How closely to match NiMh cells in a pack?

[BatteryCharger]

So I have around 600 AA NiMh cells. Yes, 600. I am attempting to build them into 12v packs, and then parallel each set of 12v until I run out of batteries.

Yes, I am insane. Even worse I'm cycling and testing ALL of these batteries using ONE 4 bay BC-900. I'm starting to think I can't even keep up with the self discharge rates....

Anyway, my question is, when building a 12v NiMh pack, how closely does each cell need to be matched? I have *SO* many different cells to choose from, I could probably find sets of 10 that match exactly, but then again, the BC-900 testing results can vary by a few dozen mAh on differnet cycles of the same battery, so it's not a precise number.

How closely do the cells need to be matched? Is keeping them in groups of 2.4xAh, 2.5xAh, 2.6xAh, and 2.7xAh good enough? If I find a battery at less than 2.4Ah I consider it bad...over 2.7, the test is wrong.

 

[45/70]

Generally NiMH cells in a pack should be within 5% capacity, or preferably better. That makes your strategy pretty good. I would try to do better than 5% though, if you can, because the closer the cells match, the longer the packs will last.

Dave

 

[LuxLuthor]

It depends on your intended purpose, how you are going to join them, brand of cells, amp draw, and how you plan on charging them once in a pack, since NiMH in general should not be charged in parallel.

 

[McAllan]

NiMH in general should not be charged in parallel.

Not in general. Should never
That's because then when NiMH charges and they're full the voltage drop a (very) little. This an intelligent charger detects.
Even though the cells might be matched when they're new it's no guarantee that they'll be matched when used say half a year.

Earlier I used to match even my AA and took great care to use them in their respective pairs as best as possible. But I soon realized that the batteries which matched perfectly as new didn't match very well after some use. The batteries which exhibited the highest capacities in many cases matched better with the ones that originally had the lowest which they still kept. The middle ones tends to still match with the middle ones.
So now I just use those together that was bought at the same time of same brand/model.

 

[BatteryCharger]

Well, ideally I'd like to arrange them in 25Ah, 12v packs. Roughly 10 parallel 12v strings. Each of these would be charged individually, and then tied together for use. Basically I'm trying to replace a 50 pound deep cycle lead acid battery with something smaller and lighter for general 12v use. (I know something like F cells would have made more sense, but I got a killer deal on the AAs)

I plan on building each individual pack so that each cell can be removed, tested, and replaced if necessary in the future.

 

[LuxLuthor]

Not in general. Should never

There are ways to manage the issue, so I don't say "never."

BC - So you are going to individually charge ten strands of 10s1p 2500mAh AA NiMH cells every time they run out, then reassemble back to being 10s10p ? And what is the peak current draw on this?

In general, I think being within 0.05 volts before connecting in parallel...but it's hard to see how this will work out on a practical basis.

 

[BatteryCharger]

BC - So you are going to individually charge ten strands of 10s1p 2500mAh AA NiMH cells every time they run out, then reassemble back to being 10s10p ?

No - I want to make a 100 cell pack (10x10) which should give me about 25Ah, and then make about six more of those and charge them individually/connect before use. I don't neccessarily have to connect all six if I don't need that much power....

Maybe not the absolute best thing for the batteries, but, that's why I'm making it easy to remove an indivdual 12v string and an individual battery from that. (not soldered/heat shrunk together)

Peak draw will be...well, whatever I can come up with. Probably at least 2C.

 

[uk_caver]

If you're charging that many cells for break-in, I'd be tempted to get some multicell AA holders and do the charging as long series strings. That way you could just use the BC-900 for discharge.

Since break-in charging is typically 14-16hrs at 0.1C, a cheap and easy to make constant current circuit run off the appropriate power supply could get a lot of cells done. Even charging a couple of strings (of 6) at a time, each string with its own CC circuit, off an old ~12V/1A wall wart would generate charged cells about as fast as the BC-900 could discharge them, just doing one overnight run per day.

You could discharge at 500mA, and as long as you pulled the cells before they started charging again, you might get 3 runs/day, and another one overnight with charging afterwards.
It's a bit of a pain that the BC900 always does the immediate recharge after discharge.

 

[Mr Happy]

No - I want to make a 100 cell pack (10x10) which should give me about 25Ah, and then make about six more of those and charge them individually/connect before use. I don't neccessarily have to connect all six if I don't need that much power....

But that does imply you will be charging 10 strings in parallel? And as has already been mentioned, you cannot effectively charge NiMH cells in parallel strings without the risk of damage.

 

[LuxLuthor]

But that does imply you will be charging 10 strings in parallel? And as has already been mentioned, you cannot effectively charge NiMH cells in parallel strings without the risk of damage.

Exactly what I have been trying to figure out your specific plans, and I did understand and say 10s10p, but the only workable way is to have a single strand of 10s or you could do 20s or 30s (and each string would technically be "1p" as in 10s1p, 20s1p, 30s1p) with the right charger, but then AFTERWARDS put them in whatever parallel setup. If you are using shrink as a tube, that's one thing...but how you have explained it so far, it's still not clear.

Again, if you are going to individually charge a single string of 10s cells, then the next string of 10s, until you have all 10 strings of 10 charged...and then combine them into 10s10p for discharge, that will work. But once discharged, you have to take it all apart and charge strings with cells in series only.

 

[BatteryCharger]

you cannot effectively charge NiMH cells in parallel strings without the risk of damage.

CAN'T or SHOULDN'T? I was thinking more along the lines of a "dumb" charger, probably not charged to 100% capacity.

Let's say I had 10 parallel batteries for 25AH. Couldn't I put it on a ~5 amp charge and stop it at about 1.3v/cell? They won't be 100% full but they wouldn't be overcharged either, would they? (especially since I'm spending 25 years to match the things!)

 

[45/70]

CAN'T or SHOULDN'T? I was thinking more along the lines of a "dumb" charger, probably not charged to 100% capacity.

Can't. Even at a 0.1C, or less charge rate. You could still damage the cells, because the current flow between cells is uneven with nickel chemistry cells charged in parallel. Mr H straightened me out about this recently, as I thought it would be possible at a slow rate, as well.

Dave

 

[VidPro]

can but 10Parellel would be rediculous trying to rebalance charge it
we do 10Sx2P here all the time , but you must keep the slow final topping rate at or BELOW 1/10c for a SINGLE cell. so basically a safe balance final topping area for a 10P x 10S would be ~1/100C (or something) , like thats gonna happen.

you can potentially low charge it to say 13.8v, but you would have to know for a FACT today and tomorrow that no cell started into V-drop on you, especially if it is going way out of balance, like from self discharge, or unbalanced charging rituals.
we do this all the time, and i have failed miserbly doing it once, when i didnt know better, you cannot terminate using a v-drop, but heck V-drop termination on large sets of series cells is not reliable Anyways.



i will demonstrate SERIES V-drop TERMINATION FAILURE in askii pics, if you want to see a real demonstration come on over

---1.20v--1.20v--1.20v--1.20v--1.20v--1.20v--1.20v--1.20v--1.20v--1.20v-- =12v
ok here is your 10x series nightmare, its magic they are all at the same voltage , they live in battery heaven.

Here is the pack when some cells are almost fully charged (well the idea anyways voltages will oft be way higher)
---1.40v--1.40v--1.41v--1.38v--1.40v--1.40v--1.40v--1.40v--1.39v--1.40v-- = 13.98 <--- The brilliant computer is watching this very closely

here is your stupid computerised battery charger missing the fact that one cell has V-dropped already
---1.40v--1.40v--1.40v--1.39v--1.40v--1.40v--1.40v--1.40v--1.39v--1.40v-- = 13.98 <--- so the brilliant (stupid) computer watching this very closely didnt see a THING.

ok so it never happens exactally like that, but your average v-drop computer doesnt realise that single items in a long seires pack have V-dropped till it either dropps hard, or more than one piece v-drops when you have normal variations. i have measured the cells in a series pack individually (the hard way) and have seen this happen regular.

So i start out with the idea of , if i really want to balance this pack when its only ONE little series set and isnt even parellel yet, that i should Finish up the charge at 1/10c and not depend on a charger that only tries to see an elusive non-existing V-drop, because reality happens.

So how is that computerised charger doing now?
---1.41v--1.40v--1.37v--1.41v--1.41v--1.39v--1.41v--1.41v--1.39v--1.41v-- = 14.01 <--- the computer is about ready to ketch on to what is happening , give it time, it will get there. in the mean time that first peaked battery is screaming.

---1.42v--1.42v--1.33v--1.40v--1.41v--1.39v--1.41v--1.39v--1.41v--1.40v-- = 13.98 <--- Okay, arent you glad the computer was paying attention there doing all your work for you. now that one cell has been tortured and 2 have an obvious drop by now. the computer is (as always) a miserable failure , unless of course you got one of them Fancy computers that slowed down Long ago, when the voltages peaked, and you know how much those cost

 

[VidPro]

so instead we toss out the computer , Fast charge this, with a SIMPLE voltage limitation,
when the voltage is low , the current is high. the needs of the battery are high.
when the voltage is high the current is low, and the needs of the battery are low.
what could be simpler?
Unrealistic Voltage Max Charging scenario

---1.20v--1.20v--1.20v--1.20v--1.20v--1.20v--1.20v--1.20v--1.20v--1.20v-- =12v <--- Pound in 14.00V into this , and it is charging around 3-6 Amps (ok so my stuff doesnt go faster than 6 so it was still current limited).

---1.40v--1.40v--1.40v--1.40v--1.40v--1.40v--1.40v--1.40v--1.40v--1.40v-- =14v <---Pound in 14.00V into this , and it is charging is ZERO (big difference now over the stupid computer huh)

---1.20v--1.20v--1.20v--1.20v--1.20v--1.20v--1.20v--1.20v--1.20v--1.20v-- =12v <---Pound in 14.00V into this , and LIMIT the current to 1/10c and its charging will never go over 1/10c no good cell item will V-drop under this condiition. charging take forever.

---1.40v--1.40v--1.40v--1.40v--1.40v--1.40v--1.40v--1.40v--1.40v--1.40v-- =14v <---Pound in 14.00V into this , and LIMIT the current to 1/10c and it is charging is ZERO still. (lets see the current controlled computer pull that one off)

but of course this is total BS because it Never happens like that, so now you gotta put A and B together , and assume C & D in the process TOO.

 

[VidPro]

so lets go back to our normally unbalanced schedule and see where were at in reality.
Voltage charging FAILURE

---1.41v--1.40v--1.43v--1.41v--1.41v--1.39v--1.37v--1.41v--1.39v--1.41v-- = 14.03 <---lets say were pumping voltage controlled 14.10v into this , we can expect that it will be self limiting, depending on your wiring and regulation and all this might be reading about 400ma of charge due to the minor voltage differential (.07v) that exists. but then again we could limit current to insure that it never causes a V-drop.

---1.41v--1.40v--1.43v--1.41v--1.41v--1.39v--1.39v--1.41v--1.39v--1.41v-- = 14.05 <-- still 14.10v in, and thanks to the voltage differential being lower (.05v) it has now gone to below the overcharge rate for the cells, and chances are fair that nothing will have v-dropped by then. but you took chances.

IF you had started a V-drop Cascade
---1.37v--1.40v--1.43v--1.30v--1.41v--1.40v--1.40v--1.39v--1.39v--1.41v-- = 13.90 <-- still 14.10v in. the voltage drops cause the huge current rise due to the higher voltage differential (.2v) and the current Keeps increasing , more batts V-drop current keeps rising ,making things worse and worse, depending on the input current the worst battery explodes. (ok it was much more fun than it sounds)
BECAUSE
we had the voltage at a risky place, and didnt have current limitations to below the possibility for a V-drop or about 1/10c.

generally i can get away with around 14.1V voltage control, and it all does exactally what it is supposed to do, but out of fear and change and unpredictability i weasil out and put on a current control of 1/10c for any SINGLE battery in the set, that way nothing will V-drop like this.

one more added factor here, if you use only a lower voltage charge, how will you get full re-Balance. Series Ni-mhy balancing is done via slow overcharge untill all the cells are peaked out. untill all cell items are being slightly overcharged (within spec) you wont have every cell in the series set balanced to thier top possible charge, and out of balance would not be as good.

 

[VidPro]

now we will begin the Parellel nightmare, stupid v-drop terminations , good for nothing i tell ya. so lets start small, because this is already hurting my eyes :wave:
Parellel V-drop Insanity

|---1.41---|
|---1.39---| <--- lower voltage , hey ? soo higher voltage differential = higher current to this one.
|---1.40---| no problem because it isnt overcharging . . . yet.

ok so here we have a happy parellel set charging up of 1S 3P, what could go wrong? they are tied together so the voltage is the SAME right? well sort of, but for our demonstration here i will be providing the batteries virtual voltage when not connected to the set.

|---1.43---|
|---1.40---|<-- so were pumping in a current control of 1/10c for all 3 batteries together naughty naughty.
|---1.41---| 1/10c for all 3 is 1/3.3 C for any ONE enough to incurr a V-drop in ONE.


|---1.39---| <--- oh oh, did you see that? it V-dropped
|---1.41---| now its overcharging , and , higher voltage differential = higher current to the CHARGED one oh no.
|---1.42---|

but it gets worse

|---1.35---| <---this lower voltage (v-dropped) battery is now taking on most of the charge current
|---1.40---| <--AND These batteries are also pouring thier current into the overcharged battery
|---1.41---| being tortured it v-drops More and more

the v-dropped battery keeps going down in voltage, the charged batteries all pulverise the one that is V-dropping, and as long as the charge current keeps going in, this battery will be baked boiled and fried.
and the V-drop termination charger didnt see any reason to terminate EITHER.

this would have worked if no battery in the set could have v-dropped, so we charge a set like this at 1/10c for any single cell item in the set, or 1/30c for the set.

 

[VidPro]

ok so we KNOW what happens, WHY did we do such stupid things, and why do computers do such stupid things , hmmm could be humans programmed em :mecry:

so lets go back and do this right, we can do this 3 ways.
control the voltage (assume that we KNOW all the voltage peaks today and tomorrow), Or Control the current , so not ONE battery can V-drop, Or control both the current and the voltage.

Proper Parellel charging


|---1.40---|
|---1.39---| <---put in 1.40V
|---1.40---|


|---1.40---|
|---1.40---| <---put in 1.40V , eventually you get this (if batteries cooperate)
|---1.40---|

but were scared, so take it down a notch, it will still charge eventually

|---1.38---|
|---1.38---| <---put in 1.38V , eventually you get this , and if you didnt buy a $5 meter you have a small safety factor
|---1.38---|


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


|---1.41---|
|---1.39---| <---control current to 1/10 c for SINGLE cell, or 1/30 C for all 3.
|---1.40---|

|---1.43---|
|---1.42---| <---control current 1/30 C for all 3. no cell will v-drop.
|---1.44---|

with no single component of the parellel set going beyond the 1/10c spec charge rate, no cell will v-drop, current will stay the same, it will slowly overcharge within spec for overcharge.


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


|---1.40---|
|---1.39---| <---put in 1.40V with current control to 1/30c
|---1.40---|


|---1.40---|
|---1.40---| <---put in 1.40V ,with current control to 1/30c, eventually you get this everytime
|---1.40---|

Even if the voltage is to high and a battery could have v-dropped it WONT, because the current was to low to cause a V-drop condition, application of both the current control, and the voltage control nothing can go wrong . . .. well you know as long as you know what your batteries peak at and what thier max and mins are and you put in a little good measure safety and check to see that they arent getting crasy over time.

 

[VidPro]

now the final Hard part, do BOTH, and keep everything balanced out perfectally so none of the failure scenarios can occur.
You can just tell the current only controlled V-drop termination computer to shove off, because it can fail in series AND in parellel, it just cant cut it without causing damage to the "matched" sets and mis-matching them even worse than before, , get windows 7 for it see if that helps

|---1.40v--1.40v--1.41v--1.38v--1.40v--1.40v--1.40v--1.40v--1.39v--1.40v--| = 13.98
|---1.39v--1.40v--1.40v--1.39v--1.40v--1.40v--1.39v--1.40v--1.40v--1.41v--| = 13.98

each of the cell items could be anyfreakingwhere in voltage including 1.0V (or lower) when you have self discharges, generally they dont TIE parellel sets together, then series it, it is usually 2 strings all over the place in voltages, tied at the ends with both ends being reasonably close in voltages. and it doesnt do any good to tie the parellels either.

|---1.40v--1.40v--1.41v--1.38v--1.40v--1.40v--1.40v--1.40v--1.39v--1.40v--| = 13.98 <--- now we know if we stay at/below 1/10 C none of these cells will V-drop , or cause a V-drop cascade.
with each single series string, current through every part of the string will be similar,



|---1.40v--1.40v--1.41v--1.38v--1.40v--1.40v--1.40v--1.40v--1.39v--1.40v--| = 13.98 <--- but now we THINK that we can stay below 1/10c for the SET of 2, and that isnt so.
|---1.39v--1.40v--1.40v--1.39v--1.40v--1.40v--1.39v--1.40v--1.40v--1.41v--| = 13.98
Once we tie in the second string current could vary 100% through either string.


Looser tighter, say one set of series items is higher in resistance
|---1.45v--1.40v--1.41v--1.38v--1.42v--1.40v--1.40v--1.40v--1.39v--1.40v--| = 14.05 <---- the ACTUAL voltage doesnt change (its tied) this is its virtual IR thingeee
|---1.39v--1.40v--1.40v--1.39v--1.40v--1.40v--1.39v--1.40v--1.40v--1.41v--| = 13.98 <--- the majority of the current flow is passing through only ONE set :-(

the higher resistance or Loose Batteries dont care about the current going through, heck they will go to 15+V before they hit fully charged.
the lower resistance batteries charged at 14V and hold thier ground , will be sucking up the current going to both sets.
Of course the actual voltage doesnt change, but if you seperated the 2 sets and charged seperatly this is what you would see, they arent the same.
the lower resistance series set is now taking on the WHOLE charge, and that could cause One item in the series set to V-drop (again).

|---1.45v--1.40v--1.41v--1.38v--1.42v--1.40v--1.40v--1.40v--1.39v--1.40v--| = 14.05 <---- the ACTUAL voltage doesnt change (its tied) this is its virtual IR thingeee
|---1.39v--1.40v--1.40v--1.39v--1.40v--1.40v--1.39v--1.40v--1.40v--1.35v--| = 13.93 <--- the majority of the current flow is passing through only ONE set :-(

as you can see, if ONE cell in any series set takes a V-drop Dive, because it has reached overcharge both (or all freaking 10) series sets will dump into the one series set to some degree, and because we are wrong trying to use a High current/ high voltage for both series sets, we can achieve another complete V-drop cascade failure.
so i figure that without voltage control, you must keep the current below what any single cell in the entire pack would v-drop over, because there could be times when the majority of the current would flow through the one low resistance set. if it V-drops everything goes out of wack and snowballs.




so lets mix it up badly
|---1.25v--1.20v--1.21v--1.09v--1.11v--1.23v--1.01v--0.00v--1.35v--1.30v--| = 12.05 <--- at this point i doubt that adds up correct :hairpull:
|---1.29v--1.20v--1.30v--0.00v--1.12v--1.00v--1.25v--1.20v--1.31v--1.21v--| = 12.08

someone went and discharged the pack too low, caused a cell reversal, and everything is really bad.

a slow 1/20c charge rate might not wake up the 0.00 cell items.
|---1.45v--1.40v--1.41v--1.38v--1.42v--1.40v--1.40v--0.00v--1.39v--1.40v--| = <--- hey you got them all charged , , , Almost
|---1.39v--1.40v--1.40v--0.00v--1.40v--1.40v--1.39v--1.40v--1.40v--1.39v--| =
it probably wont wake it up, now you have another mess to fix.

so that is where the voltage max thing can come in handy, to Zap them suckers back to awake before starting the charge, and before slow charging again.
just put it on the power supply Smack it good with 14V, then set the max slow current , and go on.

so your completely stuffed at the slow charge current of ONE cell item , if you use lower Voltage ONLY max it should work, but if things were really off one cell could easily still get pushed into a V-drop and that is always the end of the party, once that happens you might as well toss out the whole thing or call out the fire department

If you use Slow Current ONLY , you could get stuffed at the low end, where a set of cells throughout are still parked at ZERO all the way through the charge.
if you apply Both, in the proper manner you can series parellel charge a 10Px10S, but really why not use D cells at LEAST taking 100hours to charge would only work for some weird solar project

whatever can go wrong will go wrong, such is life, just adjust things to cover it.