[SteyrTMP]

When I bought my 2000 G1 MT, I was told the IMA battery was only ~32k miles, but about 7 years old. I got it to work, once, for a few minutes, but ended up tearing it apart with my dad, and after he tested the sticks, most of them were beyond bad, several of them leaking while being cycled. I ordered 21 sticks from HybridRevolt, and my dad had one unused OEM stick he'd had from his G1, and we tested those, and put the best 20 into the battery. Reinstalled battery, reset codes with my ScanGaugeII, and it worked great for just under two weeks. Battery was starting about half-way, occasionally running low and no assist, but would charge back up to a random amount, depending on the hills and such. At the end of two weeks, it dropped to empty and the IMA light went on. Initially, my CEL was 1447, but eventually 1449 came on. I jumped the OBDII, and it flashed the dreaded 78 code.

We pulled the batteries out, measured individual sticks, and took the two lowest ones out, and replaced them with the two spare sticks that measured higher than them. I was seeing a difference from 7.5V to 7.9V individually, but for not being grid charged, they didn't seem drastically different, i.e. one of them at 0.1V or anything.

Fast forward a few weeks. My dad just finished building one of Ol'Rowdy's grid chargers. We took the battery out, charged it in his workshop, up to 171.2V, but at that point I needed to head home, and it did not have time to either peak, nor to level out the sticks. However, it did have assist and read one tick from the top on the battery gauge. That lasted maybe 5 minutes. Assist and charge were doing their jobs, and after maybe 5 minutes, it threw a CEL and IMA light. I stopped, turned it off, tried to reset the codes, but the 1449 came right back on. I drove the rest of the way home like that.

When I arrived home, I plugged the charger in. It read 151V, which seems a lot for 5 minutes of assist/charging. Albeit, the overall drive was ~45 minutes, so I don't know if the DC-DC convertor pulls any power from the IMA or not. Either way, I looked up charging on here, BumbleBee, and Hybrid, and felt comfortable to leave it charging overnight (It was midnight at this point). That should have given it ~2-3 hours to peak, and another 5-6 hours to top off the stragglers. At 8:30, I got up, unplugged it, and it was charged at 181.2. However, when I reset the codes and turned it on, while I watched the battery gauge climb up to one point from the top, the 1449 and IMA lights never left. I tried again, still nothing.
This afternoon, I bought a new multimeter, and pulled the BCM cable, and checked the stick pairs. Following the order C9/C20, C20/C7, C7/C18, C18/C5, C16/C15, C15/C14, C14/C13, C13/C12, C12/C11, C11/C10, my results were: 16.95, 16.90, 16.91, 17.00, 16.96, 16.91, 16.94, 17.05, 16.90, 16.92. While there is some variation, this is within the 0.1V mentioned on previous posts.

I am running out of ideas. I'd like to do a reconditioning, but I'm not sure that's going to fix the problem. I wanted to give it a complete charge, run it until it stopped working, recondition, then run it until it stopped working again, and grid charge it a few days less than that result. Obviously, that won't be happening with the current condition.

Any ideas? I've probably forgotten something we've done, or a crucial point, but I can clarify as necessary.

 

[eq1]

Here are some comments on random things from your post:

...Reinstalled battery, reset codes with my ScanGaugeII, and it worked great for just under two weeks... At the end of two weeks, it dropped to empty and the IMA light went on. Initially, my CEL was 1447, but eventually 1449 came on. I jumped the OBDII, and it flashed the dreaded 78 code.

We pulled the batteries out, measured individual sticks, and took the two lowest ones out, and replaced them with the two spare sticks that measured higher than them. I was seeing a difference from 7.5V to 7.9V individually, but for not being grid charged, they didn't seem drastically different, i.e. one of them at 0.1V or anything.

0.4V difference is huge. Up to this point in your story it just seems like you have at least one very fast self discharge cell.

In general you can't tell much from resting voltages, you need to measure loaded, and V change...

...and it was charged at 181.2. However, when I reset the codes and turned it on, while I watched the battery gauge climb up to one point from the top, the 1449 and IMA lights never left. I tried again, still nothing.
This afternoon, I bought a new multimeter, and pulled the BCM cable, and checked the stick pairs. Following the order C9/C20, C20/C7, C7/C18, C18/C5, C16/C15, C15/C14, C14/C13, C13/C12, C12/C11, C11/C10, my results were: 19.95, 16.90, 16.91, 17.00, 16.96, 16.91, 16.94, 17.05, 16.90, 16.92. While there is some variation, this is within the 0.1V mentioned on previous posts.

I take it that 19.95 value is just a typo... In the past I've sometimes had a hard time clearing the P1449-78, even though there wasn't anything wrong with the pack. Are you pulling the underdash #18 fuse, and/or the 12V neg cable to clear? There's nothing about these tap voltages that jumps out...

 

[SteyrTMP]

I am using my code reader to reset, and its worked many times. I'll see the code go away, but then come right back. I'll have to pull the fuse and see if it changes. Yeah, I'll fix the typo, only a 0.1V variance. I'll have to look into how to figure out which one is fast discharging.

 

[hurricos]

You need to test the voltages of your sticks under substantial load. The deviations will only become apparent once you've done that; the cells that are effectively dead -- that go to 0V when 50A pass through them -- will often stand right back up on their own in a few minutes, and will almost universally jump back to >1.2V after being trickle-charged by the IMA system, but not before you get P1447 (74) due to single cell deviation at idle, or P1449 (78) due to groups of ailing cells all prematurely indicating the "pack bottom" and "pack top" to the BCM.

The IMA system cares about those voltage drops under assist / jumps under charge much more than it cares about resting voltages; basically, you have to reproduce (at least some of) the conditions the IMA sees in order to make it happy.

Here's a summary of the process I did -- it links back to a bunch of other posts I made about it: Too good to be true?

In particular, I'd suggest you try to construct a load as similar as possible to mine; that way, you can get a gauge of how strong your batteries actually are, or whether under similar current ratings, your batteries plummet instead.

 

[eq1]

^ You don't need to test sticks at a big load. Start by discharging all cells, completely, at a low current. And then test for fast self discharge cells. If these two things are taken care of, you probably won't have to do anything else. After these, you can test more, big current, small current, whatever, but you'd probably be better off just putting the pack back in the car. Unless you want to play with batteries all day.

 

[hurricos]

@eq1: you said it yourself that the downside to looking for excessive self-discharge is that you have to wait for a long time.

But I think high-current testing has been overlooked. Cells that have higher internal resistance (a good sign of degradation) may maintain >1V for 20 minutes @ 2.5A but go flat when drawn for 5 minutes @ 10A because those cells are losing more power to heat.

In my view, I don't think it makes sense to measure cells by some proxy measurement like self-discharge; I don't know about any banks of evidence which link self-discharge to degradation. Degraded cells are going to be mistreated by the MCM when the car is running and that's all the BCM cares about when checking for cell failure or measuring the top and bottom of the pack.

 

[eq1]

There's nothing "proxy" about testing cells for fast self discharge - you're trying to find cells with fast self discharge, because, as far as I can tell, that's what makes packs totally unusable (unless you do manual stuff that is, forgo the BCM's management).

High 'internal resistance' is usually a result of the cells not being discharged completely in the car. It's not really internal resistance, but it will look similar - voltage sag and spike. Hard to say exactly what the electro-chemical process is that causes this, even battery researchers can't quite figure it out. But it's akin to 'memory effect', might have to do with the cells' vario-stoichiometry, and/or might have to do with the development of large crystals which don't support high current. The Panasonic NiMH manual says the cells need to be discharged completely, to 1V, or they'll develop low voltage and 'might' develop limited capacity. My own experience is consistent. But, the way the BCM does its thing, these cells probably never get discharged remotely close to that. If/when they do, eventually, it's probably only a cell or two, not all of them. And then that cell or two develops a different voltage profile from the rest and the pack becomes forever stunted...

I don't know. High load testing - ain't gonna hurt anything. One can't simply look at resting voltages from stick to stick and expect to be able to build a good pack with that 'data'. And if you're discharging sticks, and unfamiliar with the voltage behavior/targets, you could easily mistake what looks like 'good' voltages or good behavior in general if you just, say, discharge them at 2.5 amps - a high load will make the bad sticks (most likely sticks with a 'bad' cell or two) pop out, that's for sure...

You do need to discharge all cells though, completely, preferably at a very low current, otherwise, you end up testing the performance of the sticks/cells in a degraded yet mostly completely reversible condition. You need to 'reset' them, 'normalize' them. Then you can do whatever tests. Personally, I would just do/have done a few things, no high load, and put the pack together and test in the car. My high load testing is done with/in the car, I guess. I'm always testing at high load, every drive!

 

[eq1]

... I'll have to look into how to figure out which one is fast discharging.

If you get your P1449 issue squared away, and get the pack more or less working, yet it still fails after so many days or weeks, then here's a method to identify which stick-pair has the cell with the fast self discharge:

-drive and use assist until the BATT gauge is almost empty. Preferably you'll have some sense when the pack is going to 'neg recal' (reach effectively empty) and you'll stop draining the pack before that point.
-pull into garage with headlights ON, let car go into auto stop.
-you can turn headlights OFF (you usually need headlights ON for the 12V drain to be sourced from IMA pack going into auto-stop)
-measure tap voltages (time 1)
-measure again some minutes later and/or when you see the BATT gauge plummet/go neg recal (time 2)
-subtract values at time 2 from time 1.

The tap with the largest V drop has the cell that's going empty, and it's probably a fast self discharge cell.

 

[SteyrTMP]

On Monday, I pulled fuse #18 for an hour while I ran errands, came home, and let it charge for 12 hours. I had only had it balancing for maybe 6 hours sunday, and apparently, that was not enough. I woke up, unplugged, and drove to work with full battery. Today (Wednesday), I'm down to about 3 or 4 bars. I wonder if I should have gone for a full 24 hours of charging, but I did not have 24 hours to charge. Either way, I'll do as eq1 suggested, and I'll see what my voltages are. The car is warming up right now in the driveway, with 4 bars remaining. I'm not sure how happy it's going to be with my removing the BCM connector to read the tap voltages. I dunno if sitting is going to kill 4 bars in any reasonable time, but I'll find out. As for headlights, being a Canadian car, I'll see how it behaves with DRL.

I would like to have a small LCD with constant tap voltages read out. It would make this so much easier.

 

[SteyrTMP]

Well, that didn't work. I tested voltages with the car at 4 bars, auto start. 16.22, 16.14, 16.21, 16.25, 16.24, 16.16, 16.10, 16.33, 16.13, 16.09. I plugged the connector back in, but the car was not very happy. I reset codes, and drove around again, as it was down to 2 bars, and wouldn't auto-stop, so I let it charge up to 4 bars, but when I went to stop and auto-stop, it threw a code 1449 instead. I pulled into the driveway, and was unable to remove the code. Because I had to drive again, and couldn't use auto-stop, it charged, not discharged. I had 16.91, 16.73, 16.87, 16.87, 16.81, 16.73, 16.66, 16.94, 16.67, 16.65.

I'm not sure what negative recalibration is. It's when it goes from full to empty, one click at a time, no? I have the opposite. It'll go from empty to full, with the IMA light on, and no assist/recharge.

At this point, I suspect charging it again would not be benificial. I'll leave it the way it is with the IMA for a few more days, and will be able to get to my dad's next week. Hopefully, we'll have time to add the discharge feature to the charger, and will try the high load testing when the battery out of the car, and will share results. Would it be better for me to charge it and then to high load testing, or test it from it's current state?

 

[eq1]

...I'm not sure how happy it's going to be with my removing the BCM connector to read the tap voltages... As for headlights, being a Canadian car, I'll see how it behaves with DRL.

You leave the connector in place when reading voltages, easier to back-probe the terminals with it connected. Removing it though doesn't mess anything up, as long as the car isn't turned on when you do it. I have a Canadian with DRL as well, that's good, your autostop discharge current should be close to what mine is - usually about 1 amp when the 12V battery is charged near max, no other accessories on.

 

[eq1]

Well, that didn't work. I tested voltages with the car at 4 bars, auto start. 16.22, 16.14, 16.21, 16.25, 16.24, 16.16, 16.10, 16.33, 16.13, 16.09. I plugged the connector back in, but the car was not very happy...

I see. No, you need to leave the connector in place when you measure tap voltages in auto-stop...

I'm not sure what negative recalibration is. It's when it goes from full to empty, one click at a time, no? I have the opposite. It'll go from empty to full, with the IMA light on, and no assist/recharge.

"Neg recal" - I probably shouldn't use that term, it's just habit at this point. But, it simply means your pack is effectively empty. The BATT gauge can be anywhere from 3 to 20 bars and it will drop to 1 bar, bar by bar - that's a neg recal. It often looks like it drops to 3 bars though, because it starts charging pretty fast and repopulates the 2nd bar quickly, you often don't notice it.

The idea with "negative recalibration" is that the BATT gauge is somewhere high and then drops to the bottom, then the IMA charges, and once it finds 'full' it will do the opposite - a "positive recalibration", moving from low-ish bars to the top (or 19 bars rather)... If everything's fine, you don't see big BATT gauge movement - you can start from full bars and go all the way down, and you can start all the way down and go all the way to the top, incrementally, no 'recalibration' needed. The recalibration aspect is all about the big jumps, where the BCM seems to find an empty pack unexpectedly, and then it charges to see how much it needs to do to get it back to full. If it's not much there's some kind of 'recalibration' that happens. If it's really not much you get the P1449-78...

 

[hurricos]

Hey @SteyrTMP,

You won't be able to clear IMA DTCs using a scantool; instead, reset your IMA (disconnect / reconnect the 12V battery).

Hopefully, we'll have time to add the discharge feature to the charger, and will try the high load testing when the battery out of the car, and will share results. Would it be better for me to charge it and then to high load testing, or test it from it's current state?

Grid-charge for as short or as long as you want, let them sit for at least a few minutes to settle, and then load-test. The sticks don't have to have identical starting voltages; they'll be close enough if they've been in the same pack and poked by the IMA system. (On the other hand, after replacing sticks, you should discharge and then charge a bit; think what you do to align a just-shuffled deck of cards).

During the load test, you're looking for sticks that have see particularly fast voltage drops, not necessarily sticks that end at the lowest voltages. In particular, you'll probably find sticks where some constituent cells run out of charge earlier than the others.

Under this kind of load, a cell that runs out of charge will drop from 1V to near 0V relatively suddenly. This appears a few times in my measurements:

2001 CVT IMA Potentials

Load, not-load, pair+tap test Load and measuring order,06/09, S/R,12/15, Q/P,18/19, O/N,16/13, M/L,10/07, K/J,04/01, I/H,02/05, G/F,08/11, E/D,14/17, C/B,1, H,3, T,4, I,20, A 6,18,3,9 Off, forward,9.84,12.6,13.58,12.36,13.28,13.92,13.72,12.92 Off, backward,10.24,13.24,13.65,12.42,13.35,13.93,13....

docs.google.com

2001 CVT IMA Potentials

Load, not-load, pair+tap test Load and measuring order,06/09, S/R,12/15, Q/P,18/19, O/N,16/13, M/L,10/07, K/J,04/01, I/H,02/05, G/F,08/11, E/D,14/17, C/B,1, H,3, T,4, I,20, A 6,18,3,9 Off, forward,9.84,12.6,13.58,12.36,13.28,13.92,13.72,12.92 Off, backward,10.24,13.24,13.65,12.42,13.35,13.93,13....

docs.google.com

... occasionally with two cells in the same stick:

2001 CVT IMA Potentials

Load, not-load, pair+tap test Load and measuring order,06/09, S/R,12/15, Q/P,18/19, O/N,16/13, M/L,10/07, K/J,04/01, I/H,02/05, G/F,08/11, E/D,14/17, C/B,1, H,3, T,4, I,20, A 6,18,3,9 Off, forward,9.84,12.6,13.58,12.36,13.28,13.92,13.72,12.92 Off, backward,10.24,13.24,13.65,12.42,13.35,13.93,13....

docs.google.com

2001 CVT IMA Potentials

Load, not-load, pair+tap test Load and measuring order,06/09, S/R,12/15, Q/P,18/19, O/N,16/13, M/L,10/07, K/J,04/01, I/H,02/05, G/F,08/11, E/D,14/17, C/B,1, H,3, T,4, I,20, A 6,18,3,9 Off, forward,9.84,12.6,13.58,12.36,13.28,13.92,13.72,12.92 Off, backward,10.24,13.24,13.65,12.42,13.35,13.93,13....

docs.google.com

I think that under much higher loads -- closer to the 80A the car is designed to move -- this drop below 1V happens to a lot of cells, and the IMA verifies these slightly weaker cells aren't actually dead by requesting 3 bars of charge momentarily to avoid a misleading P1449 (74). That code would occur due to a >1.2V difference between any two sticks in low current conditions (potentially following high-current conditions) for more than 25.4s -- see page 180 of the DTC manual for a description: http://gershon.ucoz.com/HONDA/HONDA_DTC.pdf

In general, when certain cells stay low in a pack which is otherwise relatively well-charged, it causes trouble. While there's documentation that specifically describes how that would cause P1449 (74), we don't know for sure whether the same is true for P1449 (78). Nobody has dumped the BCM's firmware yet, but since drawing from empty cells or charging full ones damages the pack, I'm guessing the BCM watches for both the weakest taps and the strongest taps for determining the overall pack's top and bottom charge.

Edit: I had added:

I want to see whether you pull blink code (74) as well as (78). Before reading the rest of the explanation, please take a brief look around page 180 onwards of the Honda DTC manual, which explains what causes the system to set these P1449 'battery degradation' DTCs: http://gershon.ucoz.com/HONDA/HONDA_DTC.pdf

... but I forgot that it's winter, and P1449 (74) requires the battery module be >= 25*C to trigger. Oops.

I had also added:

Since P1449 (78) can only be detected when the IMA detects the bottom AND top of the pack during a single drive cycle after a reset, try discharging it a bit by getting it into autostop and leaving the lights on. After that, reset your IMA; then, you can recharge your battery the Honda-approved way by applying your parking brake, putting it in neutral, and revving above 4K RPM. Since you're starting from low charge, the IMA will be forced to track battery charge upwards.

Once you have a working battery, this use-the-car-to-rise-from-low-charge is the same procedure you'll need in order to get the BCM to agree the battery has a large capacity (even after replacing sticks, my IMA light would come back on immediately after finding the top the pack post-IMA-reset).

... but this is mostly irrelevant; you're not going to be getting P1449 (74) because, again, it's winter.

 

[hurricos]

One more comment: The OlRowdy guide has a lot to say about the dangers of high voltage. They're all true. I have a slightly different philosophy about how to best deal with them while doing doing stick-level work.

- Exposure is important, but is a product of both time and space; 10 minutes with the right tools can be as bad as 1 minute with the wrong ones.

- Flipping the IMA switch off halves the peak potential and the number of pairs of points on the pack across which a voltage potential is present -- it's another major exposure factor, don't forget it.

- Leather gloves are better than nothing, esp. when working with a pack that is "off".

- Trickle charging indefinitely is safe because of NiMH's inherent behavior (big fan of Globtek's notes: Nickel Metal Hydride Battery Safety Notes). Go wild (but stay <500mA).

- Discharging has no such internal protection. I believe it is best to leave discharging to manual intervention to avoid damaging working cells.

- The car runs on high currents; high current testing shows problems that the car also cares about; high currents deplete the IMA battery quickly (<<<1h), making manual intervention easy.

- Expect to have the IMA out of the car for a while. You can, of course, continue to drive the car; you only need to reconnect the DC bus on the IPU to the DC-DC converter in order to continue charging your 12V. You will need to pull the fuse over from the junction board of the IMA Battery module -- I can't find a thread on IC, but this page shows you the resulting wire connecting the IPU with the DC/DC: https://99mpg.com/mikestips/makingasimpleimaby/

 

[eq1]

...I think that under much higher loads -- closer to the 80A the car is designed to move -- this drop below 1V happens to a lot of cells...

fyi: I think there's at least two major forms of low voltage 'protection'. It seems like the MDM (the 'left side stuff') must have its own low voltage mitigation, or maybe the MCM sends data to the ECM which then lowers the commanded power. I don't think its the latter, I vaguely recall watching that and commanded power can stay the same under low voltage). But to me it looks like - we'll call it the 'MDM', not sure what component it actually is - the MDM will relatively quickly reduce power demand to keep voltage above 120V. This looks like it's much quicker than the BCM's 'tap watching' can deal with. After the MDM reduces power demand, voltage should jump back to at least 120V, and then the power commanded is only really high, above ~6.5kW, for 4 seconds, so normally voltage will bounce back higher after the 4 seconds, when power commanded is even lower, max of ~6.5kW...

Point is, cells can drop below 1V during the 4 seconds of max max assist, but it should be very short lived - maybe a second. When MDM reduces power, trying to maintain at least 120V, cells should come up to at least 1V. And then when power is reduced again, to ~6.5kW or lower, cells should really be bouncing back up. And if they don't, 'something' will reduce power demand again, probably the 'MDM'.

The BCM will be watching tap voltages for true empty cells, or cells that look empty perhaps do to degradation. As far as I can tell, after watching this stuff for a few years, taking measurements and so on, it looks like the BCM measures/calculates voltage drop at tap level, and/or maybe it's looking at moving averages across taps. A 'tap' is empty when its voltage drops relatively quickly under load and doesn't rebound quickly. Not sure how it's actually implemented by the BCM. In practice I'll watch tap voltages near empty, during auto-stop ~1 amp load, and the tap with an empty cell usually drops about 0.3V more than others - when the BCM says "empty/neg recal." Knowing that normal cell voltage is about 1.2V and that they're empty at about 1V, a quick 0.3V drop is about what you'd expect.

When I'm driving, I'll see the "assist limit parameter" on the OBDIIC&C trigger a few percentage points before neg recal; I think that too is probably tap voltage slope detection/measurement/calculation. After 'Alf' triggers, assist will be throttled back, and if I keep assisting, at these lower currents, a given tap's voltage will start to tank again - and then the empty/neg recal will happen...

So, many things are happening that are trying to prevent discharging with empty/low voltage cells. I guess I'm suggesting that a lot of cells probably don't drop below 1V, or when they do it's short-lived. Or perhaps put another way: a lot of cells probably do drop below 1V, during full-on full assist, but the car will throttle. I think the main reason this happens is because cells aren't discharged fully in the car, they need to be discharged. Over time and usage, and even in the short term, voltage performance drops, and it's usually reversible. You don't want to be testing sticks/cells if they haven't been fully discharged, because you're not really uncovering the truly 'bad' cells, you're just seeing the impact of this 'cruddy' phenomenon...

In general, when certain cells stay low in a pack which is otherwise relatively well-charged, it causes trouble. While there's documentation that specifically describes how that would cause P1449 (74), we don't know for sure whether the same is true for P1449 (78)...

I don't think it's likely/common/possible for cells to stay low in a pack and still have it seemingly somewhat functional in the car. You'll have some code or dysfunctionality if there truly is a low cell or more. Really, it's pretty basic: if the cells can stay within a certain voltage range within a 'sliding scale' of power demands, and a really quite narrow capacity window (~10% of 6500mAh), the pack will continue to 'work'. If not, it will be considered a failure. The P1449-78 is the car's way of calling the pack a failure - it can't charge at least 10% within the voltage range its supposed to operate within. That's pretty much it. My hunch is that for whatever reason, at least a single cell has dropped to the bottom, probably mostly fast self discharge, it's voltage behavior becomes inflated, and it ends up triggering both 'empty' and 'full'. It can't charge high enough without making the tap it belongs to trigger the 'high voltage' limit; and then it's empty first, again and again. Meanwhile, all the other cells stay high, they're never discharged low, they develop slumpy voltages, etc etc...

Overall, the car hides failing packs by throttling power demands, constantly lowering the bar, moving the goals posts. And then the only thing left, the thing that really matters, is capacity. You need that miniscule (10%) amount of capacity in order to maintain a semblance of a functioning pack. But the pack has been long gone way before that point, long gone in the sense that it has been underperforming for a long time, but not in the sense that it's toast. It seems like most of the common degradation is easily reversible, simply with a full, very low current discharge on every cell. You can get degrees of better functionality with the partial methods, like grid charge and full pack discharges. Grid charge is...missing the point. Full pack discharge - hard to get every cell discharged that way, there's always some that hang on a long time and never get fully discharged, and you can end up deeply reversing other cells and that makes the whole operation fail... Ideally we'd each have a cell-level discharge rig for full packs. But we know that's not going to happen.

 

[hurricos]

I've just never seen anyone take a batch of random 'dead sticks' and restore capacity in them through power cycling. I'd like to see an A/B test against Martin Kennedy / IMA Battery Assisted Profiler done, once before and once after this process is done, to see the capacity or at least the voltage curve changes.

 

[eq1]

^ I don't know what "power" cycling is, but just cycling sticks is mostly a waste of time. It's the first method I did when I came around here 10 years ago. And ever since, everything I've done has been an improvement over that. To be honest, I think I'm beyond the 'nit-picky gather all this data and test all this stuff' stage (not to mention long tired of it). There's a couple major flaws in the way the BCM manages packs vis-a-vis the way Insight NiMH work. It's not too difficult to work around them. I've been using a code-throwing original 2002 pack for the last 5+ years, of course, it doesn't throw codes any more. It's the culmination of all my 'battery' work around here. It puts out full power. Etc. So, in my mind, I don't need to do anything other than what I've been doing (knock on wood).

I do have a lot of arcane questions that still linger, though. There's things that don't make sense, that I don't understand. But just getting a pack to work is child's play to me... The car doesn't require a lot, and the cells seem to be able to perform way better than you'd think they can, even after they've seemingly become "toast" in the car. Just discharge all cells completely, at low current, charge it up, put it back in the car. After that, if you still have problems, then you'll have to look for truly failed cells. But you still won't need to do extensive testing, high power discharges and the like. You could say discharge a stick at 6 amps and take cell voltage measurements, the failed cell/s won't hold the right voltage/s. You could charge-up a stick and let it sit for two weeks, test for self discharge. There isn't much more to it than that...

hmm, OK, maybe I'm missing something critical: the BCM. You have to know how the BCM does its thing and avoid the problems. To me, the problem isn't the cells, it's the management. Once you know what to look for - in terms of management problems - you can avoid them. Basically, it doesn't make sense to subject perfectly good, working cells to a management that will destroy them. Or try to make your cells just so, so they'll play nice with stupid management. Forget it. Waste of time. Waste of good batteries. If you know what to avoid, then you can use practically any cells.

 

[eq1]

...I'd like to see an A/B test against Martin Kennedy / IMA Battery Assisted Profiler done, once before and once after this process is done, to see the capacity or at least the voltage curve changes.

Not positive what exactly you're most interested in, but as far as A/B tests go, or being able to gauge pack performance goes, here's a thread I made a couple years ago that shows graphs of pack output et al, under a number of assist events, using OBDIIC&C. Do you have one of those? Seems like this kind of logging would be good enough to do comparisons:

OBDIIC&C data logging: What is 'good' pack...

I decided to drag out my old laptop and do some data logging with the OBDIIC&C. In general I've been trying to hone-in on some differences in behavior/performance depending on how you use the pack, here I figured it'd be worth posting some graphs and describing them. There's a lot that can be...

www.insightcentral.net

 

[hurricos]

@eq1:

I don't know what "power" cycling is, but just cycling sticks is mostly a waste of time

I was trying to describe your deep discharge method. I obviously didn't do a great job describing it.

To be honest, I think I'm beyond the 'nit-picky gather all this data and test all this stuff' stage (not to mention long tired of it)

OK.

You have to know how the BCM does its thing and avoid the problems.

I am getting multiple BCMs and an MCM shipped to me in order to attempt to dump the firmware (h8/500 machine code) from them to then disassemble. The MCUs used in each are rewritable; see BCM Hacking. Convert 112 pin QFP Hitachi H8 CPU to.... If I succeed, we'll have a precise idea of how the management makes its decisions.

 

[eq1]

I was trying to describe your deep discharge method. I obviously didn't do a great job describing it.

I don't know. Deep discharging isn't even really a method. Not even sure 'deep' is totally necessary.. Full discharge seems necessary, like every cell to 1V under low current. But beyond that it gets murky... My main schtick here is/has been to try to impress upon people that a lot/most of the dysfunction in packs is due to stuff that's easily reversible, in the first instance. I spent a few years banging my head on all the walls until I realized this. A lot of wasted time. I guess I want people to be able to move forward, avoid all that head banging. It's not at all obvious, hard to see, hard to grasp... But, I've been crowing about this for years, nothing has changed, people still do the same things, ask the same questions, pose the same answers, so, it's not like I think anything is gonna change now... I haven't been around here much lately but recently I've had a few spare moments to revisit some of these ideas, and sometimes, I guess, it can seem 'fun' to me, in my head, to beat a dead horse...

I am getting multiple BCMs and an MCM shipped to me in order to attempt to dump the firmware (h8/500 machine code) from them to then disassemble. The MCUs used in each are rewritable; see BCM Hacking. Convert 112 pin QFP Hitachi H8 CPU to.... If I succeed, we'll have a precise idea of how the management makes its decisions.

Sounds promising. I hope you succeed, keep us posted...