Impact of 'deep' discharge prior to grid charge

[eq1]

Anyone with an OBDIIC&C gauge who has tried multiple SOC resets when discharging prior to grid charging, or others who have tried any method of deeper-than-normal discharge prior to a grid charge, have any observations to share?

I've only grid charged 4 times, first two times from around 75% and not too long, second two times longer and after doing multiple SOC resets to discharge as much as possible. For those not familiar, the OBDIIC&C has a state of charge reset function that allows you to reset the state of charge to 3 different levels - low, medium and high. Among other things, you can discharge the pack using assist until you get a negative recal (for those with weak packs) and then reset the state of charge to say 40%, 60%, or 75% and discharge some more, until you get another negative recal, and reset again, and again, until you basically get a negative recal immediately after the reset. This allows you to discharge the pack more than you'd normally be able to.

Anyway, I think the deeper-than-normal discharge, followed by a slow grid charge, improved my pack performance. It's hard to say for sure because I wasn't flogging the pack after my two modest grid charges. I've been 'flogging' it more lately, and at lower states of charge than I normally keep the pack, yet it doesn't seem to threaten punking-out on me. The pack voltage seems to go lower than it used to, plus it seems to hold lower voltages under relatively higher amperage assists than it used to...

Anyone experience anything similar? 'Deep' discharge, grid charge, better performance?

[retepsnikrep]

It makes sense. Cycling the pack a bit better is probably restoring some capacity.

[joecvt]

I had two out of three Insights giving me IMA lights (P1449) on a near daily basis in November and part of December. Even with light IMA use, I would get the IMA code of death. I had to pass inspection in Janaury (check engine light on with IMA light is a problem in the state I live in) so I added a wiring method to discharge the battery with the ignition turned off.

I manually deeply discharged both cars (one at a time to see if positive results). I used a 40W / 60W light bulb and brought the voltage down to what some claim were dangerous levels (below 100v) and they think I possibly damaged some cells n doing so. Since then, I have not even had a sngle recal and hence no IMA lights or codes for more than a month. I even tried to "abuse" the batteries in both cars and so far, it has worked like a charm. I think there is really something to this SLOW deep discharge that s really worth looking into. Perhaps my approach was more careless compared to what others would have done. But considering my pack was reporting to be "toast" with IMA lights in every day driving with inspection coming up in the next month, I had to try something. So far, this technique has been quite a remarkable fix. Time will tell but so far, so good.

JoeCVT

[RedJellyBean]

Thanks for the post Joe. I think that I should do a good deep discharge of my battery as well..

[Eli]

Deep discharges are the key to rejuvenating Nickel based cells. However, to do it safely, you must do it at the cell level.

When doing it at the pack level, this is the question: Is the damage sustained during cell reversal worth a balanced pack?

I guess the answer could be yes, if you don't have anything to lose?!

Regarding the OP - yes, deep discharges by resetting the SoC like you've done will give marked improvements. Graphs to follow if I can find them.

[eq1]

Thanks all for your input... Lately I've been thinking more about the significance, importance, of discharging for pack rejuvenation mainly because of a technical explanation of the 'memory effect' in nickel-based batteries that I read... Plus, the idea that a deep discharge fixes the memory effect has been floating around for ages, I guess.

In that 'technical explanation', the author describes how overcharging likely produces an additional phase of hydrogen, nickel, and oxygen - HNi2O3 - which results in a lower voltage plateau at 0.78V. The more you overcharge, the more this HNi2O3 evolves, and the longer the lower voltage plateau becomes. This locks up capacity at the lower voltage, and all you see is the diminished capacity at the upper, normal voltage plateau. The fix is a "deep discharge."

Problem is, I don't know what constitutes a deep discharge. I wonder if going below say, this 0.78V, would be needed, necessary, warranted... The author points out that a deep discharge followed by a charge makes the HNi2O3 disappear, and thus so does the low voltage plateau...

Here's a link to the few pages from the book about the memory effect. Maybe someone can look it over and glean a little more understanding from it than I can:
https://dl.dropbox.com/u/20136699/MI..._pp226-233.pdf

[3-Wheeler]

Originally Posted by eq1

......Problem is, I don't know what constitutes a deep discharge. I wonder if going below say, this 0.78V, would be needed, necessary, warranted... The author points out that a deep discharge followed by a charge makes the HNi2O3 disappear, and thus so does the low voltage plateau...

eq1,

I would agree with Eli, in that the only way to *safely* get to a cell voltage of 0.78 volts would be to monitor all six cells per stick, when performing a discharge, and then watch for the lowest voltage cell to start dropping out.

Maybe if your current was low enough, one could stop the process it time and still get below 0.78 volts?

I would venture a guess and say that the current draw would be way less than 10 amps.

I can see one other scenario happening:
- the level of the cells starts to drop below 0.78 volts, but the other cells are still going strong.
- how do you get all the cells in the stick to get below the 0.78 volt threshold without reversing the lowest voltage cell?
- I suppose one would have to charge up the lowest cell first, and then continue on the discharge for the entire stick?
- Oh, now it dawns on me: one could simply discharge each cell of the stick, one at a time, and get each one below the threshold.

Jim.

[eq1]

I'm not fanatical enough to go down to the cell level at this point... I'm not even prepared to heft my pack out (again) and go down to the stick level... I'm mostly interested in potential 'quick' fixes, full pack stuff...

I'm currently more or less exploring the low end of my pack, or at least the middle. Well, OK, so the distance from top to middle to low-end isn't very far - so I'm still exploring it all... Just that, for a long time I was keeping my pack charged high, high as possible. And now I'm staying away from the top, that is, at least, on successive trips after an initial grid charge. It's really a pain to get my pack 'low' because I get a background charge at about 65% soc and then have to fight that...

There's like 3 'strategy' levels to my pack: the top level is above 72%, so 72% to 81%, but mostly 75% to 81%, keep soc from falling below 72%.

I drove for the longest time keeping the pack mostly in this range. And on top of that I was doing state of charge resets, more or less shoving in another 5% roughly every other day or so. This overall strategy could help me pull more capacity, yet that's merely because I could charge the pack up a bit more than I otherwise would letting the car do its own thing.

In the end it didn't do anything to help the pack perform better; I couldn't get higher amperage assists; I couldn't take the pack down any lower; etc. Plus, after reading that book and other things I started to realize that I should stay away from the top; I'd likely get better performance in the middle, albeit perhaps in lieu of that extra capacity...

The second level is like 70% to 75%.

Here I basically let the state of charge fall below 72% and let the car positively recal, seek its own 'natural' 75% SOC level. Really, what I was doing with state of charge resets and keeping above 72%, and what the car does on its own with positive recals at 72%, is very similar.

The car by itself pumps in something like an extra 2% capacity daily; I think part of the programming accounts for self-discharge, likely a set level, not a dynamic sort of thing (i.e. the programmers perhaps believed NiMH batteries had a self-discharge rate equal to say 2% - so they programmed the car to put in an extra 2% per 24 hours, on average)...

Anyway, this 'strategy', this level, keeps state of charge below 75% but doesn't go too far down, such as below 65% where my background charge usually kicks in...

The third strategy level thingy is keep SOC below 72% -- basically, background charge quits at 70.1% so keep it about there and below... Of course, problem here is that background charge: with it kicking in at 65% or so I've got only this 5%-7% window between 65% and 72% - for optimal driving... The last two times I grid charged was after drawing the pack down as low as I could with SOC resets. Since then I seem to be making some headway, seeing better performance, so I wonder how far I can take it, whether others have done or are doing the same...

I want to draw my pack down, utilize the middle range, but background charge gets in the way. I want to experiment with deeper-than-normal pack discharging and slow grid charging to see if I can gain anything at the low end of the pack...

I wonder if pack voltages I see when I get 'down there' reflect few individual cells dropping out or mainly most of the cells just being weaker than normal cells... I mean, when I draw my pack down I'm seeing pack voltages like 145V at rest, at say 55% state of charge. On a good pack I'm pretty certain pack voltage would be way higher; on a good pack I'm guessing that at about 55% SOC, after a good strong, long assist event, pack voltage would be popping back up to say 158V, not lingering at 145V...

Someone else should read that chapter I posted up there...

[3-Wheeler]

Originally Posted by eq1

.....I wonder if pack voltages I see when I get 'down there' reflect few individual cells dropping out or mainly most of the cells just being weaker than normal cells... I mean, when I draw my pack down I'm seeing pack voltages like 145V at rest, at say 55% state of charge. On a good pack I'm pretty certain pack voltage would be way higher; on a good pack I'm guessing that at about 55% SOC, after a good strong, long assist event, pack voltage would be popping back up to say 158V, not lingering at 145V...

Someone else should read that chapter I posted up there...

eq1,

I would say from both 1) measuring voltage on the entire pack and using MIMA, that a good pack will drop to a certain voltage, let's say 144V under load, and pretty much stay there after the load is removed. Yes it will pop up a little bit, but not much.

2) Same thing for testing individual sticks on the bench. The good, high capacity sticks always showed less "bounce" in voltage when applying the load, and after it is removed.

The poor, low capacity sticks "bounced" from a loaded 144 volts up to a much higher number after the discharge stopped.

Jim.

[Hugh-Falls]

Originally Posted by 3-Wheeler

eq1,.............The poor, low capacity sticks "bounced" from a loaded 144 volts up to a much higher number after the discharge stopped.Jim.

On the INDIVIDUAL CELL LEVEL a discharge of the "bounce" and then discharge of the subsequent "bounce" followed by a third "bounce" discharge seems to improve the ability of the cell to retain charge in my limited experience.

This phenomenon, if true, may be related to the reactions described in the book excerpt provided by eq1. How this relates to what has been called "cell reversal" would also be worth more study.

eq1: Is there any detailed analysis of "cell reversal" in your book?