SOLVED! Blown Controller Mosfets!Is it worth replacing?

SakisMx

1 mW
Joined
Jun 27, 2020
Messages
17
Location
Volos
Hi!
After fixing an overheating issue,I forgot that I left my controller on,after troubleshooting.When I put the cover back on I shorted the first mosfets' pins resulting in it exploding and other 2 mosfets shorted between drain and source.Is it worth replacing the damaged mosfets,or the whole controller is damaged now,beyond repair?

Can I measure something with a multimeter on the board so I can check what is damaged and what works? So far I have measured the mosfets and except for these 3 that got damaged the others are okey with high resistance and no shorts! Visually nothing seems damaged except for the mosfets
 
There could easily be other things in there that got fried by the voltage transients. But if you're game to swap MOSFETs, that's at least a cheap thing to try and see if it works.
 
I agree, give it a go,
I have replaced blown mosfets a couple of times on a dead controller.
Most of the circuitry round there is pretty robust as you have high spikes of voltage and current so you may be lucky
I've also blown a control chip for mosfets so that might be the next thing to go - there will be a smaller FET, probably a little three pin surface mount thing driving the bigger one or a multi pin (maybe 8 pin) driver chip (less likely).
I would compare the blown channel with the good ones. It is likely that voltages on the control circuits will be similar and you have 6 sets of the same circuit so hopefully can see if something is different.
Check for exploded tantalum capacitors maybe - visible if you look
Look for burn marks and see whether you brushed against anything else
Good luck
 
SakisMx said:
Hi!
After fixing an overheating issue,I forgot that I left my controller on,after troubleshooting.When I put the cover back on I shorted the first mosfets' pins resulting in it exploding and other 2 mosfets shorted between drain and source.Is it worth replacing the damaged mosfets,or the whole controller is damaged now,beyond repair?

Hard to say.

One thing you might try is to cut all the MOSFETs off (assuming they are leaded) and try running it again; also spin the wheel by hand (if it's sensored.) You won't get any output. But if you have access to a scope, take a look at the gate drive pin for each of the MOSFET legs. If at least the bottom side is wiggling up and down* then the FET drivers are probably OK.

(* - to check this use a scope, or if that's not available, use a regular voltmeter. It should be some intermediate voltage, like 6V.)
 
Hi and thank you all for your help!

This is where I short circuited the board(at the red circle).
Also there is a little burn mark on the second mosfet's drain pin but it didn't get damaged
Screenshot_20210504-112307_Gallery.jpg

This is what my controller looks like
20210504_115108.jpg

BobBob said:
I would compare the blown channel with the good ones. It is likely that voltages on the control circuits will be similar and you have 6 sets of the same circuit so hopefully can see if something is different.
Can you guide me as to what I have to check? I only have a multimeter but I can give the board to a shop to get checked.

JackFlorey said:
One thing you might try is to cut all the MOSFETs off (assuming they are leaded) and try running it again; also spin the wheel by hand (if it's sensored.) You won't get any output. But if you have access to a scope, take a look at the gate drive pin for each of the MOSFET legs. If at least the bottom side is wiggling up and down* then the FET drivers are probably OK.

(* - to check this use a scope, or if that's not available, use a regular voltmeter. It should be some intermediate voltage, like 6V.)
So If understood correctly, If I remove all the mosfets and run it like that I should get 6V on each mosfets gate pin,right? Also when I spin the wheel I should look again for 6V on the gate pins, right? My motor has hall sensors and controller works both sensored and sensorless.My multimeter however is for 10A maximum and if I pass more current through I will fry it,so can I check gate voltage or I will fry it?
 
SakisMx said:
BobBob said:
I would compare the blown channel with the good ones. It is likely that voltages on the control circuits will be similar and you have 6 sets of the same circuit so hopefully can see if something is different.
Can you guide me as to what I have to check? I only have a multimeter but I can give the board to a shop to get checked.
Set the meter to volts
measure the volts on the FET next door, then the one next to that and work out the pattern
see if the blown one is the same or different.
The way the controller works is there is a microcontroller that is the brain but has weak signals - it switches some little tiny switches (small FETS) which are three pin black rectangles connected to the big FET gate.
To figure out whether it is working you need to know if the brain is working, if it is signalling the little FETs and if they are signaling the big FETS. You can start with the big ones and work back. If you find that the micro (lots of pins big chip) is signalling but the little FET is not then you may need to very carefully replace that FET - I've soldered lots but use lead solder and you will need a steady hand and either good eyesight, glasses or a magnifier.
Best of luck
SakisMx said:
JackFlorey said:
One thing you might try is to cut all the MOSFETs off (assuming they are leaded) and try running it again; also spin the wheel by hand (if it's sensored.) You won't get any output. But if you have access to a scope, take a look at the gate drive pin for each of the MOSFET legs. If at least the bottom side is wiggling up and down* then the FET drivers are probably OK.

(* - to check this use a scope, or if that's not available, use a regular voltmeter. It should be some intermediate voltage, like 6V.)
So If understood correctly, If I remove all the mosfets and run it like that I should get 6V on each mosfets gate pin,right? Also when I spin the wheel I should look again for 6V on the gate pins, right? My motor has hall sensors and controller works both sensored and sensorless.My multimeter however is for 10A maximum and if I pass more current through I will fry it,so can I check gate voltage or I will fry it?
First, read up on how to use a multimeter and don't use the amps setting until you understand why I am saying this or you will indeed blow the (hopefully fuse in the) meter.

Use the volts setting to check for voltages, if you ever need to check amps then suggest you get a clamp meter

All the mosfets? I assume you mean the blown ones?

Easiest thing may simply be to replace the blown mosfets and see if it works if it's a cheap controller

or check if the voltages are the same on the replaced ones as on the others, if not replace the drivers, if so maybe it's fixed.

Another alternative is, as you say, take it to a shop - pretty simple circuit for them to figure out - might be best to take the bike so they can check with the motor and battery

Get a quote for the work first etc, insist on good quality mosfets to replace the blown ones.

Note that this is not a detailed troubleshooting guide and assumes you are not too bothered about learning the detail of how it all works. If you are, dive into some online tutorials first about how to use a scope, how to diagnose electrical faults etc. Lots to learn.
 
SakisMx said:
So If understood correctly, If I remove all the mosfets and run it like that I should get 6V on each mosfets gate pin,right? Also when I spin the wheel I should look again for 6V on the gate pins, right? My motor has hall sensors and controller works both sensored and sensorless.My multimeter however is for 10A maximum and if I pass more current through I will fry it,so can I check gate voltage or I will fry it?
Well, you will get a PWM signal (high to low to high, over and over again.) But it's too fast for the meter to see, so you will probably see something in between ground and the board's housekeeping supply (usually 12V or so.) So if you hook it up and spin the wheel if you see anything OTHER than 12V or gnd that's a good sign - especially if it seems to change with throttle setting and/or wheel position.

But as I mentioned a scope is a better way to look at it. You might consider one of the smaller USB based scopes. This is a great one:

https://downloads.saleae.com/specs/Logic+8+Data+Sheet.pdf

You might even consider something like this but I have no experience with these:

https://www.amazon.com/KKmoon-Intelligent-Oscilloscope-Multimeter-Resistance/dp/B07WJXHV26/ref=sr_1_36?dchild=1&keywords=usb+oscilloscope&qid=1620160360&sr=8-36

However after looking at that picture I have a different suggestion. You shorted the gate of one FET, which means (if you are lucky) you only blew out one phase. The things that were likely to have blown are:
-2 top FETs
-2 bottom FETs
-the gate driver (it's also connected to the FETs)

So an alternative might be to:

1) Replace those 4 FETs (it's the 4 that you can see in your bottom-of-board picture. Try it again. If that works you are good to go. If not:

2) Replace the gate driver in that section. There will be a trace from the pin you blew to a small IC. (There may be a resistor along the way; the signal goes through the resistor.) Replacing ICs is somewhat difficult, so you may want to get a local shop to do that. (If it doesn't have a separate gate driver for that phase, and instead uses one controller IC, then you have to replace that. Hard to see from your picture.)

I'd give you a 50/50 chance of fixing it by doing just that.
 
JackFlorey said:
2) Replace the gate driver in that section. There will be a trace from the pin you blew to a small IC. (There may be a resistor along the way; the signal goes through the resistor.) Replacing ICs is somewhat difficult, so you may want to get a local shop to do that. (If it doesn't have a separate gate driver for that phase, and instead uses one controller IC, then you have to replace that. Hard to see from your picture.)

I'd give you a 50/50 chance of fixing it by doing just that.

And for the sake of clarity, we are both giving the same advice using different explanations :)

I was guessing that the circled parts at the top may be pairs of transistors, driving (turning on and off) the main FETs (Darlington) though I could be wrong

The device at the bottom looks like the micro controller to me.
I might expect a connection from the microcontroller pin to the small transistors above and another from the small transistors to the big ones
A multimeter set to ressitance Ω or ohm with the controller switched off and left to discharge might confirm this, the main thing is to see what the big FETs connect to and see if that bit is also blown
Controller.JPG
 
BobBob said:
JackFlorey said:
2) Replace the gate driver in that section. There will be a trace from the pin you blew to a small IC. (There may be a resistor along the way; the signal goes through the resistor.) Replacing ICs is somewhat difficult, so you may want to get a local shop to do that. (If it doesn't have a separate gate driver for that phase, and instead uses one controller IC, then you have to replace that. Hard to see from your picture.)

I'd give you a 50/50 chance of fixing it by doing just that.

And for the sake of clarity, we are both giving the same advice using different explanations :)

I was guessing that the circled parts at the top may be pairs of transistors, driving (turning on and off) the main FETs (Darlington) though I could be wrong

The device at the bottom looks like the micro controller to me.
I might expect a connection from the microcontroller pin to the small transistors above and another from the small transistors to the big ones
A multimeter set to ressitance Ω or ohm with the controller switched off and left to discharge might confirm this, the main thing is to see what the big FETs connect to and see if that bit is also blown

I measured these smal fets (They are the gate drivers right?) and none of them are shorted,they all have the same corresponding resistance per circuit(out of these 6)
I also measured each resistor and capacitor on the board and nothing is shorted, everything has resistance!
Moreover I found out that the second mosfet had also popped from the beginning, so I replaced that.
Now I can't find any more issues with the board but it unfortunately does not work. When I connect battery, motor phase wires and throttle it does nothing not even a single sound.I even connected self learn wire to rule out possible throttle issues, so that narrows it down to the controller.The positive though is that nothing popped or exploded.

Where do I go from here? Can I check something else?
I gave it to a shop and they said they didn't find any issues, but they couldn't understand the circuit(they said they don't have the schematic).If it is something in particular that can be checked I can give it to them again so they can check for that.

However, I also have my old controller which is for 60 volts (I have 72v 20s).If I put all mosfets and big caps of this controller to that, will it work?
 
BobBob said:
I was guessing that the circled parts at the top may be pairs of transistors, driving (turning on and off) the main FETs (Darlington) though I could be wrong
Yep, they may be discrete transistors rather than a single IC. Cheaper that way. But harder to identify what the replacement parts are. For the most part almost any NPN/PNP pair that meets the open circuit voltage requirements will work for the lower gate driver. The upper gate driver is more complex, since it depends on how they implemented that.
 
I finally found what the issue is!
There is a shorted gate driver on the negative side of the green phase mosfets!
1HC Is written on top of the transistor.
I also found 1HC transistor on my old controller can I take that and replace the shorted one on my current controller?
 
SakisMx said:
I finally found what the issue is!
There is a shorted gate driver on the negative side of the green phase mosfets!
1HC Is written on top of the transistor.
I also found 1HC transistor on my old controller can I take that and replace the shorted one on my current controller?
Great :)

According to a quick google, that may be a silicon NPN SMD transistor with a current gain rank Q which is an Hfe / a gain of 150 to 300, from example datasheet: https://www.taitroncomponents.com/catalog/Datasheet/MMBT8050.pdf
As we only have three letters this may refer to several different devices

I would ask for a second opinion as to how important matchin it is, but as we don't know who the original manufacturer is, it may not be quite the same but it might work

There is variation due to manufacturer and poor transistor performance tends to kill controllers though generally the power FETS. Slow switching due to a slow or underpowered driver can cause overheating but with the same spec you could be ok

If it runs I'd check whether the repaired phase heats up more than the others.
 
I replaced the little gate driver with the one from my old controller and it now WORKS!!
Thank you all for your assistance!
The overheating issue persists however.
The only thing I know is that the phase wire ends are pretty much destroyed from fiddling around, but I think they make good contact as only the controller gets hot, not the phase wires!
I will fix the wire ends and report here if the overheating issue persists!
 
SakisMx said:
I replaced the little gate driver with the one from my old controller and it now WORKS!!
Thank you all for your assistance!
The overheating issue persists however.
The only thing I know is that the phase wire ends are pretty much destroyed from fiddling around, but I think they make good contact as only the controller gets hot, not the phase wires!
I will fix the wire ends and report here if the overheating issue persists!

:bigthumb: Great, glad it works. :bigthumb:

Can you tell what is overheating?

If it's FETs then bigger heatsinks will help as can thicker traces on the board (solder and copper wire)

If it's wires or connectors then normally bigger better wires and well made connectors should help but it depends on where the problem is.

For wiring and tracing where it's safe to do so, I like to put a one amp curent through the wire with a CC supply and then measure mV, as at 1A 1mV equates to a mΩ so you can find problem areas with good sensitivity (no chance of measuring that on a cheap multimeter otherwise)
Cheapest I could find (not a recommendation) CC supply https://www.aliexpress.com/item/1005001588863528.html
An alternative might be to measure voltage drops at a constant speed.
Another is simply feel for what is getting hot
 
I strengthened all phase wire ends and the overheating issue stopped!After a long uphill test the controller was just warm!
The only thing left now is to also strengthen the hall sensor wire ends where they plug in the connector from the motor side.When I accelerate from a stop there is a strange sound(like hitting 2 pieces of glass together) which sounds only once and right then I don't have that much torque(after like 2 seconds its back to full torque) so I think its the hall sensor wire not making good contact!
 
Back
Top