Mike,

Keeping in mind that I'm an electrical idiot,

What would this be used for? Charging the battery overnight? Could it be used to restore a weak battery?

Thanks!

 

It would be used to keep the Insight pack from going out of balance and developing P1447 errors. You would simply top up your battery every month or so.

 

grid charger

I the install is fairly simple, count me in on a group buy.

Kevin

 

What would happen if you used this every night? Would it damage the battery?

In my daily drive I go about 120 miles but over a 9 hour period. What would happen if I had a 12 deep cycle battery with an inverter powering the power supply to keep the pack up while the car is at rest between drives? I have an inverter going anyway to keep my laptop alive all day.

Thanks,

 

Battery Powered Grid Charger



That's what I have done. I use the Hybrid Automotive's Grid Charger
Grid Chargers - Hybrid Automotive - Honda Insight & Civic Hybrid
The inverter is:
AIMS PWRI18012S 180 Watt Pure Sine Wave Power Inverter, 12 Volt - Smarthome
The battery is:
EverStart 27DC-6 Marine Battery: Automotive : Walmart.com

I was reading some other posts about using a standard high voltage power supply to trickle charge the Insight's battery pack.
I think that using a simple high voltage power supply will overheat the pack, because the pack's fan is not being run. The Hybrid Automotive charger turns on the pack's fan during charge.
The inverter is a pure sine wave inverter, not a simulated sine wave, which may not work well with the 110v input of the grid charger.
The Walmart battery is 50lbs, 100 amp-hr and it takes two applications of this system to top off a relatively unbalanced battery. I use a little luggage carrier from Walmart to shuttle the deep discharge battery from my condo to the car. Condo's boards don't like extension cords.
I think this system would be useful for condo dwellers and anytime 110v is not available.
2000, Red, First Heated Air Intake Syatem

 

I would be interested in the charger as well, so please keep me in mind for a group buy.

I understand that your immediate selection in a power supply would be fixed at 350 mA, but if I could chose, my selection would be more like 200 mA or so, or at least one that could taper off at higher voltages. That way it wouldn't be quite so critical if I forget to disconnect at the proper time, or fuss with a timer to do the same.

As ogregev mentioned, my intent would be to use it once a month or so to keep the cells balanced, and maybe replenish things on an occasional heavy use day.

Jim.

 

I will talk with the TRC engineer, to see if he feels this would be safe, and then I will buy a set so I can build a prototype, and define some safe ways to tie it into the pack and bring out the charge port on a safe HV connector.
I can make battery fan connectors like I use for MIMA, that will allow plugging into the fan easily.

With the fan running, 350MA will be safe for days and would top off a pack in less than 24 hours.

We could use a simple 240V light bulb as a discharge load, and put in an automatic cycler, so you could leave it unattended for a weekend, and start off the week with a refreshed battery.
Another use would be to grid charge the insight pack with it so you could start most days with a full pack. a mini PHEV.

I just got this report from Allert in Holland: Who rebalanced a really weak pack.
See his users profile at the bottom of this page:
MIMA Users Page - MIMA Honda Insight Modified Integrated Motor Assist

"Allert said:
I just want to let you know that my battery is doing fine, I’ve made a trip to the Cote d’ Azur in France and drove almost 2000 miles.

Not one recall and better mileage than ever in the hills in France, battery capacity is what it should be and MIMA is very good in rolling hills. I did 85mpg on the whole trip and managed to get a little over 100mpg for a considerable distance from time to time."


This was after 2 cycles with 250MA


This charger will help cars that are having recalibrations frequently, and some that have an IMA code, but there will be others that will need a stronger cycling at the subpack level to bring them back. This is only for the full pack.
I will post a preliminary schematic on my website before the weekend, and do a blog on how to build it and how well it works. Then we can figure out how many we want to buy parts for.
This will be pretty simple to wire up, and if we stick with the 350MACC, no special components will be required.
A nice DIY project for the technically inclined. This supply will kill you just like the battery will, so don't try this if you are not qualified.

 

The TRC engineer does not see any reason that this will not work, so I will order a set of power supplies tomorrow and build the prototype.
The price breaks for 10, 25, 50, pieces are pretty good, so this could be quite reasonable.

 

Mike, are you going to be able to detect full and stop charging?

 

Ron,
We can do that, but that would add complexity and cost that may not be required. 350MA with the fan going will be a pretty gentle charge, and would not heat the cells or build up internal pressure for many hours after the pack is topped off. A built in digital or even analog volt meter will show when the voltage has stabilized at the end of charge, after it is stable for an hour or so, we switch to the discharge cycle with a light bulb, could be monitored with the same meter, and when the voltage drops to about 1V/cell, we would switch back to the topping charge, wait till it again stabilizes, and off you go.

If we wanted a fully automatic system, a small microcontroller could look at the terminal voltage, determine when the voltage has been stable for say 30 minutes, and automatically switch to the discharge. The number of cycles could be set, and when the cycles were complete it would turn off and set an indicator.

More components more time and more cost.
I have a lot of other projects that I need to be working on, so I am thinking that the simple totally manual system should do the job. Add a cheap programmable wall plug timer, if over charging is a concern.

Even the discharge half of the cycle may not be required.
The company does not have the RS-25-48 in stock but will be recieving 100 next week, so I will not have the parts in hand till the end of next week.

 

Ron,
We can do that, but that would add complexity and cost that may not be required.

You know why I want them. I've got to keep packs charged on the shelf.

 

I like the idea of starting the day with a full charge with a mini PHEV method!! I'm definitely interested.

 

I Like! Count me in Mike.
Wayne

 

We still need to source a nice safe male/female connector, and a 1A 200VCD in line fuse and holder.
Anyone want to look into that?
I will have the power supplies in hand first week in May.

 

a regular household connector (plug and socket) should be perfect as it goes above 166V (Peak AC) and can handle 15 amps.

 

Sounds good Mike.

My Insight was acting strangely after 3 weeks of non-use.
As mentioned before, I parked with a full SOC.
When I came back and started up, I still had full SOC showing, so the car would not background charge or take much regen. But in reality, I think the SOC was really low, as I could not get more than 1 second of assist either.

The car would NOT recal during several short trips, and disconnecting the 12v battery did not help either.

I decided to take a long trip and see what happend. After about 40 miles of highway, the car did a recal and slow recharge, and I drove 1000 more miles, and the pack is back to normal. I now get at least 10 minutes of 50% assist, and the SOC goes the entire range of all 20 bars.

I could use something safe for 3 weeks of plug time. Or, I'd set it on a daily timer to run it for 4 hours per day for my extended garage periods, or whatever the "experts" recommend.

 

Gents

I'm a little concered about the discharge cycle mentioned here. I appreciate it's not part of the project at present but it is being discussed.

Assuming we use the low current grid charger to top off an equalise the pack, then using some sort of discharger without some quite sophisticated monitoring of the pack voltage could easily lead to a reversed cell. Just discharging to say 1v per cell or (120v) assumes all the cells are of equal capacity. Which is almost certainly a false assumption.

Draining the cells with a 100w light bulb is a pretty low current discharge and if a/the weaker cell/s reach empty first they/it could be reversed before the pack reaches whatever nominal cut off voltage you/we select, as the strong cells maintain a false high pack V.

Or am I overly pesimistic?

 

My reasoning is that if the pack is brought to 100% SOC and held there for an hour or so before the first discharge, the cells should be nearly matched at 100% SOC at the start of the discharge cycle.
While there is a slight possibility that a cell could discharge to 1V much more quickly than the others, it is more likely that the full charge had already balanced the cells enough so it should not happen?

This is why the process needs to be tested carefully.

I would agree with the testing part .... but that's just me... I like to test everything.


There are some pieces missing from the 100% SoC thinking.

#1>
It assumes all 120 cells have equal Ah of capacity ... which they might not.

#2>
It assumes all 120 cells have equal internal resistance ... which they might not.

#3>
It assume all 120 cells will experience equal temperatures ... which they might not.

Assuming we use the low current grid charger to top off an equalise the pack, then using some sort of discharger without some quite sophisticated monitoring of the pack voltage could easily lead to a reversed cell.

I agree... but ...

The sophisticated monitoring you describe ... isn't extremely complex ... a bit but not crazy complex... as long as the whole process is allowed to complete from beginning to end... this is not a on the fly during operation system ... that is when it gets crazy complex.

After the NiMH pack is 100% SoC topped off with a nice gentle ~300mA or less current ... you let it rest for an hour or so for the cells to stabilize a bit... ~12 hours would be better ... but ~1 hour should be pretty good... then you discharge at a constant current... and at a slow rate ... under ~1 Amp rates ... ~300 mA would be better... don't use voltage alone as a terminating data source ... it isn't reliable enough for a 120 cell series NiMH pack... like you fear ... voltage alone greatly risks putting individual cells into voltage reversal... but if you combine the voltage with a time slope when you are under a set and known CC slow discharge rate ... you can identify when any one cell in the 120 cells series stack does its last 1V dive ... the slope at the last 5% to 10% of the NiMH battery when discharged slowly ... will drop off its last 1V pretty clearly ... if you are looking at a voltage over time plot of when a single cell goes into reversal ... you would see a sudden 1V slope drop and then the voltage would level off again as the other cells which are not discharged yet keep the voltage from dropping any more and then begin to force that first cell into voltage reversal.... the control electronics would just have to see that ~1V slope drop no matter what voltage it happens at for the whole battery pack.

---------------

The alternative is to skip the discharge part ... and just balance the pack with the slow trickle charge top off on occasion.

2) Discharge to a known 50% SOC (using voltage measurements; pick a value)

The only problem with that line of thinking is that there is not a single Voltage that one could pick that would actually = a 50% SoC state for all 120 cells simultaneously.

But other wise ... yes the idea is sound.

---------------

This jpg attached picture of a graph might make it more clear... it is of a 10 cell NiMH battery pack being discharged... Because all 10 cells are not 100% identical one of the cells ran out of Ah before the others did... you can see the ~1V sudden drop as that weakest cell ran to 0% SoC... then the other remaining 9 cells which still had Ah left ... were able to keep the total pack voltage above up... 10V... the total pack voltage continued to drop much slower after the one weakest cell got pulled to 0% SoC... then that one weakest cell begins to go into voltage reversal .. as the other remaining 9 cells force current through it... which ends up charging that weakest cell backwards... which can do permanent damage to the capacity of a NiMH cell.... the only reason cell reversal is not a death sentence for NiMH like it would be for Lead Acid or Li is that NiMH are so tollerant of abuse.... it still does damage ... but it doesn't instantly kill the battery.

This is why it isn't safe to just pick any one particular voltage ... if you had ... for a 10 cell battery pack you might have picked 10.5V ... but you still could have pulled one of the cells into reversal ... as shown.

 

My reasoning is that if the pack is brought to 100% SOC and held there for an hour or so before the first discharge, the cells should be nearly matched at 100% SOC at the start of the discharge cycle.
While there is a slight possibility that a cell could discharge to 1V much more quickly than the others, it is more likely that the full charge had already balanced the cells enough so it should not happen?

This is why the process needs to be tested carefully.

Even Jims subpack level breakout with a monitor on each tap can still let a single cell reverse if it is out of balance enough, since we do not have the ability to see each cell.

A grid charge followed by a full MIMA discharge, repeated several times may be safer as the packs built in cell reversal test are still in place?
First lets work out the charging, then we worry about the discharging.

 

Peter,

If I were to perform this cycling on my battery, I might be inclined to:

1) Charge fully to 100% and hold as Mike mentioned
2) Discharge to a known 50% SOC (using voltage measurements; pick a value)
3) Charge fully to 100% again
4) Discharge to 40% SOC
5) Repeat pattern

At some point, hopefully one could say "OK, now I can go down to 1.0 V per cell safely" and go for it.

The above would get the entire pack ready for an 'almost' complete drain cycle at some point in time. Does this sound reasonable??

Jim.

 

I would like to get in on a group buy on this also.
Thanks
Ed Mc

 

IamIan,

I suppose one *could* use the two cheap but effective tools to get the job of cycling the pack done efficiently.

Since I am not inclined to remove the pack and cycle it on a test bench with the proper tools in hand (dV/dt cycle testing equipment), I could instead:

1) top of the battery pack with the charger discussed here.

2) drive the car normally to work and home using most the battery charge by *not* letting the BCM charge it back up; effectively using the IMA disable (CalPod) switch that is installed in the car. If the battery gets too low, the BCM will let me know, and I can immediately stop the discharge process (*a).

3) after getting back home, another charge cycle can be implemented.

4) this could be done multiple times to fully exercise a limp battery pack if needed, depending on how far down it goes before the BCM decides the battery has had enough of the discharge process.

*a) - IamIan, I have witnessed this sharp cutoff of the battery pack at low SOC, by purposely discharging the pack on the way to work several times. Just as you describe in your graph, I noticed when my SOC was about 40%, it suddenly dropped to about two or three bars in under one minute. And that happened while sitting at a stop sign!! So this is something that I am at least familiar with. The BCM was smart enough to quickly start a forced regen to get things back up again.

My battery pack is actually working fairly well that removing the battery is not required at this time, and it is still under warranty. But I am attempting to understand all the mechanisms that will help me to keep the battery in good working order when the warranty has lapsed.

And batteries are definitely in our future, thus my interest in properly caring for them now.

Jim.

 

4) this could be done multiple times to fully exercise a limp battery pack if needed, depending on how far down it goes before the BCM decides the battery has had enough of the discharge process.

I agree... one could use such a strategy ... and it will give the battery pack some exercise ... specifically if done on occasion ... it could help with keep the pack balanced... but I would not say 'fully'.

part of the 'fully' process involved bringing individual cells down to as close to 0% SoC as you can without going into voltage reversal... this part is the rub / difficulty that gets worse the larger the number of series cells connected ... if it was on big single cell... it would be pretty easy to do... but with 120 cells there will always be 1 cell that is the weakest ... and that 1 cell will always go into voltage reversal before you get the rest of the 119 cells down to 0% SoC.

This is the distinction between balancing a pack ... which one can do just by topping it off at the top... vs the very different process of re-conditioning a pack... if you aren't trying to re-condition the pack... then the discharge cycle can be skipped.

Besides the BCM tries to avoid the pack going bellow 20% SoC.

*a) - IamIan, I have witnessed this sharp cutoff of the battery pack at low SOC, by purposely discharging the pack on the way to work several times. Just as you describe in your graph, I noticed when my SOC was about 40%, it suddenly dropped to about two or three bars in under one minute. And that happened while sitting at a stop sign!! So this is something that I am at least familiar with. The BCM was smart enough to quickly start a forced regen to get things back up again.

Don't make the mistake of thinking the dash SoC gauge is accurate.

At best... it gives you a vague idea of what the BCM is currently willing to give you from the battery... this is not = to what the battery is for SoC ... it is not even what the BCM thinks the battery is.

My battery pack is actually working fairly well that removing the battery is not required at this time, and it is still under warranty. But I am attempting to understand all the mechanisms that will help me to keep the battery in good working order when the warranty has lapsed.

good plan... and I am in a similar boat.

I got a IMA light and error code about ~7 months ago... when I was at 120,000 miles and ~8 years old ... Honda techs wanted to replace the battery module under warranty ... They didn't even say a thing about my MIMA system ... but others might not be soo lucky.

Instead of using the warranty ... I opened up the battery module... I rebalanced my pack... logged some data along the way ... and have been driving for the last ~7 months without issue... car has returned to normal pre-IMA-error operation.

Fixing the balance is only treating the symptom ... and the root cause that created the imbalance is still there ... so I expect the pack to get out of balance again eventually... but I have been curious how long it will take ... and I also eventually want to re-condition my OEM pack.

And batteries are definitely in our future, thus my interest in properly caring for them now.

eventually ... perhaps ...

I figure I shouldn't have much issue using the OEM original pack out to 15+ years or so... as long as I am willing to do the occasional once or twice a year maintenance on it... by fixing the balance and or reconditioning it.

 

Does anyone know why this happened?

I started to recondition a profoundly dead pack, and found that my chargers were detecting delta V immediately. They wouldn't charge at all. I had to hit the cells with a 12 amp charge for about 2000Mah before the main chargers were able to function. This is an HCH pack and is yellow. The sticks are coming up to 7300Mah charge / 6500Mah discharge after only a few cycles, so they are good. IR is 55-60 Mohms (normal).

 

It is one of the known and common errors in many 'smart' NiMH chargers.

One of the charge termination methods the 'smart' chargers look for the change in the voltage vs time slope... when you first start charging a Battery there is a sudden rise in voltage under the charging current... this sudden rise in voltage then tapers off ... sometimes a 'smart' charger confuses this change in the voltage vs time slope with the change that happens at the top of the SoC slope when the battery is full and the additional charging energy just gets converted to heat... and so it incorrectly thinks that the battery full , even though it is not.

The frequency of this will vary from 'smart' charger to 'smart' charger... and from battery to battery even on the same 'smart' charger.

The other well known and common error for 'smart' chargers comes from the rate of voltage rise vs time slope ... if you have a 100Ah battery it will rise in voltage slower under 2Amp rates of charging than a 1Ah battery will... the 'smart' charger again incorrectly thinks that the battery has reached its peak voltage condition when it hasn't.

In short ... 'smart' chargers ... are very dumb compared to a human ... just because the charger says it is full does not mean it is full yet ... and it does not mean that the 'smart' charger didn't over charge the battery... but are a simple method to try and automate the process so you as a human do not have to just sit there.

similarly a 'smart' charger that does cycles ... generally 95% of the time would miss the voltage reversal event int he picture I posted above... as 95% of them discharge to a set Voltage level.

 

I have been fabricating a trailer for the Insight on a stand and the other demos, so I have not been on line and following this thread. I see that our battery experts have added a lot of information to consider.
The stock system:
The stock BCM only charges to 80%, never charges to 100%, and stops discharging at 20%. This has not been confirmed, but has been accepted as gospel, and is the cause of battery pack imbalance.

Charging:
The proposed charger will not stop charging when one or 10 or all cells are full, and will just keep 350MA going into the pack forever, which will fully charge all of the cells, no matter what SOC they started with.This gives us a reliable 100% SOC.

The "smart chargers" that have been used to recondition subpacks try to determine 100% SOC based on the slight dip in voltage that happens when a cell reaches 100% and starts to turn charge current into heat. On the first cycle, this dip will be difficult to detect accurately, since we are looking at 6 or more cells, and some of the cells may be at quite different SOC than the rest, so the dip may be masked by an adjacent cell in the same subpack that is not quite at the 100% SOC, so it may still be rising, while others may be fully charged and into the dip.
That means that the 100% SOC point for all cells may not be reached with the "Smart chargers"
If the proposed 350MA charge is allowed to continue, until the voltage stabilizes for at least 1/2 hour, we will know with a high degree of confidence that all cells are fully charged.

I suspect that some of the AH discrepancies that the "smart chargers" are seeing are due to this premature 100% SOC determinimg effect.

Discharge:
The danger of reversing a cell during discharge is a real issue that we need to avoid. Based on some simple experiments, I think that we are pretty safe for 90% or more of the packs just using a light bulb and voltmeter and stopping the discharge at 120V(1V/Cell), after topping off the charge on the whole pack. The key is that we are bringing all cells to a reliable 100% SOC.

The "Smart Chargers" look for 1V per cell as the zero SOC point, and would likely not see a dead or reversed cell, and could terminate the discharge prematurely, skewing the AH determination.

There may be a pack that has a subpack with a cell that is far below the average AH, and this will damage the cell. To do this so that cell damage possibility is minimized, we would need to monitor the ten subpack taps and look for a rapid drop off of a cell like the car does in normal operation. The car may be the safest discharge method we have, since it already does the monitoring at the 12 cell level.
We can speculate on the best procedures, but at this point, we can get the charge part nailed down, and then work on the discharge, and prove the system with carefully run experiments.

 

The stock BCM only charges to 80%, never charges to 100%, and stops discharging at 20%. This has not been confirmed, but has been accepted as gospel, and is the cause of battery pack imbalance.

Agreed... it is an assumption.
Eventually it is something I plan to verify... one of the many things on the list

But, until it has been verified it is a known method of extending the service life of NiMH cells and Honda said they are using it... so it is a reasonable assumption... but I'll 100% agree... eventually having some real data to know one way or the other, would be preferred.

I agree, we do not yet have confirmation / hard data to know the % probability distribution among the possible IMA error issue causes... so we are making a few assumptions until we have more data about the distribution among different causes... but we can put our best effort forward ... make some educated guesses... do some testing ... and just do the best we can one step at a time.

 

If we use a 100w load/light bulb to discharge the cells this is a current drain of about 700ma at 144v (120 x 1.2v).

Assuming a 6.5ah pack it would take about 9.25hrs to discharge to zero capacity.

I agree 1v per cell is a good target for the discharge cut off, but the problem is if we charge the pack to it's theoretical 6.5ah capacity using our low current charger to 100% we are still assuming cells are all equal capacity in an old pack.

So as the pack discharges if one of the cells is down only .5ah in capacity it will reach or go below 1v 45 minutes before all the others.

As the others will maintain a terminal voltage of about 1.1v during the discharge period x 119 = 130v the drop of one cell to zero will be effectively masked using a simple 120v as a cutoff point. The weak cell will then be reversed and probably damaged.

The discharger IMO must as Ian has mentioned be looking for any sudden change in voltage of say .5 to 1v rather than the gradual decline towards 1v per cell. This would indicate when the weakest cell has reached empty and that should be the cut off point.

I must admit I'm starting to like the idea of a full pack charger and discharger rather than the design I had been working on where individual subpacks were cycled.

Certainly for me the dummy load and the high voltage supply are not an issuse, and the monitoring of the standard cell taps with a plug in device polling them every second or so which controls the charger/discharger looks quite appealing now. It would certainly be simpler. It would be easier to detect a cell going low at the standard cell tap level as well.

Thinking cap on again.

 

The discharger IMO must as Ian has mentioned be looking for any sudden change in voltage of say .5 to 1v rather than the gradual decline towards 1v per cell. This would indicate when the weakest cell has reached empty and that should be the cut off point.

But isn't it too late then? Can you react fast enough to keep from damaging it? Or has damage already been done?

 

IamIan,

Don't make the mistake of thinking the dash SoC gauge is accurate.

At best... it gives you a vague idea of what the BCM is currently willing to give you from the battery... this is not = to what the battery is for SoC ... it is not even what the BCM thinks the battery is.

I understand what you are saying about the BCM SOC level not being the most accurate, and my purpose in draining the battery was not to check the SOC level in which it happened, but to purposely 'exercise' the battery after reading much about it on this site.

Mike,

The car may be the safest discharge method we have, since it already does the monitoring at the 12 cell level.

I like what you said above, as that gives me a good option to cycle the battery and maintain it over the long haul.

1) Purchase the 'top off' charger being discussed here.
2) Effectively use the IMA disable switch during normal driving to work to purposely drain the battery when accelerating and/or going up hills, then flipping the disable switch to keep any regen from happening until the car is home, and can be charged fully again.

This could easily be arranged to happen on a weekend, such that the car can be monitored for voltage level and the charger shutoff after getting to let's say 170 VDC or so and hold for 30 minutes, depending on what the consensus is here.

Jim.

 

On the stock BCM, the 10 taps are required because of the widely varied charge and discharge rates that can be encountered. Changing from no assist to full assist can instantly drop the pack terminal voltage by 20 volts of more, so seeing a cell drop out needs the 14V tap to reliably see a cell drop.

The big cost item for determining that a cell has started to drop out, is the 10 voltage taps and the electronics and isolation required to read them.

With a known and repeatable discharge current, I expect that we may be able to simply look at the whole pack voltage with enough resolution to see a cell drop out.

If we looked at the whole pack voltage during discharge with a 12 Bit or higher resolution A/D, and we looked at the rate of change in voltage, we should be able to reliably see this cell drop and could open a relay to stop the discharge.

We are also making some assumptions about the different failures modes.

If the pack is simply unbalanced with a cell or more that have say 50% SOC when the rest are at 80%, simply charging them all to 100% may solve the problem.

If the reason some of the cells are lower than the rest is internal leakage, the same cells will continue to fall behind, but a monthly full charge may keep them in the zone, where they never cause a recal or IMA code?

If AH capacity on some cells gets reduced with temperature and aging, this would also cause a recurring unbalanced condition, which may or may not be helped with a simple charge to 100%.
While it is the best technique we have, I suspect that depending on a "smart charger" to determine AH and cycle termination may be misleading?
how about a discharge experiment using a known balanced pack with a planted subpack with a cell that has been purposely discharged to 50% of the rest of the cells.
We could data log the discharge voltage during the lamp or resistor discharge, and see the signal signature of a dropped cell, and we could then determine if it is feasable to reliabilly detect the single cell drop out while monitoring the whole pack.

 

How about a discharge experiment using a known balanced pack with a planted subpack with a cell that has been purposely discharged to 50% of the rest of the cells.
We could data log the discharge voltage during the lamp or resistor discharge, and see the signal signature of a dropped cell, and we could then determine if it is feasable to reliabilly detect the single cell drop out while monitoring the whole pack.

Good idea. In fact we could try that with a single subpack on the bench which would be safer just to get a feel of the characteristics. We could certainly charge it up to 100% and then by making a small hole in shrink wrap gain access to one single cell from the stick. Discharge it a bit and then do the discharge test on that one stick. Noting voltage drop.

The 10 bit adc on the picaxe chips is 0-1023 over a 5v range which over a range of 0-200v with a potential divider gives a resolution of 5 per Volt. That might be enough to detect it?

By monitoring at the subpack points it would give 50 per volt and also some info on where the bad subpack was in the pack. Mike you did some of this with a relay monitor a few years back?

A 16 bit adc would give over 300points per volt over 0-200v

 

I did do some battery monitoring with a labview based DAQ system, but that was before we had a lot of information about rebalancing and pack failures.
I still have the pogo contact system I made at the time,that allowed clipping on the test harness, and may still have the Cad drawing so more could be made.
The clicking relays were noisy and I was not having any battery issues (still no recals after 146K), so I dismantled the electronics and used the relays for the first MIMA systems.
I still have the 16 bit DAQ card, so I could set up a single monitor for a whole pack, just don't have the time this week, as I need to get a custom trailer built and painted before next weekend.
Ron,
I realize that at this point, you have cycled more subpacks than anyone, and have data that leads you to your conclusions, but your data is generated with a smart charger that if used on a single cell, would yield exact and reliable data, but you do not cycle on a single cell basis, therefore charge termination on the top end is a bit blury, and as more cells and higher currents are used, the fuzzyness increases.

At 250-350MA with the fan running, the heating will be negligible, and the slight overcharge condition will bring even a weak cell to 100% SOC. This procedure has been used in multiple battery systems for many years.
Any EV user will do equalization charges on their pack several times a year.
Most simple NIMH or NICAD chargers will use the same technique, just put a safe current on the pack, and wait so many hours to assure the pack is balanced, and you are ready for another cycle.
As Peter suggest, a 16bit or even a 24bit A/d converter should easily see a rapid cell dropoff, and can stop the discharge before the cell actually reverses.
The advantage we have over the stock BCM with a charge/discharge system that is run with the car off, is that we should have minimal noise, and the signal should be clean and quiet.
I won't get the chargers until the end of the month, and if I can clear the decks then, I will take my spare not so healthy pack and run some test. I suspect that Peter will do the same, and then we will see where this gets us.
Allert's pack in Holland came back and is running great, as is at least 4-6 other people that have followed the whole pack charge technique using a variac.
Since there are more than one failure modes, I expect that the more comprehensive cycling and testing that you do will prove to be required once a pack has gone too far, but IMHO I think this simpler balancing system needs to be developed to possibily prevent packs from getting that bad.
To give credit where due,the idea of using small power supplies in series to make a CC charger was built and is still used by Armin Kusig to top off his pack each night. He also has a bank of 10 DVM's watching the taps. He started this procedure at least a year or so ago and as of the last time I communicated with him the pack is going strong. He puts a timer on the charger, and sets the time based on his SOC when he gets home, so he was the first mini PHEV Insight.

 

At 250-350MA with the fan running, the heating will be negligible, and the slight overcharge condition will bring even a weak cell to 100% SOC. This procedure has been used in multiple battery systems for many years.

You can make any system foolproof, but you can't make it damnfoolproof.

As a programmer, I learned long ago to protect the user from themselves. What would be the ramifications of someone starting this charger going with a full battery (perhaps a power blip would reset the timer)? What if they leave it on charge overnight and the power keeps blipping due to a storm?

What if they put it on charge Friday night and forget to take it off charge all weekend?


BTW, miscellaneous bit of data for you - running the fan at 6V keeps the battery cool even while four sticks are simultaneously running through 10 amp cycles (10 amp charge and 10 amp discharge). An HCH pack, which has no fan, will overheat if a single stick is charged at 10 amps. I'm trying to adapt an Insight fan and duct to the HCH to solve this problem (I have two of them arriving tomorrow). If I can't, I'll have to gut Insight packs and slide in the HCH sticks just for cooling.

Since there are more than one failure modes, I expect that the more comprehensive cycling and testing that you do will prove to be required once a pack has gone too far, but IMHO I think this simpler balancing system needs to be developed to possibily prevent packs from getting that bad.

I'm in full agreement with that. From the packs I've seen (which are all terminal cases), I think this charger would eliminate most of the problems and perhaps fix some of the worse pack problems (after a few cycles) - I don't think the packs can go so far that this wouldn't fix them (but if the pack is that bad, then it will need this maintenance - periodically - forever).

I'm all for it.

 

I guess I need to be more specific about the timer I had in mind. The clockwork type that are used to cycle lights when away from home, or to time your Christmas tree lights, cannot get reset accidentally, and if the power goes off, the timer stops where it is as does the charger, only to pick up right where they left off when the power returns.

The extra 12V power supply will run the battery fan at nearly full speed, so even if left on over a weekend after fully charged, I expect that no significant heating or cell damage would be done.
This is the reason to go with a Constant Current power supply.
Another aspect of using adjustable constant voltage supplies in series with the constant current supply is that the three constant voltage supplies are adjustable so the max charge voltage that the system can reach can be adjusted.
All of this will be looked at carefully as part of the initial prototype testing.
Now I need to get out and finish the trailer.

 

Another aspect of using adjustable constant voltage supplies in series with the constant current supply is that the three constant voltage supplies are adjustable so the max charge voltage that the system can reach can be adjusted.

First page.

 

I have not been reading this thread and just discovered it so I had to go back and read the entire thread. So some of these comments will refer to earlier posts in the thread.

1) I have added 20 point monitoring to two of my packs so that I can read the voltage on every stick. It has been very useful.

2) I have discovered some "recovery" techniques that can be applied to a failing battery to make it useful.

a) First I have to determine at what level the battery does a recal. recently on my Lazarus battery it started recaling at 13 bars after almost 5 months of trouble free operation after a 5 cycle discharge to 5.4V per stick Charge to full charge (seen when the voltage curve reverses at the top) which is about 8.4V per stick.

b) with that information you then watch the SOC and when it drops near that value, turn on the headlights.

c) If you get an IMA light there is a way to clear it and continue...if the problem is a weak stick....just remember to turn on the headlights BEFORE starting the car and the IMA light will not activate. The reason for this is that, in my case, the one stick was dropping to 0Volts when the IMA battery was used to start the car and then before it could begin getting charge current to bring it back up, the computer detected it and lit the IMA light. BUT by turning on the headlights first, as soon as the engine began to develop power it was also charging the battery so the weak stick came immediately back up before the computer could detect that it was low.

d) If it really gets finicky, just leave the lights on anytime the ignition is on, in fact I was thinking about wiring it like my Caddy where the lights are always on.

I was able to keep going for a month without ANY IMA lights and only one RECAL (when I wasnt watching close enough to turn on the headlights in time) Even though this battery...that I am now testing...has one completely DEAD stick and two that have extremely low capacity, less than 1Ah (0.4Ah, 1.1Ah, and 0.0 Ah) and as you may recall this was given to me with less than 1 volt on every stick. That is why I call it the Lazarus pack. I could have taken it further (yesterday it was still at 19 bars with no IMA light or RECALS), but my refurbished pack was ready so I am bow going to swap them out and rebuild this one. These sticks were verified today while discharging, and suspected for 5 months because I have the ability to read all 20 sticks individually.

I am currently discharging it, and except as noted above on the three sticks, all the rest are perfectly balanced, just as they were when it returned from the dead 6 months ago.

 

As to safe idle charge to prevent heating:

Mike discovered the reverse voltage curve of the sticks when they achieve full charge.

I have found that 140 ma never shows the reverse curve of a full stick. But at 240ma that reverse curve of the peak voltage shows itself when the stick is fully charged. That reverse curve is explained by the fact that as the stick gets to full charge it begins to dissipate the excess energy as heat which reduces the internal resistance of the stick causing a voltage dip for the constant current input.

This would lead ME to believe that the current of 240 ma DOES cause heating and that 140 ma does NOT cause heating and that the point where heating is first generated is somewhere between 140ma and 240 ma So I would aim for a lower number than 350ma for down here in the south where our night time temps are often over 80F. It might be safe in a cooler climate, but why risk it?

 

That reverse curve is explained by the fact that as the stick gets to full charge it begins to dissipate the excess energy as heat which reduces the internal resistance of the stick causing a voltage dip for the constant current input.

a bit more detail.

The -dV at peak SoC... is a result of the chemical reaction in the battery cell changing from one type of reaction to another... in the first chemical reaction you are storing energy... once at capacity... that chemical reaction ends and a different one starts... the 2nd chemical reaction does not store energy so all the energy is converted to heat and pressure ... if you look at the temperature log of SoC vs time during a CC charge ... you see the temperature starts to curve up before the -dV ... this happens because even in 1 individual cell... not all molecules inside that cell reach the same point at the same time... so as more and more % of a cell and more and more % of cells begin to change over from the 1st chemical reaction to the 2nd you get more and more of this reaction... thus under a constant current charge rate ... the curving up temperature vs time slope ... then Peak Voltage ... then the dV ... then the -dV ... then the Constant V.