MOSFET AND IGBT fail in pwm speed control for 260VDC 2200W universal motor

Thread Starter

kandilx

Joined Oct 21, 2011
69
i did my last try on the uni lab it worked ok and peaks didn't exceed 12v volts for 140v supply untill one mosfet fail this was new where the gate and the source are short circuit which result in high current is drawn from the mosfet driver whcich burn it
i now don't know the reason for this failures since the max spikes was around 150v only

edit: the square wave are ok and only small value of ringings occur less than 10v
 
Last edited:

praondevou

Joined Jul 9, 2011
2,942
i did my last try on the uni lab it worked ok and peaks didn't exceed 12v volts for 140v supply untill one mosfet fail this was new where the gate and the source are short circuit which result in high current is drawn from the mosfet driver whcich burn it
i now don't know the reason for this failures since the max spikes was around 150v only

edit: the square wave are ok and only small value of ringings occur less than 10v

Did the MOSFET get hot before failing? What's your heatsink size?
 

praondevou

Joined Jul 9, 2011
2,942
no not at all it didn't get hot , it fail just when i connect the positive terminal before even increasing the speed
When you reconnected the positive terminal, had you increased the voltage?
Did you modify your circuit as I suggested in my previous post?

edit:
edit: the square wave are ok and only small value of ringings occur less than 10v
10V on top of the 12V gate signal? That would be too much. Did you measure it at the mosfet or at the driver output?
 
Last edited:

Thread Starter

kandilx

Joined Oct 21, 2011
69
When you reconnected the positive terminal, had you increased the voltage?
Did you modify your circuit as I suggested in my previous post?

edit:


10V on top of the 12V gate signal? That would be too much. Did you measure it at the mosfet or at the driver output?
yes i was just added 2 battiers in additon to rest(around 8 batteries)

the scope probes were put on the D and S
 

SgtWookie

Joined Jul 17, 2007
22,230
Where is the "small value of ringing" occuring? The gates of the MOSFETs? Or the drain?

How is the ground of the 260v supply connected to the ground of the 12v supply? I am wondering if you may have a "ground loop". There was a fellow about a year ago that was having problems due to issues with multiple grounds and long wires. You are using very small gauge wires, and it looks like they might be long; I don't know, you don't show much of them.
 

praondevou

Joined Jul 9, 2011
2,942
yes i was just added 2 battiers in additon to rest(around 8 batteries)

the scope probes were put on the D and S
So these are 24V batteries, right?


Summary:
The MOSFET doesn't get hot, maximum Vds and Vgs voltage is not exceeded.

It's getting difficult to suggest anything to you. Unless there is something wrong with the layout or you didn't measure correctly there is not much more to do over the internet.

I'm sure there is a simple solution to this problem because the circuit is far from being complex but without having it actually in front of me it's really hard to say what could be the problem having only the information you provided.

I would measure the current at the moment when connecting the battery. (you could put a shunt resistor or current probe in series)

You say, you connected the main voltage and the MOSFET blew immediately? That means the motor wasn't even turning? How many MOSFETs burnt? All of them, only one? Which one ( location)? Are all MOSFETs correctly being driven by the 2110?
Was there no gate pulse at all or small pulses (PWM completely off or not) when you connected the motor voltage? How are the MOSFETs connected (wire length, layout), are you sure they are sharing the current equally?
These are questions difficult to evaluate from here.

We could try to implement an overcurrent protection for the MOSFET using the shutdown feature of the 2110. But this will take some time...
 

Thread Starter

kandilx

Joined Oct 21, 2011
69
Where is the "small value of ringing" occuring? The gates of the MOSFETs? Or the drain?

How is the ground of the 260v supply connected to the ground of the 12v supply? I am wondering if you may have a "ground loop". There was a fellow about a year ago that was having problems due to issues with multiple grounds and long wires. You are using very small gauge wires, and it looks like they might be long; I don't know, you don't show much of them.
well i have 2 ground wires the 1st is going to the source it is about 30 cm long (from the ground of the first battery to source of all 4 mosfets ~1mm^1 wire)
2nd is the ground going to the 555 and the mosfet driver circuit it was small telephone wires about 50cm long connected to the ground of the first battery
bothe grounds meets in the mosfet driver circuit as this shematic


the small ringings occur between the drain and the source

one more thing when the mosfet driver was brunt i try to use the 555 as mosfet driver for only one mosfet i did work ok but the rise time was very big and the mosfet did get hot but there were no spikes or ringins at all
 

Thread Starter

kandilx

Joined Oct 21, 2011
69
So these are 24V batteries, right?


Summary:
The MOSFET doesn't get hot, maximum Vds and Vgs voltage is not exceeded.

It's getting difficult to suggest anything to you. Unless there is something wrong with the layout or you didn't measure correctly there is not much more to do over the internet.

I'm sure there is a simple solution to this problem because the circuit is far from being complex but without having it actually in front of me it's really hard to say what could be the problem having only the information you provided.

I would measure the current at the moment when connecting the battery. (you could put a shunt resistor or current probe in series)

You say, you connected the main voltage and the MOSFET blew immediately? That means the motor wasn't even turning? How many MOSFETs burnt? All of them, only one? Which one ( location)? Are all MOSFETs correctly being driven by the 2110?
Was there no gate pulse at all or small pulses (PWM completely off or not) when you connected the motor voltage? How are the MOSFETs connected (wire length, layout), are you sure they are sharing the current equally?
These are questions difficult to evaluate from here.

We could try to implement an overcurrent protection for the MOSFET using the shutdown feature of the 2110. But this will take some time...
the batteries are 12v each i have bought 20 batteries to be connected in series

only one mosfet at a time burnt or fails it was the 2nd from right to left
in these picture the mosfet wasn't installed yet there places are the 4 green rosettes

one more thing is of that the 4 mosfets have the same gate length when you check the picture you will realize that from node 1 of the ir2110 2 gate resistor are splitted equally and each one is driving 2 mosfet at the same distances
i don't know where is the problem but it maybe with the gournd lood as Sgt Wookie said may be i can disconnect the COM of the ir2110 from being connnected to the source and instead i leave it connected to to ground of the circuit itself? i don't know maybe this will help?
 

praondevou

Joined Jul 9, 2011
2,942
may be i can disconnect the COM of the ir2110 from being connnected to the source and instead i leave it connected to to ground of the circuit itself? i don't know maybe this will help?
I don't know if I understand this question, but if you mean to disconnect Vss and COM you will need a second power supply for VCC/COM. Right now you are using the same power supply for VDD/VSS and VCC/COM, right?


If you meant to disconnect the COM pin from the MOSFET's source, you won't be able to charge/discharge the gate.
 

shortbus

Joined Sep 30, 2009
10,045
If kandilx is just installing the mosfets in the terminal blocks, How is there a heat sink being used? The small heat sinks that screw on to the mosfet or other to-220 device aren't much help.

I still don't see why it is OK in this type of a circuit to tie the Com and Vss together. The circuits are at different potentials. Even if they are being operated from the same battery bank, the timing and switching are different circuits. The timing is only on 12V of the total voltage of the battery bank as I understand the circuit, in the whole circuit the grounds reconnect but shouldn't they be kept separate until then? That way the batteries act as a 'buffer' for spikes from the mosfets and motor.

In another type circuit it wouldn't cause as much problem (like a SMPS) but this is a brushed motor drive.
 

praondevou

Joined Jul 9, 2011
2,942
If kandilx is just installing the mosfets in the terminal blocks, How is there a heat sink being used? The small heat sinks that screw on to the mosfet or other to-220 device aren't much help.
Good point. I asked that too, somehow I can't get an answer.;)

I still don't see why it is OK in this type of a circuit to tie the Com and Vss together. The circuits are at different potentials. Even if they are being operated from the same battery bank, the timing and switching are different circuits. The timing is only on 12V of the total voltage of the battery bank as I understand the circuit, in the whole circuit the grounds reconnect but shouldn't they be kept separate until then? That way the batteries act as a 'buffer' for spikes from the mosfets and motor.

In another type circuit it wouldn't cause as much problem (like a SMPS) but this is a brushed motor drive.
I was assuming that the power supply for the primary and secondary of the 2110 are separated from the 260VDC and not part of it...
It's hard to get a complete circuit diagram + layout...
 

Thread Starter

kandilx

Joined Oct 21, 2011
69
hi there everybody i am in vaction till friday
i will draw a complete schematic includes the connection of the batteries and il will post it soon
 

praondevou

Joined Jul 9, 2011
2,942
sorry for late reply i can only access the internet in my day off i edit the circuit drawing to show the batteries connections i hope it is clear
Hi, I hope it isn't too chaotic in Egypt? :rolleyes:

As for your schematic:

IR2110: Pin 3 is +, pin 2 is COM.

AAC Member Shortbus pointed it out in several other threads, and I agree now... Since we cannot verify your layout and your are working with high voltage/current, use a separated 12V power supply on pin 3 and 2. COM/VSS and VDD/VCC will then be separated. (green 3. and 2. in the picture)

green 4. in picture
--» Use a zener higher than the power supply voltage. The gates of the MOSFETs can sustain voltages up to +-20V, so if you use a zener of 16 to 18V, that's ok. Zener diodes have tolerances, 12V doesn't mean necessarily that it will shunt at 12V.

green 5. in picture
--» Is that 100Ohm, 20W? That really shouldn't be necessary. As stated before. Start with lower DC bus voltage , increase it slowly and observe the voltage spike on the MOSFETs drains at varying PWM.

green 1. in picture
--» If your MOSFET max VDS is 500V then you don't need to protect it with a transzorb of 275V. Increase that value to about 450V. They react in a VERY short time. Also, they are not supposed to absorb constantly overvoltages. It's a protection device, just in case.

Where is the freewheeling diode?
 

Attachments

Thread Starter

kandilx

Joined Oct 21, 2011
69
Hi, I hope it isn't too chaotic in Egypt? :rolleyes:

As for your schematic:

IR2110: Pin 3 is +, pin 2 is COM.

AAC Member Shortbus pointed it out in several other threads, and I agree now... Since we cannot verify your layout and your are working with high voltage/current, use a separated 12V power supply on pin 3 and 2. COM/VSS and VDD/VCC will then be separated. (green 3. and 2. in the picture)

Where is the freewheeling diode?
assuming i get it right i will disconnect 2 and 9 from each other and i will connect pin 3 to 12+ and pin 2 to the ground instead of pin 2 connected to the 12+ and pin 3 to the ground ?! and the pwm will be connected as it is yeah?
one other thing should i connect both ground together after separation ??(the ground of 260v and for the 12v )
 

praondevou

Joined Jul 9, 2011
2,942
Use one power supply for the logic part, +12V to pin 9 (VDD) and 0V to pin 13 (VSS).

Use a separated power supply (isolated from the first one) and connect +12V to pin 3 (VCC) and 0V to pin 2 (COM)

These are the power supplies for the chip, don't connect COM and VSS.

The last one is the 260V, if that is a battery bank you can use the first battery (the one that is connected to the source of the MOSFET) and use it as the +12V power supply that goes to pin 3.

Edit: Don't forget the other items I mentioned:
- replace the 12V zener diodes with 15V types
- don't use the transzorb or use one with a higher voltage
- put the freewheeling diode in
 
Last edited:

Thread Starter

kandilx

Joined Oct 21, 2011
69
hey remember this thread ! well i still didn't figure out a solution but somehow i get to a conclusion
in the past 2 years i have done many attempts using G20N60 IGBT , this always was failing at the begining just after switching on the batteries (i used 4 pin 4 contact switch) i concluded that the switch was part of the problem first because after switching off the there were arching inside the switch for a second or two until the switch was burned after couple of attempts , second ; switching from two points (positive and negative terminals) somehow created a very powerful spikes at the IGBT (because of the long wires from the batteries to the motor and to the control ) this will cause only open ciruit point which is the igbt to fail,
i think a solution to this problem is to use high and low mosfets /igbt or igbt chopper through which there will be several points open on the circuit and separated physically.
well because i am suppose to travel for a long time after less than a week i will make one last attempt using the above conclusions so any suggestions ? :D
 

papanasos

Joined Apr 18, 2013
1
The general rule of thumb is that the turn-on and turn-off time should not account for more than 10% of your PWM period. 5% for turn on and 5% for turn off. You didn't specify said:
ΙI think there is an error on calculations. 288mA needed for 200KHz switching frequency, and not for 20KHz
 
Top