutral), so I am tempted to leave them for now; the car kind of ignores them and bases the 'SoC' on the rest of the cells, so none of the cells go anywhere near overcharge, or cell reversal.

I'm going to remove the original battery pack today, and fit one of the two spare packs I have (one of the spares has been sitting for months but was apparently good - bought by someone who ended up having a failed inverter, so they didn't use the pack..... and the other is very good with one cell in one module failed, otherwise a 90,000km battery from a very well cared for car).

While it isn't possible to know the true state of either of my spare packs, I'm going to put the 'apparently good' pack into the car, fire it up with EngineLink going, and check SoC etc. If it is within the normal range I will just let the car charge it as it warms up (e.g. the slow charge it does when warming up the ICE). If it is low or out of balance I will just shut the car down (after taking freeze frame data) and put that battery aside.

My original battery probably has a lot of corrosion and electrolyte leaks inside, yet it still functions fine; my plan for this battery is to fully dismantle, examine, and once happy with it generally, do a cycle or two with the grid charger (and do a slow discharge with a DIY discharger), then recharge it to approx 50% SoC with the grid charger, refit it to the car, and use EngineLink to watch the voltages during a slow 'full' (80%) charge, and slow discharge down to about 25% true SoC - cheating but quicker than using my hobby charger. I also plan to repair the ECU just to prove it can be done (despite the destruction it did actually work still).

The end result is I have three batteries, one known faulty, one with the 'HV ecu connector fire', and one 'apparently good used battery'. I also have three battery ECUs, two perfectly good, and one not so good. I like the Gen 2 (and NWH11 Gen 1) Prius, and would love to get an NWH10 (they are now too rare here in NZ but when they appear they cost almost nothing due to battery failure). I have a working non-hybrid (gen 1 insight) for long journeys, and a Leaf for most of my daily usage, so I can muck around and generally recondition my Prius over time as a hobby

 

I am going to do a lot of research, experimenting, reverse engineering of the battery ECU, battery, and then more on my Gen II Prius over the coming months. I managed to make two threads covering different aspects of this problem, so I am going to put everything into a new thread as follows:

Landspeed's HV battery + ECU teardown, analysis, experiment, upgrade thread | PriusChat

I would be grateful if a moderator could lock this thread - thank you!