Nicely done Eq1.

One thing that you might have mentioned though is that if electrolyte has off-gassed there will be patches that are dead and unrecoverable. They will never come back.

Where you discuss "time cell reversal" you mention low amperage drain as another variable. I think that is at least as important at time spent reversed. I have started to drain packs with only a 40 watt bulb. I would use a 25 watt bulb if I had one. I do an initial grid charge because I figure the cells could be way out of synch but after that it's down, down, down. I stop the drain periodically to allow the cells to recover then continue.

Usually I then disassemble the pack and continue the drain at the stick level. Some sticks go on and on at .1A to the bottom. I had one stick yield over 2000 mah. That must have been chemically locked up capacity: what else could it be?

Before I reassemble I charge the sticks but not right to their maximum. I let the grid charger do a finishing charge. I am considering letting the grid charger do the whole recharge and skipping all that recharge handling of individual sticks.

Rick

 

For my HCH2 packs, I use 500W halogen to 120V then a 40W to near zero following an initial topping grid charge.

https://docs.google.com/spreadsheets/d/1j1mFp6DKPmuQGnrwF-BjXwDEGWcPJtIF-7jofdFVYMI/edit#gid=0

First sheet is the 500W, and the 2nd is the 40W with 360 Ohm resistor numbers for comparison.

I'm happy with the 40W as it feels like a good balance between low current and duration of reversal. I don't remember exactly, but the whole trip from 120V to 1.8V was maybe 5 hours.

If you're looking to save yourself the effort on the stick recharge, I see no harm in using a ~24 hour grid charge instead. I do both because I'm impatient, and I don't want to wait the 24 hours. With the HCH2 sticks, I use regenerative discharges, so the prior discharged stick-pair is recharged by the next up to about 70% capacity (leaving only the last two pairs fully discharged). My total recharge time on all sticks is about an hour with the two Reaktor chargers. Even after a full stick recharge, I'll run the grid charger for at least a few hours until I get peak voltage and start getting a little heat off the cells.

I've done this with 4 packs, and the recovered capacity under a 4A load (500W halogen) has been 25-33% between the first discharge to 132V and the discharge following a deep discharge and recharge.

EDIT: BTW, eq1, nice summary and graphics.

 

True, there's things that happen that result in capacity lost forever, like out-gassed electrolyte. My illustrations just aim to paint the broad strokes... Also, I agree that discharge rate is as important, if not more important, than time when it comes to cell reversals. I don't really know for sure, but I presume high rate reversals not only engender unwanted, bad reactions, but also a lot of heat and pressure, like a triple whammy. Low rate reversals probably just entail some unwanted reactions - create hydrogen gas - and maybe, actually, if it's slow enough the hydrogen can be re-absorbed in some manner, not pose a serious threat of irreversible damage...

 

Here's another strategy I was just about to post somewhere else but decided it might go here:

The truly low-no-budget, easy partial pack fix might dispense with grid charging - you'd deep discharge with a light bulb, reset the computer, charge with the car while parked, holding engine speed at 3000rpms. I've never tried it and never heard of anyone trying it, but it'd probably work, it'd probably do more for the pack than the grid charges people have done for ages... You need to spend almost nothing for this approach - you attach a light bub to clips and cord, attach the clips to the main IMA batt terminals, turn key on, and let it discharge...

The downside of this approach is, one, you probably should have grid charged first to limit reversals, and two, the charge after the deep discharge should be to 100% full at a low rate. Charging in the car will be about a 7A charge if you hold engine speed at 3000, yet it's kind of tricky, touchy, it could be 5-12A 'at 3000', i.e. very minute engine speed differences effect the charge rate... The state of charge will 'positively recalibrate', it will jump up to the car's full mark, whenever the car feels like it - you'd rather charge too 100% full and keep the car's wishes out of the equation, at this point.

BUT, in a pinch, on a lark, on a whim, if one wanted to he or she could do a deep discharge then charge with the gas engine and probably have good results... No grid charge harness to mess with, just a light bulb discharger that you can fabricate from stuff probably lying around the house... If it doesn't help enough you could just do it again, the discharge, but charge with a grid charge the second time around. Someone should try this...

 

I agree IMO grid charging only without deep discharge is mostly a waste of time. I've recovered a pack that was from a junk yard so it sat for a long time and it's voltage was less than 40VDC total output and was usable in my car after and didn't throw an IMA light (car is still being driven 6 months and 10,000 miles later with no battery recals and no IMA light). IMHO the best way to get max capacity regained from an aging battery is to do a stick level reconditioning. This has yielded me great success and is the best and longest term fix until you can buy a new battery.

The most obvious fix to aging hybrid batteries is to just plan on replacing the pack every 5-8 years or 150,000 miles like the manufactures originally suggested.

 

....IMHO the best way to get max capacity regained from an aging battery is to do a stick level reconditioning...

If the "stick level reconditioning" involves data logging/graphing, internal resistance calculations/measurements, finding matched cells and creating the pack from those, then I would agree. Otherwise there isn't a difference, certainly not one that justifies the extra time...

 

Nice graphics. Puts the mental picture on paper.

 

I think one thing people should realize is that there isn't a whole lot of difference between overcharge and overdischarge. Both are bad and damage the cells in similar ways.

The idea that overcharge is somehow OK and overdischarge must be avoided at all costs is incorrect.

I hope to post something similar to this overall, showing people exactly what happens when you grid charge/discharge, and outlining what the process should really be to minimize damage.

It should definitely be a combination, and I think it should be a gradual thing. There's no need to slam the packs for 24+hours initially, I don't think. Which is contrary to the note I posted not long ago, still trying to figure this part out..

I've experimented with deep discharging and then letting the car charge the pack up by itself. For whatever reason, the results weren't favorable - the pack gave more output consistently after the grid charge. Need to do more experiments.

 

....there isn't a whole lot of difference between overcharge and overdischarge. Both are bad and damage the cells in similar ways....It should definitely be a combination, and I think it should be a gradual thing. There's no need to slam the packs for 24+hours initially, I don't think....

My understanding is that overcharge - whether due to a real overcharge, continuing to charge when the cells are 100% full, or due to too high a charge current no matter how full - causes the reactions that produce an amorphous hydrogen-nickel-oxygen phase that depresses discharge voltage and increases charge voltages (i.e. voltage depression, 'memory effect'). I've never read anything similar about reactions on over discharge, though maybe that's simply because I've never come across anything...

I don't have a good enough facility with the electrochemistry to actually deduce what reactions will happen during either process. Gaseous oxygen is created during overcharge, and gaseous hydrogen is created during overdischarge, that's about all I know. Understanding what can happen to these would be a decent next step, for me... Also, I have more firsthand experience with overcharge and very little with overdischarge... But in any event, in my mind overcharge is worse, or at least it's the problem that we have to deal with more often...

On not needing to 'slam the packs for 24+ hours', I agree with that. I don't think it should be necessary to charge more than 6500mAh tops for that first topping-up charge before the deep discharge...

I don't know, contrary to what the common knowledge seems to be, I don't find the Insight cells very inefficient; you can charge about 6500mAh and pull nearly that much out of them. It seems to me that the extent to which people think the cells are inefficient has more to do with erroneously overcharging them in the first place, like waiting for 'delta V', by which time one has already overcharged the cell. If you charge them more than needed of course it's gonna seem like the cell is inefficient - you 'put in' 7200mAh and 'pulled out' 6400mAh, about 89% efficient. But if you only 'put in' 6500mAh you'd probably pull out that same 6400mAh, 98% efficient... I'm still kind of working with this, but in general, I'm finding that I can pull just about the same amount out of cells whether I charge to the typical standards or to what would seem undercharged by comparison to those same standards...

....I've experimented with deep discharging and then letting the car charge the pack up by itself. For whatever reason, the results weren't favorable - the pack gave more output consistently after the grid charge. Need to do more experiments.

Good to know. What rate did you charge the pack up with the car? Was it consistent?

In general, yeah, a slow, consistent rate should be better, like the slow grid charge rates. Yet I wouldn't think there'd be too much difference if the pack were charged at say a 1C rate. Maybe if there is a big difference it might have more to do with not charging the pack all the way - if the car stops charging before completely full, as we generally understand how it's done...

My conception is that the deep discharge drops the positive electrode potential low enough that the 'amorphous HNi2O3' breaks apart and that, upon charge, it's no longer there to obstruct the regular reactions. But during this process, the electrode/s are being reformed in a manner - the reactions will take a bit longer and lower current will force a more complete conversion.

If the current is too fast, it seems like actual locations along the electrode will get skipped over... It's not totally random; the Ni(OH)2/NiOOH interface, where the reaction takes place, moves 'across' the plate, moves 'along' the plate - it exists in an actual location depending on state of charge, whatever that means with rolled-up electrodes. So once a 'sector' of the electrode is 'processed' - it's done, the reaction interface has already moved on.

I imagine it's not so cut and dried, like one second the reaction is taking place along a given sector and then, boom, it's moved onto another sector. But rather, most of the reaction is taking place along the interface and once that region is charged, once the electrode has reached the proton-deficient limit, the interface has moved and most of the proton-stripping is happening somewhere else.

SO, if the current is too much, the interface moves too fast for the messed-up areas of the electrode to be re-processed. The spacing is larger where the HNi2O3 once occupied; the interstitial sites, the galleries, where normally protons reside, have been cracked open more than usual. The reaction speed depends in part on that spacing, for diffusion or something; the larger the spaces, the longer it takes for the stuff to diffuse... So after the deep discharge, the HNi2O3 or whatever has been removed from locations on the positive electrode. And the spaces that need to be re-processed upon charge are larger than normal, so the reactions, the normal reactions, will take longer; hence, you need a lower current... Something like that...

 

EQ1, nice summary and graphs. I've read it twice and still a little puzzled (only 8 years of college so please bear with me). If a car has been sitting for a year, like many of mine are, which is better:
1. Grid charge/Discharge/Grid charge
2. Discharge/Grid charge
1 or 2?
And next question:
A. Discharge as low as possible
B. Charge/discharge to 100V/charge/discharge to 75V/charge/discharge to 50V/charge
A or B?
Answering the above would be a big help to me.
Thanks
Gerald

 

....If a car has been sitting for a year, like many of mine are, which is better:
1. Grid charge/Discharge/Grid charge
2. Discharge/Grid charge

I'd probably check total pack voltage and base my decision on that. If it's low, like below 130V or so, I'd just deep discharge from there. Actually, I'd probably just deep discharge first no matter what the voltage was, but chances are, voltage will be low and the cells are already almost completely discharged. The pack is almost there already - Why muck it up with a full charge? Just discharge so you're sure that all the cells have reached the 'critical' low voltage threshold, and then do the full charge... If total pack voltage is below 60V, you probably don't need to discharge anymore, though...

Some of this answer is based on what I've seen with sticks that sat for over a year. They already self-discharged and upon full charge, a massive amount of voltage depression was wiped out... If the pack's below 60V, then chances are that self-discharge, at a very very low rate, has essentially totally discharged the pack...

And next question:
A. Discharge as low as possible
B. Charge/discharge to 100V/charge/discharge to 75V/charge/discharge to 50V/charge
A or B?

This kind of stepped discharge seems too cautious and takes too long, plus it doesn't go low enough. Though it's not like I have a vast database to consult for answers...

I want to try to avoid getting into the weeds of 'pack reconditioning'. I have a general theory and some working experience, and then there's a lot of 'this' and 'that' that will depend on individual circumstances and individual risk tolerance...

The stepped approach aims to reduce risk of cell reversal while still gaining the benefits of deep discharge. The problem I see with it is that 50V at the typical discharge rates doesn't guarantee that all the cells have dropped low enough. Ideally you'd be doing each cell alone and dropping it to something like 0.4V at 0.2A. If you knew your pack were perfectly balanced, then that'd be a total pack voltage of 48V. But of course your pack isn't perfectly balanced, that's why you're working with it. The only way to know that all the cells have dropped low enough is to drop pack voltage to near zero (or do cells individually)...

Stepping in general would be a decent way to minimize reversal time and frequency, so if you're concerned about reversals you might do a stepped approach. But even after stepping, 50V doesn't seem low enough to guarantee all cells drop below the critical threshold. I'd go lower. Personally, I wouldn't bother with a stepped approach...

I've dropped one of my packs to near zero twice now. Before that I dropped it to about 22V. And before that it had undergone all sorts of the typical IC reconditioning regimes, even stick level. None of the fiddling helped as much as the near-zero drops. Even the 22V drop wasn't low enough - the pack was clearly 'fixed' only after the first near-zero drop... Granted, it's not like it's a new pack - it's old and worn and its age shows. But it obviously isn't being effected by voltage depression - the difference is clearly felt, and since I have an OBDIIC&C I can see differences in the IMA data values...

Anyway, my proclivities are probably clear by now. If your pack has sat for a year, give it a short discharge to be sure all cells drop, and then charge it up. Forget the stepping, go lower than 50V... If you still have problems, call hybridautomotive or bumblebee...

 

Can you clarify how you arrive at the generalization that charging past 100 % SOC doesn't restore 'bad' (lost) capacity, yet discharging below 0% SOC (~<0.8 V) does? NiMH doesn't fair well when over-discharged, particularly to cell reversal. Your explanation is clear, but I'm not sure it's factual.

I'm approaching this from a purely academic level, as I've not empirically tested your scenarios, but as an electrical engineer specializing in DC power distribution, I remain skeptical. Fortunately, the scientific method doesn't care if I'm a skeptic, so I suppose I should do this, as I've always been quite skeptical of overdischarge routines 'restoring' batteries.

I know you take a lot of data on your G1 pack, but have you ever tested a single cell (not stick) to better understand the chemistry without worrying about the pack BMS? Under controlled conditions?

Fortunately, I've got 6 or so sticks lying around that I don't care about, so I'm going to set up a test with a programmable source measurement unit (power supply + programmable load) and run it on each individual cell with the following profiles:
-Individually Charge all cells to full @ 4 A (as indicated by mV drop at 100 % SOC)
-Individually discharge 100 % of cells to 0.8 V @ 30 A, then
-Individually discharge 66 % of cells to 0.0 V @ 4 A, then
-Individually discharge 33 % of cells for another 2 Ah @ 1 A (to simulate a reversed cell condition)

I'll then bulk charge each cell @ 1 A until 0.8 V, then bulk at 25 A until mV drop, then I'll float for another 10 minutes @ 500 mA (not really important with a 1S 1P 'pack' mV-dip measured).

Finally, I'll discharge each cell at 30 A (CC) until it hits 0.9 V and note ampacity. I've got six sticks, so 36 cells... 12 of each.

Let me know what changes you'd like me to make... my test equipment is fully automated, so it'll just be switching cells out. Once we have the individual cell data, we can extrapolate to the pack level. I imagine the 3rd party battery guys have already done this?

 

Can you clarify how you arrive at the generalization that charging past 100 % SOC doesn't restore 'bad' (lost) capacity, yet discharging below 0% SOC (~<0.8 V) does? NiMH doesn't fair well when over-discharged, particularly to cell reversal. Your explanation is clear, but I'm not sure it's factual.

I'm approaching this from a purely academic level, as I've not empirically tested your scenarios, but as an electrical engineer specializing in DC power distribution, I remain skeptical. Fortunately, the scientific method doesn't care if I'm a skeptic, so I suppose I should do this, as I've always been quite skeptical of overdischarge routines 'restoring' batteries.

I'll save eq1 the trouble...

It's not a secret that deep discharging Nickel cells 'restores' them - it's been known since the NiCd days, and was practiced extensively on NiCd aircraft batteries.

http://batteryuniversity.com/learn/article/how_to_restore_nickel_based_batteries

However, I think why is pretty new knowledge - Read this, courtesy of eq1:

https://dl.dropboxusercontent.com/u.../u/20136699/insight/Advanced_Batteries_Robert_A_Huggins_TOC_Ch1_10C_11B_11C.pdf

Ultimately, NiMH is difficult to manage because the 100 % SOC condition isn't a function of voltage.

It's probably time this piece of misinformation is laid to rest. I know I've perpetuated it in the past too, but I now know that it's not true. It has done the Insight community in general a huge disservice. Mike never paid attention to voltage. His grid charger only has 0.3V resolution. Nobody has ever paid attention to voltage......... Could that be the reason that "voltage doesn't matter"?

That's all I'm really going to say about it lol.

 

Ultimately, NiMH is difficult to manage because the 100 % SOC condition isn't a function of voltage. Lithium is actually much easier to reliably charge, as the cathode charge function is self limiting at known voltages (e.g. 3.65 V indicates 100 % SOC for LiFePO4). This is one possible reason a drop in Lithium battery might be a better long term solution. I'm still really wanting to design a true drop in lithium replacement... 6 Ah for under $2000 is a reasonable goal. No increased capacity, but much higher cycle count, much lower self discharge, MUCH longer lifetime, etc.

 

Yeah. In my understanding, it's the gas production in both cases that is detrimental. Even though most can recombine if the rates of production are low, some water is still lost to corrosion and other unwanted reactions. The physical gas production itself damages cells mechanically, forcing structures apart, etc.

You could almost say that deep discharging is better than overcharging because it doesn't cause any nefarious chemical states - it only fixes them.. Until cell reversal, I suppose.

It's all Mike's fault. I mean that in a truly endearing way. It's just funny, it could have easily happened that deep discharging was what was stumbled upon first rather than grid charging, and that would be our world.

The Prius folks are all about the discharging, but are weary about the grid charger - it's a new concept to them.

 

Hmm. If we are talking about IMA light throwing packs here, if the pack voltage is below say ~155V, there are already cells at ~0% SoC.

So if the goal is to minimize reversals, you do need to charge first.

We should probably clarify the goals. The first goal is to balance the pack. To balance 'properly' - without the need for [excessive] overcharge or over discharge - takes time, but it can be done. Think about cycling sticks on hobby chargers - they only get taken down to ~0.9V or so, but they do eventually balance, it can just take 6 or more cycles. The more the overcharge[or overdischarge, but nobody ever does that on a hobby charger], the faster the balance...

Once the pack is balanced, you can deep discharge with as little reversal as possible, and grid charge with as little overcharge as possible.

If one doesn't care about these things, then the whole process can be done in a single deep discharge/recharge event - for better or worse.

It should be understood that "reconditioning" packs in situ will never be ideal. It will never make a pack "perfect" again. In almost all cases, it's a band-aid and the real reason the pack went out of balance in the first place will continue to cause it to go out of balance. Though I suppose getting 6mo-1 year between having to do procedures is certainly meaningful to a driver, I guess from my point of view I'm trying to make packs that last 3+ years.

The reconditioning process really becomes a concern when trying to rebuild a pack with that goal in mind. You have to pick which sticks go together on a very high level.

Remember to get rid of the idea of "good" and "bad" stick. There are, of course, bad sticks. But equally, there are just sticks that don't belong together.

It's entirely possible to have a pack full of 20 good sticks that throws errors, lol. Think about that one for a bit.....

I'm playing with the concept of $/mAh for used packs. So you would know exactly what you're buying, with a fanciful chart and stuff.

Something like $0.35/mAh, and warranty tiers based on mAh. So a "perfect" 3900mAh pack would be $1365, if I could find one - and carry a 2 year warranty. A 2000mAh pack would be $700.... and carry a 6mo warranty. Or something like that....

 

....The first goal is to balance the pack. To balance 'properly' - without the need for [excessive] overcharge or over discharge - takes time, but it can be done. Think about cycling sticks on hobby chargers - they only get taken down to ~0.9V or so, but they do eventually balance, it can just take 6 or more cycles....Once the pack is balanced, you can deep discharge with as little reversal as possible, and grid charge with as little overcharge as possible....

We definitely have different conceptions of what's going on. I see the deep discharge as an essential part of balancing - that's partly what's depicted in my graphics. In order to fully charge and balance the cells, one needs to 'get rid of the yellow squares', and that's what the deep discharge accomplishes.

Contrary to what you must have experienced, or at least what you believe, NOT being able to balance the cells by just cycling between the typical voltages is what I've experienced. I guess I can't be much clearer about that: Given a pack of cells on the fritz, one cannot balance them without deep discharging them...

So my strategy says get that out of the way asap, deep discharge them, eradicate the bad stuff, get rid of the yellow squares, and then charge and cycle till your heart's content...

 

Deep Analysis!

For 99.9% of people on here including me the process is pretty clear.

If the pack has sat for >6months and has self discharged then discharge it further with a load to zero volts.

If it hasn't sat then charge it for 24hrs then discharge it.
Then charge for 24hrs minimum and enjoy new lease of life.

If that doesn't work and it's obviously not a single failed cell then bin the sticks and buy some new ones.

 

I guess I don't 'draw the same lines'... In my experience, I've cycled sticks the 'hobby charger' way and the cells still remain unbalanced. They only become balanced after the deep discharge. Cells will have different amounts of voltage depression, i.e. if you graph the charge or discharge curves the curves are different, some cratering earlier than others upon discharge, for example, or darting up faster, sooner upon charge. That's unbalanced to me... Isn't that unbalanced to you?

Yes, it is... mostly. 0% Soc is still where the line is drawn - like I said, the car will deal with voltage imbalance - it will not deal with capacity imbalance. Obviously the two are interrelated. Hmm. You must be testing pretty crappy sticks. Want another box to play with? lol

Well, OK. So let's back up here a little bit.

You understand the concept that these chemical reactions are probabilistic, not absolute. So the same concepts you were applying to charge absorption also apply to discharge. As the voltage falls towards the end of the discharge curve, some of the "overcharge stuff" is converted back into active material as related to the current carrying capacity of the cell.

So there is an element of discharge balance even with a 0.9V cutoff while hobby cycling. It just takes longer - I've seen it take 8 cycles. The hobby chargers actually tend to do a "better" job, since they use stick voltage. My equipment can see individual cell voltages and negative deltaV, and stop appropriately, while the hobby chargers cannot. What typically happens is that the 80% out of balance cells hit negative deltaV but the 0% cells are still rising in voltage, masking it. That's how they end up venting cells.

In the car environment, cells do actually reverse - specifically because of the higher currents. That means they hit - and pass the 0.4V threshold, if only for a moment. It doesn't have to be at 0.2A, it's the voltage that matters. It just has to be at low current to sustain at the low voltage. But any reduction is progress.

The cells at 0% SoC are effectively deep discharged as far we're concerned, because they exhibit no, or negligable voltage depression upon being charged to 100% SoC - especially in comparison to the cells that were at 80% SoC, after a 24.. or even 16 hour grid charge. They have been run to 0V over and over again, with each negative recal. Balanced packs don't recal, cells don't drop out.

I haven't seen a pack in recent memory that didn't have cells at both 0% SoC and 80% SoC... +/- 20%. That is the defacto standard of an IMA light throwing pack.

And regarding the business aspect - we're getting down into the pretty nitty gritty of pack rebuilding. I'll post some data to back up my claims, but I likely won't be able to elaborate much further lol. We have tens of thousands of sticks in our database now though. The trends are very clear.

Now we will replicate my theories of the failure process with the beta rebuilt packs and if they verify, we should be able to start offering quality rebuilt packs soon.

I'll find a core stick out of the... 720 currently on the floor that exhibits what I'm talking about strikingly, couldn't find a good example in the database with a quick search. Will be done testing in 6 hours, with my HDD 250 megabytes smaller.

 

....You understand the concept that these chemical reactions are probabilistic, not absolute. So the same concepts you were applying to charge absorption also apply to discharge. As the voltage falls towards the end of the discharge curve, some of the "overcharge stuff" is converted back into active material as related to the current carrying capacity of the cell.

So there is an element of discharge balance even with a 0.9V cutoff while hobby cycling. It just takes longer - I've seen it take 8 cycles.... In the car environment, cells do actually reverse - specifically because of the higher currents. That means they hit - and pass the 0.4V threshold, if only for a moment. It doesn't have to be at 0.2A, it's the voltage that matters. It just has to be at low current to sustain at the low voltage. But any reduction is progress. The cells at 0% SoC are effectively deep discharged as far as we're concerned, because they exhibit no, or negligible voltage depression upon being charged to 100% SoC....

Lots of ideas here... As far as reactions being "probabilistic" and me understanding it, I barely understand that. I mentioned something related earlier. I said something to affect of 'it's not totally random'. The two phase NiOOH/Ni(OH)2 interface, from what I understand, moves across the electrode - the main reaction takes place at that interface. So, although probability I guess determines some aspect of the reactions, it doesn't totally determine the location of those reactions, I think.

Huggins has a page or two that describes 3 types of insertion reactions (6.5 "Types of inserted guest species configurations"). One entails "random diffusion of guest species into the gallery space," another entails "motion of two-phase interface when guest species is not ordered upon possible sites," and the third entails "motion of two-phase interface when guest species IS ordered upon possible sites in the gallery space." NiMH positive electrode is mostly the second type - not totally random, but not totally ordered...

What this suggests to me is that, as charge or discharge takes place, whether a reaction happens or not might be probabilistic, or, whether one reaction at one location happens vs. another of the same type of reaction happens at another location is probabilistic. But, perhaps we might say that, at the two-phase interface, the probability of reaction is way higher versus the probability of reaction at locations far removed from that interface...

As far as I understand it, when we're talking about this moving interface, its location along the electrode, we're actually talking about the state of charge; so the interface would be in one area when the state of charge is high and in another area when the state of charge is low... I don't know any of this for sure in the real-world; I don't quite understand how this works when you have electrode plates that are coiled-up like they are in the Insight cells. It's hard to picture. But it seems like, given the type of reaction and the existence of a two-phase interface, there must be some element of localization, an order or sequence to the reaction in space vs. state of charge...

Anyway, this discussion of the two-phase interface and random vs. ordered etc. mainly has to do with deep discharge and what it takes to get rid of the HNi2O3 (I'll just call it that now, instead of 'bad stuff').

You say that voltage is what matters, that the discharge rate doesn't have to be low. But I haven't found that to be true... On the other hand, I'm just now sort of getting where you're coming from: you'd prefer to do 8 cycles, taking a long time, getting rid of most of the voltage depression, opposed to risking damage due to cell reversals...

Regardless, I was gonna say that I started with normal, relatively high discharge rates and going to low voltage for the few seconds it takes, and it didn't do anything. It wasn't until later that I realized the most effective treatment would have to be low current. It's sort of counter-intuitive, cuz at higher rates you easily reverse cells or drop the voltage to zero quickly, and you think that the cell should be fixed, that the low voltage is all that matters. But it's not like that.

There's a voltage range that needs to be 'burned', and I generally accept that it falls around the secondary low-voltage plateau that Huggins hypothesizes, at 0.78V... But unlike when I first started with this stuff, I realize now that 0.78V isn't a fixed threshold of some sort, but rather, it's a bit like how the normal voltage plateau works in our cells, where charge and discharge happen more or less around those voltages (like 1.34V and 1.25V)... The big difference is that the normal reactions can sustain at high current, whereas the reaction that gets rid of HNi2O3 apparently cannot - that's what I've found, and it makes sense to me that it wouldn't...

Anyway, I totally get what you're saying now. You're saying that you can still achieve some reduction of HNi2O3 with cycles to 0.9V (you're actually saying total reduction). It may take 8 cycles, but it can be done. Ouch, yeah, I kind of agree. I'm not sure you'd get it all, and I've never had much luck with these kinds of cycles. But some reduction should/does happen... I've seen something like a 70% reduction going to 0.9V at 0.6A (but keep in mind that most 'cyclers' are going down to 0.9V at like 6.5A or 10A, not to mention they're working with 6 cell series sticks, with cell SoC all over the place)... It seems like a waste of time when you can get it all in one low current discharge and one charge... You must really be paranoid (ok, concerned) about cell reversals. I'll have to sit down at some point and really think about reversals, cuz my gut tells me low current reversals are no big deal... Maybe I'm wrong...

 

where does huggins talk about H2 gas evolution on over discharge?

if you look at the tie line he talks about that follows the discharged from the over charged state where nickel has the high valence above 2 and is over to the tight of the normal 1.334V tie line you can see that as that discharge goes back down towards the H corner then there is no stable chemical state that is in equilibrium with the resultant products form the discharge.

this is where i am assuming the excess protons which can no longer be stored in the interstitial spaces are forced into the solution and recombine with oxygen to produce water. since i am assuming there is an excess of water surrounding the electrodes at this point i did some slight mechanical agitation at this point of total discharge also.

i am thinking that some of these problems of limited capacity have to do with how the electrodes may be 'shielded' from the electrolyte by insoluble byproducts of the charge/discharge cycle. these products could form around the electrode and impede access of the electrolyte to the electrode which would create high internal resistance at the local level and impact the charge/discharge profile.

so at maximum discharge, maximum water content, i used a small engraver tool which i had altered so that it had a blunt tip instead of a pointed tip. i used a small 2mm screw with a flat head, threaded into the end piece that normally holds the sharp tip. then i used that vibrator to vibrate the sticks from each end and then used it on the sticks i had out of the case by running it up and down the sticks and vibrated them slightly also from the sides. i also put a block of wood on top of the pack and went over the entire bottom side of the pack and whacked the block of wood while the pack was discharged to send shock waves through the sticks in case that could help to separate the insoluble layer from the electrical connection they may have to the electrodes. i was hoping that if this insoluble layer, something like ice layer on water or similar in my imagination, was disconnected from the electrode then it would not be at the same potential as the electrode and that would allow it to dissolve in the electrolyte again.

i intend to do something similar on the civic hybrid pack.

so the over discharge is not forcing gaseous hydrogen out of the can imo. but this is where the discharge is risky in that the valance state of the nickel may be reduced enuff that it can become resistant to oxidation again on the charge cycle.

i recommend that the first deep discharge not be done while the pack is intact and in serial form. i recommend the sticks be combined in parallel and then discharged together in parallel down to the .2V level slowly. i think that is the safest way to force the high cells inside the sticks to not force excess current through the cells in the stick that have reversed.

as each stick reaches the .1-.2V level there will be always one or two cells which still are holding on to some voltage while the other cells in the stick are reversed. at this point i tried to keep the discharge current under 200mA so the damage could be limited.

i have finished restoring my insight pack and put it back in the car with the balance charger already but i am gonna restore the IMA pack in my civic hybrid next and then i am gonna use a special discharging probe so that as all of the sticks are being discharged in parallel, i can probe the cells with voltmeter probes through the shrink wrap to find the high cells and then discharge them more as the stick is drained.

when i find the high cells in the stick which are holding on to their 1.4V charge and the other cells in the stick are in reversal then i am gonna use this probe made up of two voltmeter probes with a 5 ohm power resistor across the leads. so i can then stick the probes through the plastic and short out the individual high cells in the stick and drain them off to the same low voltage as the others in the stick.

then once all of the cells in the stick are drained down to less than the .3V level or so then i can go recharge the sticks back up again but not to full charge.

then discharge again to low voltage from this partially recharged state back down to totally discharged again. so the chemistry never follows that 1.334V tie line over to the full charge state. it is discharged back down that tie line past the stoichiometrically appropriate state for the Ni (OH)2 back to the level where excess protons are forced out of the lattice again and water created from the excess.

i will do two cycles like that and then charge the pack back up to full charge using the imaX B6 to charge the sticks up in parallel up to about 8.60V each. then reassemble the pack and continue charging with the balancing charger so it forces the entire pack to have every cell at full charge. the little balancing chargers, the led drivers peter found, push about 257mA of current at the 172V level so there is no risk of damage from overcharging.

then do one final total discharge of the pack into a dummy load at the .2C rate to evaluate capacity. then use those capacity numbers to determine if i wanna take the pack back apart and remove and replace individual sticks or if i then just put the pack back in the car with the balancing charger attached to the IMA pack.

instead of the spade lead doubler on the precharge resistor leg, i use ring terminals on the leads, along with a 20A fuse on the positive electrode tap, and put them under the screw that holds the copper conductor onto the plastic contactor housing on the positive lead and similar on the negative lead. once this final discharge is complete and i can decide if i wanna open it again or not, then i can screw down all of the the thermistor lead screws again. save that to the very last.

you can see the center tap i used on the top of B10 to connect the lower led driver above the diode and the upper led driver connects to the top of B10 up to B20. the second little balancing charger is from that white cased led driver and i mounted them back to back with the serial connection being a reused diode leg that i pushed through the foam holding them apart back to back. the top led driver has the diode on top and this balancing charger is mounted between B0 and B20. total cost about $17 but this unit gets very hot. too hot as it is incased inside the plastic box. i will have to change it but it still pumped the pack up to 173V.

 

where does huggins talk about H2 gas evolution on over discharge?

if you look at the tie line he talks about that follows the discharged from the over charged state where nickel has the high valence above 2 and is over to the tight of the normal 1.334V tie line you can see that as that discharge goes back down towards the H corner then there is no stable chemical state that is in equilibrium with the resultant products form the discharge.

this is where i am assuming the excess protons which can no longer be stored in the interstitial spaces are forced into the solution and recombine with oxygen to produce water. since i am assuming there is an excess of water surrounding the electrodes at this point i did some slight mechanical agitation at this point of total discharge also.

i am thinking that some of these problems of limited capacity have to do with how the electrodes may be 'shielded' from the electrolyte by insoluble byproducts of the charge/discharge cycle. these products could form around the electrode and impede access of the electrolyte to the electrode which would create high internal resistance at the local level and impact the charge/discharge profile.

so at maximum discharge, maximum water content, i used a small engraver tool which i had altered so that it had a blunt tip instead of a pointed tip. i used a small 2mm screw with a flat head, threaded into the end piece that normally holds the sharp tip. then i used that vibrator to vibrate the sticks from each end and then used it on the sticks i had out of the case by running it up and down the sticks and vibrated them slightly also from the sides. i also put a block of wood on top of the pack and went over the entire bottom side of the pack and whacked the block of wood while the pack was discharged to send shock waves through the sticks in case that could help to separate the insoluble layer from the electrical connection they may have to the electrodes. i was hoping that if this insoluble layer, something like ice layer on water or similar in my imagination, was disconnected from the electrode then it would not be at the same potential as the electrode and that would allow it to dissolve in the electrolyte again.

i intend to do something similar on the civic hybrid pack.

so the over discharge is not forcing gaseous hydrogen out of the can imo. but this is where the discharge is risky in that the valance state of the nickel may be reduced enuff that it can become resistant to oxidation again on the charge cycle.

Sorry, but I think you're way off base here. The hydrogen/oxygen recombination cycles on overdischarge/overcharge are well known/documented. They are one of the reasons the nickel cell is so robust.

It's an easy experiment to do. Reverse a cell at 20A for about 3 minutes and it will vent. Ditto with overcharge...

There is an absence of water in used cells, not an excess.... Water is consumed as a byproduct of unwanted chemical reactions, not as part of the cell chemistry. It's just a carrier.

Wear goggles - good cells are pretty juicy and can squirt fluid some distance. While bad cells just make farting noises.

 

By overdischarge I specifically mean cell-reversals. You're right in the sense that you can't "excessively overdischarge" a single cell - once voltage falls to 0, current stops flowing and that's it.

But in a pack of series cells, reversal is very easy. It should be realized that in bad packs, cells are dropping out in the 150V range resting. I think the stepped/tiered approach is better from the standpoint that you aren't leaving cells reversed for excessive periods of time. It's the polar opposite of grid charging, it's just that grid charging also has the negative of causing voltage depression.

But a single deep discharge to 0V and then a grid charge would be the quickest and most effective method to rebalance a pack.

It should be noted that when you discharge a cell to 0V, it takes much more input than usual to truly reach 100% SoC.

 

Data

Sorry for the consecutive posts. Post more people!

Here's some very basic data showing what grid charging does to 0%/80% SoC cells.

Green represents 80% SoC cells. Red represents 0% SoC cells(or close to it). Yellow represents now voltage depressed cells. Blue are the initially red 0% SoC cells that are now normal.

Now, I don't have any data on what happens to the 0% SoC cells when you reverse them for the capacity of the 80% SoC cells. We know that it doesn't cause "bad chemical states" to form, so there is that at least.

I still contend that it can be done without damage. I think we should get away from the idea of a quick fix. We all want a magic pill that will make us better. Just like they're hard to come by for people, they're very hard to come by for our packs. There really is no quick fix, it's a process...

Just like with a stick, it should take several cycles to balance the pack.

 

....I think we should get away from the idea of a quick fix. We all want a magic pill that will make us better. Just like they're hard to come by for people, they're very hard to come by for our packs. There really is no quick fix, it's a process... Just like with a stick, it should take several cycles to balance the pack.

Balderdash! I'd rather have people doing one deep discharge and a full grid charge on full pack, NOW, than have them wait around for you to come up with the plan, the B-Line packs, etc... Sure, people with newer packs, especially a nice MaxIMA, should basically avoid this discussion entirely. But everyone else should go stick a light bulb on their packs...

Frankly, old packs can be made functional, but I'm pretty sure there's a lot more to be gained from a new pack, especially if someone figures out a decent way to make the most of what they can deliver via BCM/MCM mods. A new pack is likely worth it simply on the basis of high performance, rather than one settling for 'just functional'... I think it depends on mods though, because the stock cars don't really use all the new pack capacity... 5.9kW isn't enough. 10kW+ would be...

With mods, a new pack could easily deliver an extra 5kW, 6.71hp, +17.6 ft-lbs of torque at 2000 RPM, consistently, at-will. At ~$2000, that equals $298 per hp. A bit on the high side, as I recall, but still... That power is nothing to sneeze at considering the miniscule amount that's there to begin with...

 

but i did not see huggins talk about hydrogen gas evolution so maybe there is another part of that chapter 11 i did not read?

i just cannot imagine reversing a cell with 20A. it is hard enuff for me to let it reverse at 200mA. i do everything i can to keep it from overdischarging any cell in the stick by more than .2-.3V as measured in real time. that is why i am gonna use the 5 ohm resistor to short the high cells in the stick when the stick gets down to that point.

so where does the equilibrium come from in that phase diagram when the cell is reversed? it has to be far to the left of the point where the cell is discharged which is the bottom left end of that tie line.

so over discharge is in that region that is undefined to the left and below that spot where it reaches total discharge at 1.334V on the tie line.

 

This is a great thread with lots of good information, even though much of it is way beyond me.

I've had much success in the past with the stepped methods mentioned above, and was able to watch it on a live graph display (and monitored stick taps) at Mike's while he used his universal charger / discharger. Very easy to see the sticks dropping out, and the improvement as the process went on.

I've since modified the process to just one series of charge / discharge / charge. Also with success, and saving much time. (discharge to total pack voltage of 24 volts).

I have no equipment other than a simple charger (350 ma built by Jeff) and simple discharger (double 300 watt bulb shop light wired in series, built by Mike).

The only means of measuring success (for the ordinary car owner like myself) is the pack can sustain assist (on the mountain commute) and accept regen much better ,is more 'peppy', and I can travel many many more miles before I see symptoms of less capacity (I call it rapid depletion of soc, and more frequent forced regens and re-cals). The obd helps in this observation.

What I'm gathering so far, (Peter said it above for cars being used), the charge / discharge / charge method is likely the way to go.

Looking forward to more discussion. Thank you guys.

 

Yeah, he doesn't talk about it in that chapter, I guess since it's specifically about overcharge.

The NiMH battery also has over-charge and over-discharge reactions that allow the battery to handle abuse without adverse effects. This is called the oxygen cycle for overcharge. On over-discharge the battery has the hydrogen cycle

Introduction to NiMH Battery Technology

 

I didn't meditate on it, but I can see that your scheme makes a lot of sense for 'the ultimate' full-pack reconditioning program... It'd just be hard for Average Joe to implement... Are your "Intelligent Dischargers" intended to be made only for Bumblebee packs or any pack? I think you mentioned that once, can't remember... Maybe I'm thinking of those purple pcbs, the BCM interceptor/data logger thing...

 

They're for any pack.. Honda, Toyota, ailing or not. Just different regimens based on the condition of your pack.. Which admittedly is hard to quantify... and half of the reason why I'm doing all of this stuff.. So we can know what to recommend.

I want to actually know for sure that we're doing good things to packs with our recommendations.

 

Well, it is a little extreme I guess. It's not precisely 80%/0% at all times, or maybe even ever exactly that. Like if you took a pack out tested it after an IMA-light-causing positive recalibration, there would be a little juice in the "0%" cells.

But it's around in that range. So what happens is this hypothetical user gets an IMA light(or recalibration). They disconnect their 12V battery and do a recalibration, or the car performs one. The "80%" cells hit "actual 80%" after say 100-1000mAh of input, depending on the condition of the pack...

They get in and drive.. a couple of days, weeks, months later... Or as soon as they gas it up a hill.. the "0%" cells are really at 0% again, and the whole process starts over...

 

I picked a horrible stick to illustrate how cycling can help cure voltage depression, lol. I'll try it with one of the "Case" sticks I used to illustrate GC voltage depression.

Start
Fresh out of the car.

Charge 1
First input was only 1665mAh @ 6.5A before negative deltaV, implying "80%" cells at 70-75%.

Discharge 1
Initially dead cell drops out after 1265mAh @ 18.65A.

Charge 2
Cells are starting to be woken up. Notice long strange plateau before negative deltaV this time, with 3808mAh input at 6.5A.

Discharge 2
Hello, voltage depression! Finally able to see it, in the discharge curve of the five "80%" cells. 3615mAh output at 18.65A.

Charge 3
Things are getting better, but the stick is still a basket case. 5254mAh input @ 6.5A before negative deltaV.

Discharge 3
Plenty of voltage depression still, but improved. 4982mAh output @ 18.65A.

Charge 4
6022mAh input @ 6.5A

Discharge 4
Getting better yet, but still not anywhere near perfect. You can see cell 5's high IR is now limiting the stick with a 0.9V cell cutoff. 5568mAh output at 18.65A.

I would tend to agree with your overall assessment; basket case sticks are unable to be improved beyond about 70% with cycling alone. Cells 1,2,3 and 6 will never be brought below ~1.0V without "drastic measures".

 

Nice graphs, interesting to see cell-level curves for a progression like this. I'm actually a bit surprised the curves improve as much as they do - you're charging till first cell hits delta V at 6.5A and discharging at 18.65A till first cell hits 0.9V, correct?

I guess one thing that jumps out at me is the cell 4 curve: that almost vertical slope at the beginning of the charge, right up to the 'plateau voltage', is what you want to see after a deep discharge, what you would hope to see if the cell is going to be 'fixed' after the full charge... Your other cells? - I'm immediately thrown back in time to my first days at IC, when I first cycled my sticks... It was the worst of times...

Do you really believe you 'need' to cycle the sticks, just to avoid reversals? It seems like a lot of work and time for something with shaky outcomes, to avoid risks that, at least to me, don't seem all that grave... Maybe you should do some more tests and try to quantify any damage due to reversal at typical, low rate deep discharge currents? Hopefully mudder does a bit of that, as he outlined earlier... On the other hand, I gotta admit that I'm pretty much beyond the fear and will never go back to the 'regular cycling'... Maybe I'll take some of that more seriously if I ever get a new pack. But with the used sticks I got - I'm willing to move forward, or sideways, based on the anecdotal info here and there and my own experience with that other side...

One other thing: it seems to me that you're overcharging these cells, by like 500-1000mAh each charge... But I guess that's part of the plan, the strategy, 'overcharge balancing'?

 

Nice graphs, interesting to see cell-level curves for a progression like this. I'm actually a bit surprised the curves improve as much as they do - you're charging till first cell hits delta V at 6.5A and discharging at 18.65A till first cell hits 0.9V, correct?

That's correct.

Do you really believe you 'need' to cycle the sticks, just to avoid reversals?

Hmm. On the pack level, yes - the need for cycles would be to avoid "excessive" reversals, both in time and sheer number. On the stick level, not necessarily, since I don't tend to reverse cells there anyway with cell level monitoring. Hobby chargers will though since they only look at stick voltage.

But I do believe cycling is good, as it definitely "wakes cells up". The majority of the cells active material has been inactive for potentially a long period of time. Exercise is clearly good for packs/sticks.. up to a point.

I went ahead and cycled that above stick another 4 times. Some metrics are still improving after 8 cycles, like the stick's voltage balance, though capacity has leveled off. That means that HNi2O3 is still getting converted back into active material, even with the voltage depressed cells only hitting 1.0V. I'll post those graphs here in a bit.

It seems like a lot of work and time for something with shaky outcomes, to avoid risks that, at least to me, don't seem all that grave...

That's fair. I agree, I don't think the risks are grave. I guess from my standpoint I'm trying to find both the best and most cost effective way to do it, while minimizing any potential damage, however slight it may be.

Maybe you should do some more tests and try to quantify any damage due to reversal at typical, low rate deep discharge currents?

I would like to do that. Even with our precision equipment, minute changes like that are easily lost to variability in single or even a few tests. It would probably have to be a whole process over like a month....

1) Get a baseline test
2) Reverse a cell on a program; X number of times/day, for X mAh..
3) Test stick once a week
4) Once you've determined a point where you can see clear degradation in comparison to non-reversed cells, add up your reversals and total mAh, and calculate the average loss per event...

Easy peasy eh? I'll get right on that lol

One other thing: it seems to me that you're overcharging these cells, by like 500-1000mAh each charge... But I guess that's part of the plan, the strategy, 'overcharge balancing'?

So here, how is "overcharge" defined?

Negative deltaV, or delta temperature slope, are the two traditional methods to determine a nickel cell being "full". Now, this depends on charge current. There is no negative deltaV at say, 350mA. And the temp rise would be negligible.

So at 6.5A, the cells begin reaching full when they can no longer absorb "all 6.5A" entering them. Some starts to be turned into heat. As the cell continues to fill, more and more is converted to heat, until at some point, all of it would be. I think we all understand these concepts.

So I'm not sure if the cells are really being "overcharged" here. I mean, they are - because their charge absorption rate is being exceeded as they reach 100% SoC. Technically, the same thing would happen at 1A too, but it would be much closer to actual 100% SoC. If I stopped the charge at the first sign of plateau or temp rise, you wouldn't get the same amount of capacity out had you let the cycle complete.

Charging via negative deltaV is said to fill cells to about 90%. You need to trickle charge at low rates to effectively fill up that last 10%. So perhaps more appropriate would be a current taper when this happens. That would effectively eliminate the need for a trickle charge to reach true 100% SoC...

But.... how are you supposed to know you've reached 100% SoC, until you actually do? So I think some amount of overcharge is unavoidable.. even if only for some seconds... if you want to charge cells to 100% SoC?

If you know enough about the specific cells in question, you can use voltage. But normally you can't.

 

Painful, painful...

Here's my graphic depiction of an incremental grid charge/deep discharge strategy I think generally akin to what Eli has described. There's problems in the details, but this is just something to help visualize the process.

In general, the idea is to charge the unbalanced pack first until the most charged 'cell' reaches full, plus an amount equal to an extra 10% capacity (so plus 650mAh). This brings the lowest 'cell' up off the floor by however much it takes to charge that most-charged cell plus an additional 10%.

Then, you discharge the pack until the lowest charged cell reaches empty plus an additional 10%. I don't think Eli suggests this extra 10% of cell reversal, but rather, just goes down until the first cell reaches 0.9V. But whatever - here I'm just trying to illustrate how to minimize reversal and overcharge while getting the job done...

After that discharge, you charge the pack up again until the most-charged 'cell' reaches full plus 10%, and then discharge again until the least-charged cell reaches empty plus 10%. You keep iterating up and down until all cells have reached empty, after which the full charge should get rid of all 'yellow squares' - no more voltage depression and your cells are balanced...

The thing is, I don't see how this is feasible in the real-world. For one, if you could do it it'd take a long time. But how do you actually do it? On a real pack how are we seeing the most-charged cell, or perhaps most-charged stick or sub-pack pair? How are we seeing when the 'least-charged cell' is getting empty? This strategy requires information that the typical person wouldn't be able to obtain, while overall, the benefits, if there are any, seem like they'd be pretty marginal. It doesn't seem like a practical strategy...

Note that some of the elements are just gross illustrations. For example, I use the convention that the 'yellow squares', which represent areas of the positive electrode occupied by bad stuff, can disappear only after the cell has reached zero charge and then charged back up to the level where the yellow square is placed. There's some connection to the real-world with that, but it's cartoon-ish at best... It's hard to say how much 'bad stuff' really disappears upon discharge and subsequent charge...

The "initial state" has been changed a bit to reflect a pack that's more out of balance than the original illustration depicts; a couple cells have been pushed to the extremes of full in-car charge (here 70% with 10% out of action) and empty...