Previous in Forum: Concealed PVC Conduits Running on Wall   Next in Forum: 4 to 4.5 MVA UPS for Kiln Section in a Cement Plant
Close
Close
Close
35 comments
Commentator

Join Date: Jun 2018
Posts: 57

Power MOSFET Blew Up When Increased Duty Cycle From 5% to 10%

09/13/2018 7:03 PM

First, these are the componentes I'm using on my PCB, click on them to see the datasheets:
Gate Driver
Power MOSFET
Freewheeling diode
Current Sensor
The schematic is attached below. Sorry if I'm not using some right symbologies. I took the schematic from the PCB software.
Vbus voltage comes from rectified and filtered mains (110VAC).
PWM frequency is 10 kHz
The gate driver is isolated, so GND1 is ground from rectifier, an GND2 is the digital ground.
I tested the PCB with two different motors.
Motor 1:
130VDC
25.5A
Motor 2:
130VDC
27A
These are the steps that I followed before the transistor blew:
1- I tested the circuit with a light bulb first. Everything ok.
2- Then, I connected motor 1 with no load. Everything ok.
3- I connected motor 2 with a slight load. It started ok at 5% of duty cycle. Then, I increased duty cycle to 10% and the transistor blew up, so the motor went to max speed uncontrollably.
I haven't programmed the fast overcurrent detection from the current sensor. That way I would have probably saved the mosfet.
The question is... what could be the cause of this? high dv/dt and di/dt? need an snubber? I have a lot of mysteries in mind right now.

Register to Reply
Pathfinder Tags: Gate driver mosfet
Interested in this topic? By joining CR4 you can "subscribe" to
this discussion and receive notification when new comments are added.

Comments rated to be Good Answers:

These comments received enough positive ratings to make them "good answers".
Guru

Join Date: Apr 2010
Location: About 4000 miles from the center of the earth (+/-100 mi)
Posts: 9725
Good Answers: 1118
#1

Re: Power MOSFET blew up when increased duty cycle from 5% to 10%

09/13/2018 7:39 PM

Are you sure the diodes are the correct polarity?

Register to Reply
Commentator

Join Date: Jun 2018
Posts: 57
#2
In reply to #1

Re: Power MOSFET blew up when increased duty cycle from 5% to 10%

09/13/2018 7:50 PM

Yes, D5 polarity is okay.

Register to Reply
Guru

Join Date: Apr 2010
Location: About 4000 miles from the center of the earth (+/-100 mi)
Posts: 9725
Good Answers: 1118
#3
In reply to #2

Re: Power MOSFET blew up when increased duty cycle from 5% to 10%

09/13/2018 8:20 PM

I'm confused by the "Plus", "Minus" indications on the motor. If Vbus is positive, then it looks backward. If it's labeled, maybe it has an internal flyback diode.

What is the voltage value of Vbus?

Register to Reply
Commentator

Join Date: Jun 2018
Posts: 57
#4
In reply to #3

Re: Power MOSFET blew up when increased duty cycle from 5% to 10%

09/13/2018 8:32 PM

Vbus is positive and around 150VDC. It's a PMDC motor which its direction can be changed by reversing connections. I don't want to change direction. The application requires the motor to run in only one direction. I haven't considered an internal flyback.

Register to Reply
Guru

Join Date: Apr 2010
Location: About 4000 miles from the center of the earth (+/-100 mi)
Posts: 9725
Good Answers: 1118
#16
In reply to #2

Re: Power MOSFET blew up when increased duty cycle from 5% to 10%

09/15/2018 4:16 PM

Is D5 still healthy?

Register to Reply
Commentator

Join Date: Jun 2018
Posts: 57
#17
In reply to #16

Re: Power MOSFET blew up when increased duty cycle from 5% to 10%

09/15/2018 5:18 PM

Yes, D5 is okay. The driver as well. I just replaced the FET, but I haven't tested with the motor that blew up my FET. I'm trying to carefully see what is happening before burning more FETs. They are pretty expensive. I will get some waveforms when I get my scope. In the meantime I'm just guessing.

Register to Reply
Guru
New Zealand - Member - Kiwi Popular Science - Weaponology - New Member Engineering Fields - Power Engineering - New Member Engineering Fields - Electrical Engineering - New Member

Join Date: Sep 2006
Location: Auckland, New Zealand
Posts: 8778
Good Answers: 376
#5

Re: Power MOSFET blew up when increased duty cycle from 5% to 10%

09/13/2018 9:19 PM

Have a look at the Gate driver U7 datasheet. I haven't used this type before but the GND2 pin (which is a vital reference for the output drive signal) is shown connected directly to the MOSFET source pin in the datasheet but you have connected it elsewhere (likely at a different potential voltage than the driver chip expects).

I suspect the gate drive is not being driven properly because of this which caused the MOSFET failure.

Have a look and let us know what you find.

__________________
jack of all trades
Register to Reply
Power-User

Join Date: Sep 2016
Location: SPAIN
Posts: 233
Good Answers: 13
#6

Re: Power MOSFET blew up when increased duty cycle from 5% to 10%

09/14/2018 1:28 AM

The power stage looks fine, except for the fact that no MOSFET heatsink was mentioned. If you have not done any calculations, you better install one.

If it still does not work as expected, upload some waveforms of the driving signal and the MOSFET voltage for us to be guess what has gone wrong in your design.

__________________
Building the future!
Register to Reply
Guru

Join Date: Aug 2005
Location: Hemel Hempstead, UK
Posts: 5792
Good Answers: 315
#7

Re: Power MOSFET blew up when increased duty cycle from 5% to 10%

09/14/2018 7:21 AM

I don't know much anything about motors but I'd have thought that if they're rated at 130 VDC, and, you're driving them at 150 VDC you might be introducing "unexpected" problems.

Should the motor drive voltage be "smoothed", or are they designed to be driven by a PWM signal?

__________________
We are alone in the universe, or, we are not. Either way it's incredible... Adapted from R. Buckminster Fuller/Arthur C. Clarke
Register to Reply
Guru
United Kingdom - Member - Indeterminate Engineering Fields - Control Engineering - New Member

Join Date: Jan 2007
Location: In the bothy, 7 chains down the line from Dodman's Lane level crossing, in the nation formerly known as Great Britain. Kettle's on.
Posts: 32048
Good Answers: 838
#8

Re: Power MOSFET blew up when increased duty cycle from 5% to 10%

09/14/2018 8:49 AM

OK. Coil of wire motor. Huzz a load of square-wave into it. What happens to the semiconductors when the field collapses in between pulses? Hmmmmm......

__________________
"Did you get my e-mail?" - "The biggest problem in communication is the illusion that it has taken place" - George Bernard Shaw, 1856
Register to Reply
Commentator

Join Date: Jun 2018
Posts: 57
#9
In reply to #8

Re: Power MOSFET blew up when increased duty cycle from 5% to 10%

09/14/2018 9:06 AM

I'm confused about what happens... What do you suggest?

Register to Reply
Guru
United Kingdom - Member - Indeterminate Engineering Fields - Control Engineering - New Member

Join Date: Jan 2007
Location: In the bothy, 7 chains down the line from Dodman's Lane level crossing, in the nation formerly known as Great Britain. Kettle's on.
Posts: 32048
Good Answers: 838
#15
In reply to #9

Re: Power MOSFET blew up when increased duty cycle from 5% to 10%

09/14/2018 12:14 PM

Er, huge negative-going voltages that shag semiconductors?

__________________
"Did you get my e-mail?" - "The biggest problem in communication is the illusion that it has taken place" - George Bernard Shaw, 1856
Register to Reply
Guru

Join Date: Apr 2010
Location: About 4000 miles from the center of the earth (+/-100 mi)
Posts: 9725
Good Answers: 1118
#21
In reply to #9

Re: Power MOSFET blew up when increased duty cycle from 5% to 10%

09/16/2018 10:51 AM

When the transistor turns on, the supply voltage is impressed across the motor (an inductor, L). Current increases at a rate proportional to this voltage...

dI/dt = V/L

When the transistor turns off, assuming there were no flyback diode,

V = L dI/dt

In this case, dI/dt (and thus V) is very large since the current has no place to go. But diode D5 allows the current to continue back through the diode limiting this voltage. If D5 is doing its job, there should be no problem.

Here are some thoughts. Three things come to mind as the possible cause of death of the transistor, some of which have been already mentioned.

1. Excess voltage - from the voltage spike at turn off (bad connection on D5?) or electrostatic discharge (ESD) due to improper handling of the transistor before installation. This could just weaken the transistor, and not cause it to fail right away. (Transistor should be stored in special electrostatic safe bags or in conductive foam.)

2. Excess current - As has already been mentioned, if the transistor is not properly heatsinked, it could overheat.

3. Insufficient drive - If the transistor is not turned completely on and off, but spends time halfway on, the transistor will have higher resistance to the motor current and will overheat.

Register to Reply
Guru
Engineering Fields - Electrical Engineering - Been there, done that. Engineering Fields - Control Engineering - New Member

Join Date: Dec 2008
Location: Long Island NY
Posts: 15501
Good Answers: 957
#10

Re: Power MOSFET blew up when increased duty cycle from 5% to 10%

09/14/2018 9:30 AM

My WAG is the back EMF from the motor is producing more voltage, faster than the snubbing diode can clamp. I suspect this because while a change of 5% to 10% is not a big change that will affect the thermodynamics. It is a significant increase in the amount of energy stored in the motor coils on each PWM cycle. A linear snubber with no turn ON time delay may help this. A linear snubber will also dissipate this flyback energy faster for the time constant of a coil is τ=L/R. Your only resistance in the nonlinear diode is wiring resistance (most likely a low number) in series with the diode. The coils will eventually release their stored energy somewhere, pop goes the circuit.

With this much energy being quickly switched, parasitic energy storage in wiring capacitance and inductance (transmission lines) can also become a problem. This is much more difficult to troubleshoot from a wiring schematic diagram. (Nonetheless, bravo for providing so much information.)

__________________
"Don't disturb my circles." translation of Archimedes last words
Register to Reply
Commentator

Join Date: Jun 2018
Posts: 57
#11
In reply to #10

Re: Power MOSFET blew up when increased duty cycle from 5% to 10%

09/14/2018 9:43 AM

I would like to find a solution that I can apply it on the PCB. Then, if it works, I design the fixed PCB again with the new parts. Maybe an snubber could help. But designing it is not an easy task.. I have been searching for some guidelines, but none of them have help me so far.

Register to Reply
Guru
Engineering Fields - Electrical Engineering - Been there, done that. Engineering Fields - Control Engineering - New Member

Join Date: Dec 2008
Location: Long Island NY
Posts: 15501
Good Answers: 957
#12
In reply to #11

Re: Power MOSFET blew up when increased duty cycle from 5% to 10%

09/14/2018 11:00 AM

Cornell Dubilier has a nice white paper on inductive load snubber design for power circuits. This does not strictly meet your criteria of driving a DC motor but it does apply for effectively driving just the stator of a motor. Since you do not indicate if this motor is a series or parallel rotor winding, permanent magnet rotor or even what values of coil resistance and inductance you have. Then there is the return of mechanical energy one must dissipate when stopping the rotational inertia but that goes even further down a rabbit hole.

__________________
"Don't disturb my circles." translation of Archimedes last words
Register to Reply
Commentator

Join Date: Jun 2018
Posts: 57
#13
In reply to #12

Re: Power MOSFET blew up when increased duty cycle from 5% to 10%

09/14/2018 11:06 AM

The motor is PMDC type.

Register to Reply
Guru
Engineering Fields - Electrical Engineering - Been there, done that. Engineering Fields - Control Engineering - New Member

Join Date: Dec 2008
Location: Long Island NY
Posts: 15501
Good Answers: 957
#14
In reply to #13

Re: Power MOSFET blew up when increased duty cycle from 5% to 10%

09/14/2018 11:44 AM

PMDC? Pakistan Medical and Dental Council? Undefined acronyms quickly lead to confusion.

Now that I've had some fun, the simplest Permanent Magnet Direct Current motor has permanent magnets for the stator and the rotor windings commutate via brush contacts to maintain a continuous rotation. At all times at least one rotor coil is connected to the drive circuitry via these brushes. So you can still use that paper to design a suitable snubber network once the coil inductance and series resistance have been identified. Just remember the faster one dissipates the coil energy the higher the flyback voltage generated by the coil.

__________________
"Don't disturb my circles." translation of Archimedes last words
Register to Reply
Guru
United Kingdom - Member - Indeterminate Engineering Fields - Control Engineering - New Member

Join Date: Jan 2007
Location: In the bothy, 7 chains down the line from Dodman's Lane level crossing, in the nation formerly known as Great Britain. Kettle's on.
Posts: 32048
Good Answers: 838
#18
In reply to #11

Re: Power MOSFET blew up when increased duty cycle from 5% to 10%

09/16/2018 8:48 AM

Just lob a suitable silicon rectifier across the motor terminals, reverse biased.

__________________
"Did you get my e-mail?" - "The biggest problem in communication is the illusion that it has taken place" - George Bernard Shaw, 1856
Register to Reply
Commentator

Join Date: Jun 2018
Posts: 57
#19
In reply to #18

Re: Power MOSFET blew up when increased duty cycle from 5% to 10%

09/16/2018 9:07 AM

Do you mean adding another diode in parallel with D5?

Register to Reply
Guru
Engineering Fields - Electrical Engineering - Been there, done that. Engineering Fields - Control Engineering - New Member

Join Date: Dec 2008
Location: Long Island NY
Posts: 15501
Good Answers: 957
#20
In reply to #18

Re: Power MOSFET blew up when increased duty cycle from 5% to 10%

09/16/2018 10:07 AM

You mean one just like D5 on the schematic? Assuming that diode does turn ON in the claimed 75 ns reverse recovery time there isn't much time for an overvoltage spike but an overvoltage spike still might be the root of the problem. Without a passive snubber network across D5 then only parasitic paths allow current to flow.

On another note, the diode peak repetitive surge current is tested with a 20 kHz square wave. A 10% duty cycle pulse of a 10 kHz switching frequency has the same ON time as a 100 kHZ square wave. This might limit the lifetime of the freewheeling diode but I wouldn't expect this factor to toast the FET.

With so much current being switched between multiple paths there always the possibility of coupling inducing secondary currents where and when they don't belong.

__________________
"Don't disturb my circles." translation of Archimedes last words
Register to Reply
Guru

Join Date: Aug 2010
Location: Melbourne, Australia
Posts: 740
Good Answers: 24
#22

Re: Power MOSFET Blew Up When Increased Duty Cycle From 5% to 10%

09/16/2018 6:43 PM

What is Q6?

Is it rated for this application, or are you operating outside its safe operating region?

Register to Reply
Associate

Join Date: Aug 2014
Posts: 33
Good Answers: 2
#23

Re: Power MOSFET Blew Up When Increased Duty Cycle From 5% to 10%

09/17/2018 8:29 AM

I use 2x+ as a general rule for voltage tolerance of components in a switched DC inductive load circuit.

You will always have 2x voltage spikes and you will probably have much higher.

The spikes are generated by every parasitic inductance in the circuit path.

I once traced a problem to a 1 1/2 inch wire from a switch causing an inductive kick [flyback] at 2x Vs to destroy an IC.

It take very little energy, in an overvoltage situation, to destroy a mosfet.

Register to Reply
Associate

Join Date: Aug 2014
Posts: 33
Good Answers: 2
#24

Re: Power MOSFET Blew Up When Increased Duty Cycle From 5% to 10%

09/17/2018 8:32 AM

Also, I have learned to use a unified grounding scheme wherever possible.

I only split ground when it is absolutely necessary.

Register to Reply
Commentator

Join Date: Jun 2018
Posts: 57
#25

Re: Power MOSFET Blew Up When Increased Duty Cycle From 5% to 10%

09/17/2018 11:30 AM

Update: I've blown up a second MOSFET at startup of the motor. I've replaced the FET with this IGBT IKW50N60DTP and so far it's working well. Anyways, I will capture some waveforms soon.

Register to Reply
Associate

Join Date: Mar 2012
Posts: 42
Good Answers: 5
#26
In reply to #25

Re: Power MOSFET Blew Up When Increased Duty Cycle From 5% to 10%

09/18/2018 3:53 PM

You may be suffering from the power dissipation peaks that occur during switching.

Your 10kHz switching frequency seems excessive, and a lower frequency would have proportionally lower switching losses.

If the FET is blowing up for "no good reason" and can run at lower duty cycles without getting too hot, it is probably being killed by high power pulses during the switching events.

At 100% duty cycle (always on), your I^2*R losses are around 6 watts, so a heat sink is required.

Register to Reply
Commentator

Join Date: Jun 2018
Posts: 57
#27
In reply to #26

Re: Power MOSFET Blew Up When Increased Duty Cycle From 5% to 10%

09/18/2018 4:02 PM

The IGBT hasn't failed yet at 10 kHz switching frequency. But, do you recommend going for lower frequency? How can I determine my optimal PWM frequency?

Register to Reply
Associate

Join Date: Mar 2012
Posts: 42
Good Answers: 5
#28
In reply to #27

Re: Power MOSFET Blew Up When Increased Duty Cycle From 5% to 10%

09/18/2018 5:05 PM

The system may not have any mechanical resonance you can hear at 10kHz, but ask for young people to make that assessment. Many years ago, I asked my dad if he wanted me to reduce the screaming noise (15,750 Hz) coming from the back of his old TV. His response was "What noise?"

As you reduce the switching rate to lower frequencies, you may find frequency ranges that produce undesirable mechanical (audible) resonances. The power losses (and FET heating) due to switching are directly proportional to the switching frequency. Select the operating frequency as low as possible, limited by mechanical resonance of the system and noise you can hear. The limiting factor will be affected by those resonances and by the range and resolution of speed-adjustment control you need.

You may want to use a non-contact thermal measurement device ($15 at Harbor Freight) to measure the temperature on the surface of the FET while changing the switching frequency.

Register to Reply
Commentator

Join Date: Jun 2018
Posts: 57
#29
In reply to #28

Re: Power MOSFET Blew Up When Increased Duty Cycle From 5% to 10%

09/18/2018 5:41 PM

Ok. But I guess the reason why the IGBT has not failed, when the FET failed at startup is because the IGBT handles more breakdown voltage I think. Also, the IGBT doesn't warm up. It stays cool. I can try what happens at lower frequencies for sure. I'm 22, so I probably manage to hear something.

Register to Reply
Guru

Join Date: Dec 2010
Posts: 1686
Good Answers: 116
#30
In reply to #29

Re: Power MOSFET Blew Up When Increased Duty Cycle From 5% to 10%

09/19/2018 11:40 AM

Really, a motor of this size is not used without protection against overload & overheat. Locally, electrical wiring regulations require such protection for motors over 0.37 kW, unless motor is part of an equipment which itself is made to a suitable standard. Simple resistance starters keep peak current to twice rated current - 54 amp in this case - so control to this level of current would be expected with thermal overload behaviour.

You have not given any resistance for the motor, 27 amp/125V = 4.6 ohm - say 10% of that is winding resistance makes 0.46 ohm. 170V could give 170/0.46 = 367 amp, locked rotor.

Comparing the FET & IGBT - FET can typically draw > 300 amp with Vgs =10V, say 50% of samples would be more: IGBT is typical 150 amp with Vge =12V. So the FET can draw a lot more power to cook itself. The max current the FET can withstand for 10 microsec tails off a lot once Vds is > 100-200V - the IGBT is actually specified to withstand 400V for 10 microsec at Vge =12V with short circuit load.

Supposing the motor inductance is 1 mH, stored energy at 300 amp is 0.5LI2 = 45 Joules. At turn-off, with maximum withstand for devices of fractions of Joule it does not need much of that energy to go into device to destroy it. With L/R = 2 ms there will be negligible back emf starting with stationary motor.

You should really have checked the currents & voltages seen by FET at 5% duty cycle - and estimated the time constants before pushing the limits.

Is any kind of overcurrent/overheat protection provided by the current sensor ??

Register to Reply
Commentator

Join Date: Jun 2018
Posts: 57
#31
In reply to #30

Re: Power MOSFET Blew Up When Increased Duty Cycle From 5% to 10%

09/19/2018 11:55 AM

The current sensor that I'm using (TLI4970) provides fast overcurrent detection using interrupts. I'm haven't programmed that feature yet. So far, I've been trying with the IGBT, and it works well. I'm still considering that PWM frequency and inductive kickback were the causes why I blew up two FETs, but haven't blew up IGBTs. So, it makes me rethink that IGBTs for sure are better below 20 kHz, and handle more voltage. I thought the MOSFET was pretty good for my application, but it failed. Maybe because I needed some kind of protection, like a snubber to help clamp the voltage transients. But the IGBT has been okay with only the flyback diode.

Register to Reply
Guru
Engineering Fields - Electrical Engineering - Been there, done that. Engineering Fields - Control Engineering - New Member

Join Date: Dec 2008
Location: Long Island NY
Posts: 15501
Good Answers: 957
#32
In reply to #31

Re: Power MOSFET Blew Up When Increased Duty Cycle From 5% to 10%

09/19/2018 12:07 PM

The location of your current sensor will only detect the current that succeeds to flow through your controlling switch (IGBT or FET). It is also a useless protection device if one hasn't programmed it to protect.

__________________
"Don't disturb my circles." translation of Archimedes last words
Register to Reply
Commentator

Join Date: Jun 2018
Posts: 57
#33
In reply to #32

Re: Power MOSFET Blew Up When Increased Duty Cycle From 5% to 10%

09/19/2018 12:16 PM

I'm working on the protection feature now. About its location, perhaps it would have been more suitable if I placed it at the collector leg or between Vbus and M+. Not at source/emitter leg like I did. But I'll take it into consideration for the next PCB modifications, as I already have the PCB physically with the sensor at the source.

Register to Reply
Participant

Join Date: Feb 2018
Posts: 3
#34

Re: Power MOSFET Blew Up When Increased Duty Cycle From 5% to 10%

09/19/2018 4:26 PM

I confess to not having made a detailed analysis of your circuit, but offer a couple of suggestions just the same. Take a look at the allowed safe operating area of the semiconductor you are using. Switching an inductive load can produce excursions that will exceed the SOAR. Low duty cycles may not result in cumulative heating leading to your device surviving at low duty cycles.

If the SOAR is being exceeded, try tuning the circuit with capacitors across the semiconductor switch to shape the path to stay within the SOAR.

Regards,

Myron Boyajian

Register to Reply
2
Associate

Join Date: Aug 2014
Posts: 33
Good Answers: 2
#35

Re: Power MOSFET Blew Up When Increased Duty Cycle From 5% to 10%

09/20/2018 9:01 AM

I design BLDC motor drivers and switching power supplies.

It looks like you can read a data sheet well enough.

Requirements, inductance of the motor in general, have driven pwm rates of 20kHz to nearly 200kHz in my designs.

The motor is usually not the problem when mosfets blow up

Using transistor peak capability ratings is good practice as long as sufficient margin is allowed, [50%].

Always asses the avalanche capability of mosfets you choose.

You will have 2x voltage excursions due to flyback effect working on every length of conductor in the signal path, these can even ride on top of each other to provide higher multiples. The length of the conductive path and the frequency content of the driving signal will determine the frequency content of the flyback energy. Frequency content is ~ 0.35/ton.

I fixed a particularly badly designed medium power sine wave generator circuit using 2 mosfets and inductive elements. I found the 7th multiple of the operating voltage across the mosfets just stacked up pretty as you please [after lots of oscilloscope tweaking] the top rider was only 5ns wide [200MHz]. Slowing the mosfets ton toff was a good quik-fix to ensure operation for more than the 30 minutes of the original design.

There is a balance to achieve between efficiency, turn-on turn-off time of the mosfets and pwm frequency. There usually is a lot of wiggle room here.

Slowing the ton toff reduces the frequency content of the energy fed into the flyback effect and adds avalanche immunity to the mosfets, but increases mosfet power dissipation. The motor generally has a large time constant, the wiring is the threat.

I hope any of this helps.

Register to Reply Good Answer (Score 2)
Register to Reply 35 comments
Interested in this topic? By joining CR4 you can "subscribe" to
this discussion and receive notification when new comments are added.

Comments rated to be Good Answers:

These comments received enough positive ratings to make them "good answers".
Copy to Clipboard

Users who posted comments:

67model (1); Carl Van Wormer (2); Ivanov327950 (1); jack of all trades (1); johnhotmer (3); NIKT (1); PWSlack (3); Randall (1); redfred (5); Rixter (4); WAWAUS (1); XaviPacheco (12)

Previous in Forum: Concealed PVC Conduits Running on Wall   Next in Forum: 4 to 4.5 MVA UPS for Kiln Section in a Cement Plant

Advertisement