using multiple thermistors in parallel to monitor nimh?

[monster]

hi

i lost another home-made charger transformer to overheating the other day. i should be able to repair it tho. can i use multiple thermistors in parallel to charge terminate if any one of them gets hot? this might be useful for charging lots of NiMH batteries in series if they are all different DOD, or different capacities, or you want to fast charge at the limit of your charger.

i seem to remember that resistors in parallel is better than series for this kind of thing?

at the moment i am just using one thermistor in the weakest pack.

 

[NickF23]

My next nimh pack will consist of 3 separate blocks of cell in different spaces on the bike. In total I'll have 40 cells in series. In order to have at least 1 thermistor in each block I'm going for a 2 series 2 parallel configuration, which, last time i checked, gives the same value as 1 thermistor. I think 1 thermistor has to get fairly hot if the others were cold for my charger not stop charging though. Another option is a thermal breaker

 

[monster]

series and parallel thats a clever idea.

i've got a custom built charger that uses a manually adjustable thermostat circuit to terminate the charger. i won't need to series them as well, i can adjust the set point to account for multiple parallel thermistors.

so i think the parallel thermistors is sound then? even if the others are cold, if one of them gets hot it will terminate the charge?

 

[deardancer3]

my input:

I have thought this thru several times, and nothing works like a single dedicated thermistor going to one charger and one battery.

when I do the calculations for using 4 identical thermistors- if one pack/thermistor is fast and heating up faster than the others (cause its charged up early), the other thermistors prevent the warmest one from triggering the charger to slow down- keeping the rapid chage current on too long, and heating up the hot pack too far.

try doing the resistor/voltage trip caluclations yourself using a 10k NTC thremistor. (say 10k =75F, 9k=85F and 8K= 95F)

(Yeah i know it is really change in temp over change in time, but this exercise shows how the resistor network doesnt do what we want it to do.)

so you start at all 4 equal 10 k ohms. but later lets say one thermistor is at 8k and three guys are at 9k. 8k plus 9k = 17k ohms, 9 plus 9 = 18k , and the parallel value is 8.7k, still disguising the one warm packs value of 8k to the charger. so the warm pack gets even hotter til its at say 7.5 k ( too hot) and one other gets to 8k and two at 8.5k, still not quite enough to trip the charger.

there is a fellow in D.C. working on a charger that has separate current outputs and thermistor inputs and is looking for testors because of this above issue.

dick

 

[Toorbough ULL-Zeveigh]

Well Saft is now a big believer in paralleling multiple thermistors, after their NiMH packs in the Volocci suffered from frequent meltdowns.
I'm sure that cost them a pretty penny in warranty coverage.
When reading the manual for their VHF module it's rather conspicuous that they seem to go out of their way to point out their new packs contain more than one thermistor.
As tho it's a sore point with them that they had to learn the lesson the hard way.

It's obvious that the center of a pack heats up more than the outside (I guess it wasn't so obvious to Saft) where the heat can dissipate more readily & that the bigger the pack a single thermistor isn't going to be able to cover all the bases.
The thing to remember about adding parallel resistances is that the total resistance will always be less than the smallest resistance branch.
This means that as soon as any single one in parallel starts to heat up you get a lo resistance trigger.
The thermistor isn't supposed to function as your primary charge termination anyway, only act as a failsafe along with a timer.
Since you have the ability to adjust for the lower initial resistance it looks like you can parallel four of them if Saft is anything to go by, two for sure.

 

[fechter]

In series-parallel, the resistance will reflect the average temperature of all 4.

Seems like you'd want to use diodes or some way to pick out the hottest one and use that to trip the charger.

 

[deardancer3]

In series-parallel, the resistance will reflect the average temperature of all 4.

Seems like you'd want to use diodes or some way to pick out the hottest one and use that to trip the charger.

Yes, Yes, yes. somehow devise a method to trigger on the lowest value, especially on NTC Negative temp co-efficient, whre the warmest pack ha the losest value resistance thermistor, and hence lowest voltage.

But remember, this thermistor has several functions, two being --absolute temp., and the bigger one being the temp change tracking in the "delta temp over delta time" charger calculation. (the pack can heat up kinda slowly for several reasons, but when you get to 1 degree centigrade rise per minute, the pack is real close to fully charged)

 

[fechter]

Well, you could put diodes in series with each thermistor, then put all the diode-thermsitor strings in parallel. This way all the current would be going through the lowest resistance one and ignore all the others.

Putting a diode in series would cause a voltage drop, which would throw the calibration out of whack, so it would be necessary to know a bit more about how the charger 'looks' at them. There should be some way to do it.

 

[monster]

the series parallel is good for people who want to add multi thermistors without changing the charger. if you stick two in series in the same pack and then parallel this with another two in series in another pack. then this would act as one thermistor with no need to change a charger set point that cant be changed.

i like the idea with the diodes, tho i don't fully understand it.

 

[fechter]

If you hook up the charger and start charging, then measure the voltage on both ends of the thermistor with respect to battery neg. it would give me an idea of how this might work.

The idea is the charger is looking at a voltage that's controlled by the thermistor somewhere. If you have multiple voltages going to diodes configured to 'pull down', the resultant will be the lowest of the group.

The problem is to figure out which side the diodes need to go on and to correct for the voltage drop of the diodes.

 

[monster]

so i set up the thermistors in parallel and it works.

all i had to do was calibrate it to trip when one diode reached 50*C and it worked fine for both. i just did this with a cup of luke warm coffee. the system now prevents my charger and batteries from overheating using the same temp sensing circuit (£5 from maplin). this means i can get the most out of my transformer without worrying about frying it. i have a computer fan cooling the charger. i can pull 385W from a 300W charger, where as before i was worried about just 250W.

i would recommend everyone put a fan on their chargers because i think heat is what kills the chargers so much.