Why did my MOSFETS blow up?

Go To Last Post
27 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Well they did not go out with a boom but for some reason the gate to source resistance is now 92 Ohms so they are unusable.

They were part of an H-bridge configuration that controlled a stepper motor. Type IRF510. The schematic is bellow.

Everything was working out great until I swapped the MOSFET 2V power supply for one of the same voltage.

How does something like this happen to a MOSFET? what did I do wrong? Should I be using resistors on the FET's gate?

The FETs that broke were both connected directly to an AVR output pin. The ones that broke are in the blue circles bellow.The logic circuit works on a 9V battery with a 5V regulator.

TO THE FINDER... THE ISLE OF KOHOLINT, IS BUT AN ILLUSION... HUMAN, MONSTER, SEA, SKY... A SCENE ON THE LID OF A SLEEPER'S EYE... AWAKE THE DREAMER, AND KOHOLINT WILL VANISH MUCH LIKE A BUBBLE ON A NEEDLE... CAST-AWAY, YOU SHOULD KNOW THE TRUTH!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Also have a merry xmas everyone =]

TO THE FINDER... THE ISLE OF KOHOLINT, IS BUT AN ILLUSION... HUMAN, MONSTER, SEA, SKY... A SCENE ON THE LID OF A SLEEPER'S EYE... AWAKE THE DREAMER, AND KOHOLINT WILL VANISH MUCH LIKE A BUBBLE ON A NEEDLE... CAST-AWAY, YOU SHOULD KNOW THE TRUTH!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

 

That's a ridiculous circuit.... if you want PHA connected to 12V at Q1, then the gate will need to be at least 15V min & more likely 22V from gnd

...you anit gonna get there coming from a 5V AVR pin!

 

 

If you sorta partly turn it on, that is even worse, since now you've created an electronic friction brake--so the fet just gets red hot.  you want the fet 100% or 0% on/off ONLY (or as close as possible).

 

 

You need high side drivers for the high side gates!!!   Have you not looked at your fet datasheets?

 

Please use the proper mosfet symbol that includes the mosfet diode, so you have a full idea of what is going on

 

 

Hope you still have a great Christmas...it will be a better 2020!!!

 

 

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

Last Edited: Tue. Dec 24, 2019 - 09:41 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Well, break between gate and source/drain means there was a bigger voltage between gate and source than the gate insulation layer can survive. So how it's handled voltage generated by coils when you turn off the switches?

Computers don't make errors - What they do they do on purpose.

Last Edited: Tue. Dec 24, 2019 - 09:01 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Although these transistors have reverse diodes, it is necessary to use some type of Snubber
for example a capacitor of few nF between Drain and source

Another problem of the circuit is the critical level of the control signal, very close to the minimum.
The upper traansistors may not reach saturation. Especially as the 5V battery is discharged
 

Ledo

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

And, I doubt that there is any dead-time that prevents Q1 & Q3 (or any other pair) from both being on at the same time.

 

For the OP, there is a good reason why there are bridge driver ICs out there. There are even drivers with built-in charge pumps so that you can use N channel FETs for the high-side switches with logic level inputs. I strongly recommend that you avail yourself of one of these drivers. 

 

Jim

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Or use P-Fet's for the high side FET.

 

Jim

 

 

 

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I am aware there are ready to use drivers. I wanted to do this for practice. There is dead time programmed in the IC. But is the logic above not enough? Should there always be dead time? The logic prevents any of the pairs to activate at the same time but does not offer dead time.

TO THE FINDER... THE ISLE OF KOHOLINT, IS BUT AN ILLUSION... HUMAN, MONSTER, SEA, SKY... A SCENE ON THE LID OF A SLEEPER'S EYE... AWAKE THE DREAMER, AND KOHOLINT WILL VANISH MUCH LIKE A BUBBLE ON A NEEDLE... CAST-AWAY, YOU SHOULD KNOW THE TRUTH!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I don't really know exactly currents or so, but I wouldn't bother to create it as there are some interesting stepper motor drivers out there. Like DRV8834 Low-Voltage Stepper Motor Driver Carrier

Computers don't make errors - What they do they do on purpose.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I have looked at the datasheet and I understand all you say. But the drain is not connected to 12V but to 2V.

So, given that a phase of the motor is at 2.2 Ohms the FETS don't limit the current much and don't produce much heat.

Good advice on the symbols! I am new to eagle and hope to learn better=]

Hope you have an even better Xmas ;)

 

TO THE FINDER... THE ISLE OF KOHOLINT, IS BUT AN ILLUSION... HUMAN, MONSTER, SEA, SKY... A SCENE ON THE LID OF A SLEEPER'S EYE... AWAKE THE DREAMER, AND KOHOLINT WILL VANISH MUCH LIKE A BUBBLE ON A NEEDLE... CAST-AWAY, YOU SHOULD KNOW THE TRUTH!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

It makes sense that this could be the problem. Unfortunately I don't have the equipment to test for that.

On the other hand, the FETS include a freewheeling diode so I was thinking it would take care of the coil currents.

Is there info about this diode and how much it can handle?

TO THE FINDER... THE ISLE OF KOHOLINT, IS BUT AN ILLUSION... HUMAN, MONSTER, SEA, SKY... A SCENE ON THE LID OF A SLEEPER'S EYE... AWAKE THE DREAMER, AND KOHOLINT WILL VANISH MUCH LIKE A BUBBLE ON A NEEDLE... CAST-AWAY, YOU SHOULD KNOW THE TRUTH!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

tellSlater wrote:
Is there info about this diode and how much it can handle?

 

That's what the datasheet is for. Took me 10 seconds to Google it.

 

How much current does the motor require? The IRF510 is not the best choice  - it can only pass a small amount of current with 5V on the gate. You really want logic level mosfets. As well, it is an old part - there are much better performing parts these days. If you've got some failed PC motherboards floating around, these usually have a handful of mosfets on them - you might be able to harvest some of these. 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

God this transistor is rated at 100V max drain to source. It does not add up to me.

Is it really possible to have caused a more than 100V spike by alternating 300mA of current at 2V on the motor's coils while every MOSFET has a freewheeling diode?

TO THE FINDER... THE ISLE OF KOHOLINT, IS BUT AN ILLUSION... HUMAN, MONSTER, SEA, SKY... A SCENE ON THE LID OF A SLEEPER'S EYE... AWAKE THE DREAMER, AND KOHOLINT WILL VANISH MUCH LIKE A BUBBLE ON A NEEDLE... CAST-AWAY, YOU SHOULD KNOW THE TRUTH!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

These FETS really fit into my project nicely right now but.. I just need to know how they failed. It really irritates me to not know that. It makes no sense right now

I do have the datasheet but I don't know how to properly interpret the diode info.

TO THE FINDER... THE ISLE OF KOHOLINT, IS BUT AN ILLUSION... HUMAN, MONSTER, SEA, SKY... A SCENE ON THE LID OF A SLEEPER'S EYE... AWAKE THE DREAMER, AND KOHOLINT WILL VANISH MUCH LIKE A BUBBLE ON A NEEDLE... CAST-AWAY, YOU SHOULD KNOW THE TRUTH!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

tellSlater wrote:

God this transistor is rated at 100V max drain to source. It does not add up to me.

Is it really possible to have caused a more than 100V spike by alternating 300mA of current at 2V on the motor's coils while every MOSFET has a freewheeling diode?

 

Yep - easily. Blame Lenz for that and Faraday. Note that the 100V spec is drain to source. The spec for gate to drain is around 20V. So - did your mosfets die from overvoltage or did they die from thermal problems?

How much current is available from your 2V supply? As well, the mosfets won't pass 300mA with only 3V of gate voltage. Then there's the 4001 gates used to drive the mosfets - 4000 series logic does not have good output characteristics - especially with a capacitive load. And they're dog slow. Unfortunately, you made a number of bad design choices. Now's the time to learn from your mistakes.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Thank you! That's why I am here. No, it was not thermal for sure. The FETS broke instantly.

Do you think using some additional external Schottky diodes for freewheeling is a good idea? They are quite fast aren't they?

TO THE FINDER... THE ISLE OF KOHOLINT, IS BUT AN ILLUSION... HUMAN, MONSTER, SEA, SKY... A SCENE ON THE LID OF A SLEEPER'S EYE... AWAKE THE DREAMER, AND KOHOLINT WILL VANISH MUCH LIKE A BUBBLE ON A NEEDLE... CAST-AWAY, YOU SHOULD KNOW THE TRUTH!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You didn't change the 2V supply while the 5V supply was still active, did you?

Letting the smoke out since 1978

 

 

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I may have done this... bad?

TO THE FINDER... THE ISLE OF KOHOLINT, IS BUT AN ILLUSION... HUMAN, MONSTER, SEA, SKY... A SCENE ON THE LID OF A SLEEPER'S EYE... AWAKE THE DREAMER, AND KOHOLINT WILL VANISH MUCH LIKE A BUBBLE ON A NEEDLE... CAST-AWAY, YOU SHOULD KNOW THE TRUTH!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

But the drain is not connected to 12V but to 2V.

Oh yes, now I see...please use BLACK font, not some blurry gray font & make it oriented properly, not  sideways.

Even then 5V, can be somewhat low, for this fet may require 4V just to start turning on (2+4=6V, however you only have 5).   Turing partially on is a sure way to burn up the device, internally before you may even detect a lot of heat with your finger (thought you'd probably measure some heat).

Most likely is you had 2 fets on at some point , creating a dead short...kaboom!

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

tellSlater wrote:

I am aware there are ready to use drivers. I wanted to do this for practice. There is dead time programmed in the IC. But is the logic above not enough? Should there always be dead time? The logic prevents any of the pairs to activate at the same time but does not offer dead time.

 

You always need a few tens of nanoseconds dead time between turning one fet off and the other in the totem pole on. There is a serious risk otherwise that the charge on the gate capacitance can keep both transistors on - leading to a power supply short circuit through the fets - which they're unlikely to appreciate.

 

The points others have made regarding partial turn-on are also very valid, as are those about back emf pulses from the inductors being turned off.

 

In most cases, the only real-life reason to roll your own motor driver is for production cost reduction or packaging issues (I've had to do it a time or two, and it can be critical down to the part number, family, and *manufacturer* of logic parts). Though I would not criticise your urge to learn :)

 

Neil

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Both your failed MOSFETs are on the top side, maybe this is a coincidence, or maybe it's an hint. So I will go along with what others have said and bet that the the topside MOSFETs are not turning fully on, causing them to overheat and fail.

If you look at the IRF510 datasheet, you see that the gate-source threshold voltage can be up to 4V, but in an ideal situation, where the 2V power source is fully dropped by the motor, the source of the topside MOSFET will be at 2V, so you only have 3V left to the 5V logic that is turning the FET on.

And changing to a PMOS will make it even worse, because now you only have a 2V margin to GND to turn it on. So, my opinion is that you need to replace the IRF510 by a different NMOS with lower gate-source threshold (probably about 2V max.).

 

I don't think inductive spikes are a problem, the body diodes should be able to channel spikes to GND or +2V. But if this also turns out to be a problem, you will need to parallel Schottky diodes with the body diode to handle the spikes. Some manufacturers offer MOSFETs with integrated Schottky diodes, for example: https://www.diodes.com/part/view...

 

Finally, yes, the dead time is a must, you should determine the turn on and turn off times with a scope and account for that (do actual measurements, don't just believe the timings in the datasheet that were obtained with the manufacturer's test circuit in ideal conditions).

 

edit: also, the outputs of the 4001 are pretty weak, they will probably take quite a while to charge the MOSFET gates, this will cause slow switching which is bad and may cause overheating.

Last Edited: Wed. Dec 25, 2019 - 12:55 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

"Most likely is you had 2 fets on at some point , creating a dead short...kaboom!"

The logic above prevents that for any AVR output. But activation and deactivation time intervals of the FETS are not taken into consideration.

Nevertheless, before them breaking, wave drive was used - wave drive has plenty of dead time between H bridge activations.

The heat was very low when they broke. It happened in less than a second.

Also I had the same FETS at worst conditions where I used 5V on drain and let them reach 100oC but nothing bad happened - they are rated for higher temperatures.

But yeah I get what you are telling me about them burning up inside. The truth is I had to use these cause the IRF44N I ordered are not here yet.

I would really like to know what killed these transistors...  

TO THE FINDER... THE ISLE OF KOHOLINT, IS BUT AN ILLUSION... HUMAN, MONSTER, SEA, SKY... A SCENE ON THE LID OF A SLEEPER'S EYE... AWAKE THE DREAMER, AND KOHOLINT WILL VANISH MUCH LIKE A BUBBLE ON A NEEDLE... CAST-AWAY, YOU SHOULD KNOW THE TRUTH!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

tellSlater wrote:
I would really like to know what killed these transistors...  

 

Since you measure less than 100ohm gate-source, I believe there the gate-source isolation has ben destroyed. This is a very thin layer of siliconoxide. A gate-source voltage in excess of +/- 20V, even just a few microseconds, can do that.

 

Beware, breaking current in an inductive load, even at just 2V, can create spikes at much higher voltage. Builtin diodes in mosfets may not be able to withstand that. You may need external freewheeling diodes.

 

Diodes in mosfets are there due to fabrication process, nothing else. They are generally not made for any purpose. You may use them sometimes, but they are really just a relict from fabrication process.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I'm going to go with the 'fast overheating because they were only partially turned on' theory here.  Were those IRFs actually attached to anything heat-sink-shaped, or just hangin' out in the wind?

 

If:  From a 12V rail supplying 300mA into 2.2ohms (current into the winding) and assuming the low side is fully turned on and the resistance there is negligible, then the high side MOSFET is dropping 11ish volts @ 300mA, which is more than 3 watts of power (heat).  TO-220 junctions to ambient air run about 60degC / watt, so at that point your junction temperature was about 180C - which I think you will find is a bit more than it should be.

 

S.

 

PS - I imagine you've done this, but you might want to re-confirm that your actual wiring and logic is what the schematic says it is...  smiley

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The IRF44N aren’t going to fare much better.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Transistors were at room temperature (by hand) when they broke and they broke in 1" so even at that much heat production it wouldnt have much time to surpass temperature limits. Also drain voltage is 2V and not 12V - sry for bad schematic. I did run the bridge totally shorting the PSU for 30 sec and with 5V at drain for good measure and although the transistors developed a temp of 100oC they were ok after. they limited the current themselves to 500mA during this time while PSU had a cap at 600mA. So from what I have seen this must not be a thermal issue except in existance of a sudden big thermal deviation with high temperature targeted near the broken transistor parts internaly

TO THE FINDER... THE ISLE OF KOHOLINT, IS BUT AN ILLUSION... HUMAN, MONSTER, SEA, SKY... A SCENE ON THE LID OF A SLEEPER'S EYE... AWAKE THE DREAMER, AND KOHOLINT WILL VANISH MUCH LIKE A BUBBLE ON A NEEDLE... CAST-AWAY, YOU SHOULD KNOW THE TRUTH!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Everyone thank you from the bottoms of my heart for all the help you have offered me. I am blessed to sit today for dinner with loved ones and I am very greatful. I wish you all to enjoy the people you love the most this xmas and have a good time with them. Have a great xmas :) cheers

TO THE FINDER... THE ISLE OF KOHOLINT, IS BUT AN ILLUSION... HUMAN, MONSTER, SEA, SKY... A SCENE ON THE LID OF A SLEEPER'S EYE... AWAKE THE DREAMER, AND KOHOLINT WILL VANISH MUCH LIKE A BUBBLE ON A NEEDLE... CAST-AWAY, YOU SHOULD KNOW THE TRUTH!