Blown battery pcb question, not a bms

[sometrike]

What could the trouble be?

I had a weak but working 24V NIMH battery pack for something called an IGlide wheelchair. Name doesn’t matter too much except it is a light weight chair with two brushed motors and the controller is on the chair. The battery pack has it’s own circuit board which is *not* a bms. The battery circuit board is used for “gas gauge lights” left and right wheel power, thermister, and timer. It worked immediately before I switched batteries. The charge on the NIMH was 26.0V.

I cut out the old battery and spliced in a 24V NMC battery pack charged only to 25.5V. I thought I heard a tiny pop when the wires touched. Sure enough whether or not the “pop” meant anything this system did not work. Polarity was not reversed. And now it will not work with the NIMH either. That gas gauge I spoke about, instead of showing how many lights go on, just the bottom one flashes red.

Although not connected to the chair during the mishap, the battery is fused from the chair with a 20A 32V fuse. I mention that since the spliced in battery pack is that 7s2p Samsung 2200 mAh 10A I mentioned in my previous question.

I don’t understand how unhooking one battery and hooking up a similar voltage of another would blow the pcb. Does anyone have an idea?

 

[fechter]

Can you post a picture of the board?

Can you see anything that looks burnt?

If there are capacitors in the speed controller, it would be normal to get a spark when you connect the battery.

Sounds like it shouldn't have blown from connecting the battery but obviously something is now wrong. Hard to say what would cause that without seeing what happened to it.

 

[sometrike]

Thanks for thinking about this.
I'm having trouble getting my pics small enough to post just now. Nothing appears burnt. Actually this happened twice. I couldn't believe when the backup failed, so I ended up wrecking my main board too.
The speed controller would not be involved since it was not connected. It would be in the controller attached to the chair and the battery was not connected at the time.

 

[sometrike]

These are pics of the same blown battery pcb, left and right side. There is only one MOSFET on the board.

 

[fechter]

I can see a TI battery fuel gauge chip on the board and a shunt resistor for the current measurement.

Besides the battery gauge, does it work otherwise? The FET doesn't look big enough to run much, maybe 15A max as a switch. If you measure voltage across the two outside FET legs, you can see if it is getting gate voltage or not.

By tracing out the wires, it should be possible to tell most of what the board is supposed to do. I don't have a very good idea for the overall layout just from that board. The battery meter part I get, but does it switch the power for the rest of the parts?

 

[sometrike]

That one blinking red light is the only sign of life, but all four lights should be steady on instead. Nothing works, there is no 24V+ output at the connection to the chair pigtail like there should be. And I just now hooked the battery up and measured zero voltage across the two outside FET legs as you asked. Judging by the pigtail connections this board, delivers 24V to each wheel separately, there is a thermister wire, and a timer. The way it works is that for each push of the wheel chair rims, you only get about 5 seconds of power, then have to push again. I don't know if I can post this, but here is the schematic of the battery pcb. Remember this blew without even being turned on or hooked up to the chair. And it happened to the second one too because I didn't believe it could happen to the first one.

 

[fechter]

It looks like maybe there's a fuse on the board somewhere. Check for fuses.

There is also a voltage regulator somewhere. It would be good to measure the voltages going in and out of that. I don't see it in the picutres.

 

[sometrike]

The fuse is between the pcb and the pigtail to the wheelchair. It is not affected by this since it wasn't connected. I see no other fuse.
I think the voltage regulator you mean is marked U5 on the schematic. I found that on the backside of the pcb, lower rt near where power come in, 8 legs, zero voltage on each leg.

On the opposite side of the board there is another but much smaller looking one. I don't know if it is a regulator and cannot find the key to the schematic. However those voltages are, 0.0, 2.1, 1.9, 0.0, read in the same order, top to bottom the others are 3.3, 2.8, 0.0, 0.0

Before when I measured across the FET, I did not think to check individual legs. It turns out that one side does have full battery positive voltage fwiw.

Something seems strange with that schematic to me. I wonder if it really is for the whole chair. There is a mystery box in the center that seems undefined and nowhere on the pcb. At any rate in the upper left corner the shcematic shows a switch SW1. I find neither on the pcb. But there is an on/off switch on the chair, and the chair is supposed to always have positive charge from the battery pack.

 

[sometrike]

Update, this is solved. The problem was that I switched battery chemistries which seemed to cause a pcb within the battery box to fail when I connected it to the replacement cells.
It turned out that all I had to do was plug the battery pack (original cells) back in to the chair to reset the pcb. So I'm back to square one, weak cells.

 

[amberwolf]

So since it works with the old batteries, you might measure out various points on the PCB(s) with them, and then do the same with the new batteries.

When you find a different voltage (probably radically so) then you may have found the circuit that is disabling the system (due most likely to either undervoltage or overvoltage of the new pack vs the old).

You could also test the new batteries with them charged only to the exact voltage the old batteries are at, and while the old batteries are still connected, also connect the new batteries, and then disconnec the old batteries, and see if it makes a difference.

 

[fechter]

The circuit must have some kind of over voltage protection that kicked in. Somewhere on the board, there is a little resistor you could change to increase the trigger point so the new batteries would work. With a good enough picture of the board, both sides, I could probably find it.

 

[sometrike]

Sorry I missed the last two thoughtful responses. I'll try to explain what I learned. I was using two different batteries. One was the original ten year old Nimh which I thought I had wakened up. The other is a new NMC, and it is the one that really confused me because it was charged. Both showed dead battery on the "fuel guage". But once I got to messing around further with the ten year old battery, it's charge came up okay and the wheelchair worked fine, fantastic in fact. But the new battery sucks as if it's dead even though it shows full voltage on my multimeter. So now I question what is going on with it and my ask questions about it later as my goal was to convert battery types.

 

[fechter]

It's hard to tell what the circuit board is doing exactly as it is not a typical ebike part. Just guessing it is a charge controller for Nimh. This probably won't work or be safe for lithium chemistry. I would also guess the board drives the battery meter.

It might be better to try bypassing the board and give up the battery meter as it won't give an accurate reading with NMC. Ideally you want to install some kind of BMS to protect the lithium battery if the battery doesn't already have one built in. This is an important safety feature. The last thing you want is a battery fire.

There are some inexpensive digital voltmeters that could be used to monitor the pack voltage in place of the original meter.

 

[sometrike]

You do good without seeing more than you have.

>It's hard to tell what the circuit board is doing exactly as it is not a typical ebike part.

It’s not a typical anything part.

>Just guessing it is a charge controller for Nimh.

From what I see, yes with the required dedicated Panasonic charger. The pcb has an IC with various controls as you suggest, and there are several thermister connections. I’d decided already that charging could only be done by bypassing the pcb. I’ve shock charged several old Nimh battery packs now and found that bypassing the pcb has not hurt yet.


>This probably won't work or be safe for lithium chemistry.

I agree. The original system as-is can only be used to charge Nimh.

>I would also guess the board drives the battery meter.

Oh yes, lifting the handle on the battery pack activates a magnetic switch triggering the meter to flash the charge. There is no need for the pack to be plugged in to the chair.


>It might be better to try bypassing the board and give up the battery meter as it won't >give an accurate reading with NMC.

The meter will be false with the NMC? Aren’t electrons electrons? Maybe that is my problem. Whatever is causing the meter to say “dead NMC battery” is preventing the main power switch from turning on with NMC.

I had considered bypassing the board, but there are battery/chair pinout connections. They are: 24+, 24+, com, thermister, clock, and data. The first two would be right and left for push power direction. Clock would be the timer as each push provides power for only a few seconds. Doing away with could cause a runaway chair. I have no idea what data is for. Power through the pcb seems essential to me.


>Ideally you want to install some kind of BMS to protect the lithium battery if the battery >doesn't already have one built in. This is an important safety feature. The last thing you >want is a battery fire.

I’m questioning the quality of the battery packs I purchased for this project. They do have BMS’s installed. There is one tag wire for each pair of parallel cells. And they balance.

That’s where I am. I am prone to believe the NMC packs I bought are no good. Your comment about the fuel gauge intrigues me. I have another, bigger NMC I could try but it is mounted on another chair. I just haven’t gotten that far yet.