Eliminating bad cells from a pack—how to test?

F

FawltyPlay

1 µW
Joined
Nov 7, 2022
Messages
3
Hi all, rookie here.

I've got 10 sticks of 12 NiMH cells welded together in series, each pulled from a 2005 Honda Accord Hybrid. These cells I believe to be rated at 6500mAh. The fully assembled pack should have 120 cells in series to give 144v nominally. The battery for this car is being troublesome, refusing to spit out as much hybrid assist as it should and causing some monitors to not set (according to the dealership, anyway).

Under normal operation I understand the computers in the car attempt to keep the SoC of the pack between 20% and 80%.

I am in the process of reconditioning the sticks individually and have reached a stage where I'm ready to rebuild the pack, if I can decide which cells are worth keeping and which should be removed.

From reading, it seems prudent that I balance the cells as best I can and discard any outliers. I haven't found great specificity in what an outlier is, so I thought I would check with the experts.

Using a Cadex C7400ER I was able to do two sticks at once, and gave each stick several cycles until the measured discharge capacity stopped increasing for 2 cycles in a row. Details for this can be found under "Prime" on page 25 (41 in the pdf) here: https://www.cadex.com/_files/243/c7x00_user_manual-rev11.pdf.

The results of this were as follows, for each stick as a percent of nominal 6500mAh discharge capacity:

  1. 91% -> 92% -> 92%
  2. 95% -> 93%
  3. 95% -> 94%
  4. 94% -> 94%
  5. 84% -> 92% -> 92%
  6. 21% -> 92% -> 92%
  7. 30% -> 94% -> 93%
  8. 12% -> 92% -> 91%
  9. 93% -> 91% -> 91%
  10. 21% -> 93% -> 92%

Obviously some of those are alarmingly out of line with each other, but I seem to have brought them back in to being similar capacities. A suggested capacity deviation of 2% I found on these forums would mean a maximum difference of 130mAh. The largest between any two sticks is 260mAh (sticks 3/4 compared to sticks 8/9). Elsewhere I've seen tolerances of 5 or even 10 percent. Any alarms ringing?

One thing that was suggested to me was to allow them to self-discharge to see if they do so evenly when the car is not in use, especially given the high self-discharge present in the first 24h or so for NiMH cells. I am currently on day 4, with these results coming in (same stick order, of course):

Code: 
Prime Voltage | 24h (drop)     | 48h (drop)     | 72h (drop)     | 96h (drop)     | 144h (drop)    | 168h (drop)
1.  17.90v    | 16.28v (-1.62v)| 16.02v (-0.26v)| 15.85v (-0.17v)| 15.77v (-0.08v)| 15.63v (-0.07v)| 15.60v (-0.03v)
2.  17.72v    | 16.29v (-1.43v)| 16.04v (-0.25v)| 15.86v (-0.18v)| 15.77v (-0.09v)| 15.63v (-0.07v)| 15.61v (-0.02v)
3.  17.36v    | 16.27v (-1.09v)| 16.04v (-0.23v)| 15.85v (-0.19v)| 15.75v (-0.10v)| 15.56v (-0.09v)| 15.54v (-0.02v)
4.  17.38v    | 16.20v (-1.18v)| 15.95v (-0.25v)| 15.77v (-0.18v)| 15.66v (-0.11v)| 15.48v (-0.08v)| 15.45v (-0.03v)
5.  17.49v    | 16.24v (-1.25v)| 16.00v (-0.24v)| 15.82v (-0.18v)| 15.72v (-0.10v)| 15.59v (-0.06v)| 15.57v (-0.02v)
6.  17.70v    | 16.18v (-1.52v)| 15.93v (-0.25v)| 15.77v (-0.16v)| 15.65v (-0.12v)| 15.51v (-0.06v)| 15.48v (-0.03v)
7.  17.73v    | 16.18v (-1.55v)| 15.93v (-0.25v)| 15.76v (-0.17v)| 15.67v (-0.09v)| 15.55v (-0.05v)| 15.53v (-0.02v)
8.  17.74v    | 16.26v (-1.22v)| 16.00v (-0.26v)| 15.82v (-0.18v)| 15.72v (-0.10v)| 15.58v (-0.06v)| 15.56v (-0.02v)
9.  17.30v    | 16.17v (-1.13v)| 15.90v (-0.27v)| 15.73v (-0.17v)| 15.66v (-0.07v)| 15.54v (-0.06v)| 15.52v (-0.02v)
10. 17.44v    | 16.23v (-1.21v)| 15.94v (-0.29v)| 15.77v (-0.17v)| 15.69v (-0.08v)| 15.57v (-0.06v)| 15.55v (-0.02v)

Things seem *fairly* even here, but again, not totally sure what to look for. Stick 1 has aberrantly high primed voltage, while sticks 3, 4, and 9 have lower primed voltage. But after only 24h the cells are all within roughly 0.1v of each other. This feels like a good sign but I'm far from an expert.

Are there other tests I can run? This Cadex unit is pretty awesome and it doesn't take me long at all to charge or discharge these sticks so I'm game for most things.

As an aside, would it be useful for this forum if I documented the (dis)assembly process for this hybrid vehicle battery? The HAH in particular has scant information online.
 
Generally when you have cells that aren't all remaining at the same state of charge under normal usage, they have changed in characteristics from the rest of the cells. So not only their capacity but their internal resistances (which affects voltage sag under load and ability to deliver current, which affects the entire pack's performance directly since they are all wired in series in this case) are different than they should be.

So...the "best" repair method is to eliminate any cells that were significantly different from the others, meaning the four that were at very low capacity especially. Optimally you'd also replace the ones that were at 84% too, but they havent' changed in characteristics nearly as much as the other four.

For now, you could just run the pack as it is, now that the low sticks have been reconditioned enough to (at least for the moment) recover most capacity, then when problems resume, just replace those sticks that show lower SoC or capacity.

Where to draw the cutoff line is up to you and your budget. ;)



Some notes:

Each of those sticks is made of a number of individual cells, and it's likely that only one cell in each underperforming stick is really causing most of it's problems. Your tester may be able to test individual cells as well as entire sticks, if you can open the sticks to get access to individual cell terminals inside (assuming there is no such access normally).

So if you get some new sticks to replace the worst ones, you could use one of the replaced sticks as parts to refurbish the others, assuming you have or can get the tools to reconnect the cells in the same method. I have not seen how those specific sticks are built inside, so I don't know the specific interconnect method, but suspect they're spotwelded to nickel (or other) metal strips, which are folded between each cell. If you disassemble the sticks to remove individual problem cells, you can just cut the strips between the cells, so that you have more options for reconnecting to replacement cells (which can even include soldering the strips together if you can't re-spot-weld them together).
 
Hey, thanks for the reply.

Part of my problem is that it's difficult to acquire sticks that are in a known good condition. Most people who are selling them have simply removed them from a hybrid that has been turned into parts, and the batteries are very likely to have gone.

I can get new cells guaranteed to have capacities at least 5500mAh for 40$ each from a reputable seller. If I do this and all those cells come back as being 5500mAh rated, would putting them in a pack with sticks that test out at 6000mAh be a recipe for future problems? Is that any better than just using the current sticks? I would expect problems, as the computer does not monitor each stick, only the whole pack, and the lower capacity sticks could end up being overdischarged.

On the point about separating the individual cells; I'd love to, honestly. But I just don't have the equipment, and the sticks seem to be welded in many places while having fairly tight tolerances. I could make a jig, maybe, but I'm not sure I'm willing to go that far for this car.
 
One thing I've discovered from opening up bad batteries and individually testing the p-groups: often the lowest voltage p-group has the lowest internal resistance and is actually the healthiest. Only low voltage because it was compensating for the p-groups with older cells, I suppose, and didn't sag as much under load.

So if you didn't charge everything to full, then got 30% for one discharge test, 94% for the next, and 94% again - that cell likely wasn't unhealthy compared to one that did 90%, 90%, 90%. Much the contrary, it was likely low because it is still one of the healthiest with the lowest resistance and able to contribute the most current with least sag.
 
FawltyPlay said:
I've got 10 sticks of 12 NiMH cells welded together in series, each pulled from a 2005 Honda Accord Hybrid.

I am in the process of reconditioning the sticks individually and have reached a stage where I'm ready to rebuild the pack, if I can decide which cells are worth keeping and which should be removed.

From reading, it seems prudent that I balance the cells as best I can and discard any outliers. I haven't found great specificity in what an outlier is, so I thought I would check with the experts.
Click to expand...
When you checked with the "experts" at Honda what was their advice as far as attempting to "recondition" any of the 17yr old sticks or if it's more likely none of the NiMH sticks/cells are regenerable for any sustainable length of mileage ?

Over 50% of these 17 yr old NiMH cells are outliers, IMO. Actually may be closer to 100%. Have you talked to a Honda dealership to see if they know of anyone that has been able to recondition 17yr old NiMH stick cells?

You've probably already checked out some of the options ...
https://www.okacc.com/product/brand-new-ni-mh-6500mah-144v-hybrid-car-battery-pack-replacement-for-honda-civic-1st-gen-2003-2005/
https://www.alibaba.com/product-detail/Replacement-Nimh-Hybrid-Battery-Stick-for_62078134007.html
https://www.ebay.com/itm/313912311534

https://venicehybridtech.com/battery-sales/ ...
"The Accord Hybrid is similar to Honda Insight and HCH but, like the Gen II Honda Civic, it uses special sticks which are also in pairs. These are a little longer and thicker than the other two cars, and hence these are not inter-changeable. These sticks are also hard to find.

The Grid charger has been very effective however. There are many people who face the IMA light and CEL light problems, and face the dilemma of whether they should sell the car as they cannot afford to buy re-conditioned battery packs.

Fortunately for them, our Grid Chargers are very effective in these cars, thus saving them the trouble of getting rid of their cars at a throwaway prices. In fact we have seen a lot of people buy these cars cheaply and come to us to repair them, and get them back on the road."

Hope this helps and you are able to get your Honda Accord back on the road.
 
When you checked with the "experts" at Honda what was their advice as far as attempting to "recondition" any of the 17yr old sticks or if it's more likely none of the NiMH sticks/cells are regenerable for any sustainable length of mileage ?
Click to expand...

When I wrote "experts" I was actually referring to this forum :) the folks at Honda have been a total waste of time and money to try to speak to.

But no, not specifically. I've read stories of people doing it with success, but its far from an exact science, and I imagine most don't try or give up relatively quickly.

A lot of my information comes from reconditioning of Prius or Civic batteries, actually, and in general people seem successful. Can't be sure on the 17yr old part, but I've definitely seen videos as recent as last year.

Over 50% of these 17 yr old NiMH cells are outliers, IMO.
Click to expand...

What makes you say that?

https://venicehybridtech.com/battery-sales/ ...
Click to expand...

Ah, this guy has a youtube channel that was the start of my ambitions to get this working. Helpful, but ultimately the cost of a charger is only 400-600 less than a new pack that actually comes with a warranty for 2 years or 100k miles. That same pack costs more than the car is worth, though so it's a bit sticky.

The sad part about this is the car runs perfectly fine, the smog check monitors just won't set :/

I've updated the self-discharge table with values from the last 3 days.

Stick 4 is standing out a bit, but the daily voltage change across all the cells seems pretty same-y. If there's no new advice I'll probably stick these cells into the car and give it a spin.
 
You must log in or register to reply here.
Back
Top