Author Topic: Mosfet HV regulator : 2 multimeters fried  (Read 11640 times)

0 Members and 1 Guest are viewing this topic.

Offline MagicSmoker

  • Super Contributor
  • ***
  • Posts: 1408
  • Country: us
Re: Mosfet HV regulator : 2 multimeters fried
« Reply #25 on: June 24, 2016, 06:40:22 pm »
...
Exactly, well, kind of...
I have a bachelor degree in electronics but until now I was field service engineer for 7 years. Now I try to evolve in electronics R&D so I certainly have knowledge gaps.
Thanks for the informations !
Any alternative in this case ?

An alternative for the MOSFET you selected? I would go with an IXYS L2 series part. IXTP15N50L2 looks like a good choice and isn't too expensive:
http://www.digikey.com/product-detail/en/ixys/IXTP15N50L2/IXTP15N50L2-ND/2117346

It is rated for linear operation and has a thermal resistance from junction to heatsink (shorthand: Rth[j-s]) of 0.67C/W, which is about the same as just the junction to case resistance of the part you mentioned earlier. Max. recommended junction temperature is 150C but safe operating area is guaranteed up to 75C, so you have about 50C to work with. Referring to the FBSOA graph at the end of the datasheet, you can theoretically* handle 0.5A of current with 300V of drop across the drain-source, so it should be plenty robust for your application.

* - you should now know why I say "theoretically" ;)
 
The following users thanked this post: Rutherberg

Offline ZeTeX

  • Frequent Contributor
  • **
  • !
  • Posts: 610
  • Country: il
  • When in doubt, add more flux.
Re: Mosfet HV regulator : 2 multimeters fried
« Reply #26 on: June 24, 2016, 08:27:52 pm »
...
Exactly, well, kind of...
I have a bachelor degree in electronics but until now I was field service engineer for 7 years. Now I try to evolve in electronics R&D so I certainly have knowledge gaps.
Thanks for the informations !
Any alternative in this case ?

An alternative for the MOSFET you selected? I would go with an IXYS L2 series part. IXTP15N50L2 looks like a good choice and isn't too expensive:
http://www.digikey.com/product-detail/en/ixys/IXTP15N50L2/IXTP15N50L2-ND/2117346

It is rated for linear operation and has a thermal resistance from junction to heatsink (shorthand: Rth[j-s]) of 0.67C/W, which is about the same as just the junction to case resistance of the part you mentioned earlier. Max. recommended junction temperature is 150C but safe operating area is guaranteed up to 75C, so you have about 50C to work with. Referring to the FBSOA graph at the end of the datasheet, you can theoretically* handle 0.5A of current with 300V of drop across the drain-source, so it should be plenty robust for your application.

* - you should now know why I say "theoretically" ;)
 

Offline RutherbergTopic starter

  • Regular Contributor
  • *
  • Posts: 57
  • Country: be
Re: Mosfet HV regulator : 2 multimeters fried
« Reply #27 on: June 27, 2016, 11:02:39 am »


Thanks but could you be more specific please ?  ;)

I thought back about the worst case in my application and it should be less restrictive, about Vds = 210V and Id peak = 50mA

But assuming a more demanding application, would paralleling 2 mosfets be a good solution to reduce stress ? I assume there is precautions to dispatch current evenly between the two (or more ?) ?

Anyway, thanks for the help so far
 

Offline ZeTeX

  • Frequent Contributor
  • **
  • !
  • Posts: 610
  • Country: il
  • When in doubt, add more flux.
Re: Mosfet HV regulator : 2 multimeters fried
« Reply #28 on: June 27, 2016, 04:36:19 pm »


Thanks but could you be more specific please ?  ;)

I thought back about the worst case in my application and it should be less restrictive, about Vds = 210V and Id peak = 50mA

But assuming a more demanding application, would paralleling 2 mosfets be a good solution to reduce stress ? I assume there is precautions to dispatch current evenly between the two (or more ?) ?

Anyway, thanks for the help so far
Paralleling 2 mosfets with balancing resistors to share the current across them might help with the power dissipation but I don't think that it will do any more then helping with the power dissipation.
 

Offline MagicSmoker

  • Super Contributor
  • ***
  • Posts: 1408
  • Country: us
Re: Mosfet HV regulator : 2 multimeters fried
« Reply #29 on: June 27, 2016, 06:26:16 pm »
In high density power electronics such as solid state transformers or EV inverters, silicon power devices are pushed to 125C or higher at maximum output condition. SiC devices are pushed to 150+C, and the limitation comes from not the die, but the packaging material. With liquid cooling, a TO247 can be pushed to 150+W easily in resistive mode. ...

Yes, some semiconductors are rated for a continuous Tj of 125C or higher - and I even said as much - but even those devices will last longer if Tj is kept below 100C. Not sure what your point is here, as your argument does not refute or invalidate the general rule to keep Tj below 100C for long life.

And yes, there are *some* TO-247 parts which have a sufficiently low total Rth[j-s] so that along with a suitable heatsink (think heat-pipes, vapor phase chamber or liquid chill plate) they are able to dissipate 150W or more, but *most* TO-247 parts and *most* heatsinks are of more humble capability, and thus one finds - usually with experience - that ~50W is a reasonable maximum Pd, and even that assumes forced convection cooling of the heatsink; if relying on natural convection of even a large PCB-mounted extruded aluminum heatsink you might not be able to manage more than 20W.


Paralleling 2 mosfets with balancing resistors to share the current across them might help with the power dissipation but I don't think that it will do any more then helping with the power dissipation.

Dropping 210V at 50mA is well within the DC SOA of the MOSFET I suggested; perhaps you were predicting thermonuclear catastrophe with the MOSFET the OP suggested (IRFPE50) as its SOA does not go down to DC (10ms is the longest pulse spec'ed)?

 

Offline ZeTeX

  • Frequent Contributor
  • **
  • !
  • Posts: 610
  • Country: il
  • When in doubt, add more flux.
Re: Mosfet HV regulator : 2 multimeters fried
« Reply #30 on: June 27, 2016, 07:39:50 pm »
In high density power electronics such as solid state transformers or EV inverters, silicon power devices are pushed to 125C or higher at maximum output condition. SiC devices are pushed to 150+C, and the limitation comes from not the die, but the packaging material. With liquid cooling, a TO247 can be pushed to 150+W easily in resistive mode. ...

Yes, some semiconductors are rated for a continuous Tj of 125C or higher - and I even said as much - but even those devices will last longer if Tj is kept below 100C. Not sure what your point is here, as your argument does not refute or invalidate the general rule to keep Tj below 100C for long life.

And yes, there are *some* TO-247 parts which have a sufficiently low total Rth[j-s] so that along with a suitable heatsink (think heat-pipes, vapor phase chamber or liquid chill plate) they are able to dissipate 150W or more, but *most* TO-247 parts and *most* heatsinks are of more humble capability, and thus one finds - usually with experience - that ~50W is a reasonable maximum Pd, and even that assumes forced convection cooling of the heatsink; if relying on natural convection of even a large PCB-mounted extruded aluminum heatsink you might not be able to manage more than 20W.


Paralleling 2 mosfets with balancing resistors to share the current across them might help with the power dissipation but I don't think that it will do any more then helping with the power dissipation.

Dropping 210V at 50mA is well within the DC SOA of the MOSFET I suggested; perhaps you were predicting thermonuclear catastrophe with the MOSFET the OP suggested (IRFPE50) as its SOA does not go down to DC (10ms is the longest pulse spec'ed)?
yep, well not thermonuclear explosion but something a little bit smaller.
 

Offline RutherbergTopic starter

  • Regular Contributor
  • *
  • Posts: 57
  • Country: be
Re: Mosfet HV regulator : 2 multimeters fried
« Reply #31 on: June 29, 2016, 10:07:50 am »
yep, well not thermonuclear explosion but something a little bit smaller.

You are probably right, at Vds = 300V and Ic quiescent = 10mA the case of the IRFE50 is already at 50°C. The mosfet is fixed on a big aluminium chassis with a milar film in between.
I will order an IXYS L2 and see how it reacts.

About paralleling mosfets, the balancing resistors in this case should be connected at the source of each mosfets, right ?
Is there a rule of thumb about the value of these resistors ?

Thanks
 

Offline bktemp

  • Super Contributor
  • ***
  • Posts: 1616
  • Country: de
Re: Mosfet HV regulator : 2 multimeters fried
« Reply #32 on: June 29, 2016, 10:28:12 am »
You are probably right, at Vds = 300V and Ic quiescent = 10mA the case of the IRFE50 is already at 50°C. The mosfet is fixed on a big aluminium chassis with a milar film in between.
That sounds wrong: 300V*10mA=3W. If the case reaches 50°C by dissipating only 3W, then there must be a good thermal insulator between the mosfet and the heatsink. Using a different mosfet will not change much. A better heatsink is more important than going to a larger mosfet.

Quote
About paralleling mosfets, the balancing resistors in this case should be connected at the source of each mosfets, right ?
Is there a rule of thumb about the value of these resistors ?
The voltage drop on the source resistor should be much larger (10x) than the gate threshold voltage change by temperature. This translates to a couple of volts voltage drop on the source resistor.
« Last Edit: June 29, 2016, 10:33:06 am by bktemp »
 

Offline RutherbergTopic starter

  • Regular Contributor
  • *
  • Posts: 57
  • Country: be
Re: Mosfet HV regulator : 2 multimeters fried
« Reply #33 on: June 29, 2016, 11:29:21 am »
...then there must be a good thermal insulator between the mosfet and the heatsink. Using a different mosfet will not change much. A better heatsink is more important than going to a larger mosfet.

The junction between the mosfet and the chassis is not optimal, the milar film is pretty thick and the screw is nylon type so not really tight.

Meanwhile I found an alternative candidate to the IXYS L2 : http://www.farnell.com/datasheets/1863418.pdf
DC is plotted on the FBSOA curves, the thermal resistances are pretty similar and it's almost 10 times cheaper. Thermonuclear explosion or not ?  ;)
 

Offline bktemp

  • Super Contributor
  • ***
  • Posts: 1616
  • Country: de
Re: Mosfet HV regulator : 2 multimeters fried
« Reply #34 on: June 29, 2016, 11:49:16 am »
Meanwhile I found an alternative candidate to the IXYS L2 : http://www.farnell.com/datasheets/1863418.pdf
DC is plotted on the FBSOA curves, the thermal resistances are pretty similar and it's almost 10 times cheaper. Thermonuclear explosion or not ?  ;)
It is probably as good as IRFPE50.
My rule of thumb: If the power dissipation does not decrease at high voltages, the SOA curve is bullshit. And that is the case in this datasheet: No way this mosfet can reliably dissipate more than 200mA at 900V. The junction to case thermal resistance is 0.63°C/W. So if you dissipate 198W @ 25°C Tcase, the junction will be at 150°C. That's where the 198W specification comes from. They probably didn't do any SOA test, they simply calculated the values.

30W power dissipation at 900V is a more realistic value. At 300V probably 50W, but you can only guess. Same as for IRFPE50.
It most likely works (if you improve your heat sink), but you can't know for sure.
In contrast IXYS probably did a lot of destructive SOA tests to give you a reliable SOA curve. That's why IXYS parts are more expensive than other manufacturers.
« Last Edit: June 29, 2016, 11:51:25 am by bktemp »
 

Offline tautech

  • Super Contributor
  • ***
  • Posts: 28141
  • Country: nz
  • Taupaki Technologies Ltd. Siglent Distributor NZ.
    • Taupaki Technologies Ltd.
Re: Mosfet HV regulator : 2 multimeters fried
« Reply #35 on: June 29, 2016, 11:58:58 am »
...then there must be a good thermal insulator between the mosfet and the heatsink. Using a different mosfet will not change much. A better heatsink is more important than going to a larger mosfet.

The junction between the mosfet and the chassis is not optimal, the milar film is pretty thick and the screw is nylon type so not really tight.
Might I suggest that's a large part of your problem?
Get a mounting kit with metal nuts and bolts and quality silicone heatsink pads.
Avid Rabid Hobbyist
Siglent Youtube channel: https://www.youtube.com/@SiglentVideo/videos
 

Offline RutherbergTopic starter

  • Regular Contributor
  • *
  • Posts: 57
  • Country: be
Re: Mosfet HV regulator
« Reply #36 on: July 26, 2016, 04:06:48 pm »
Hi,

After few tests of the implemented system (attached schematic), the FQA6N90C seems to be OK, even in the worst case it never exceed 70°C.

Now I'm facing an other weird problem. When the amplifier is pushed, so when the current increases, the system tend to oscillate.
The result : when the signal peaks, I can hear a loud hum/hard clipping (at 100Hz..?) in the sound. It's getting worse when the switch is on the low position (so when Vds is the bigger).
The peak current is well below the current limitation of R67/D5 so it is not the problem. I also tried without C40, no change.

The weirdest : when I connect my DMM between GND and the positive terminal of C24, it works like a charm, no oscillations and weird noises.

I suspect a layout problem, the oscillation created by the radiance of another component but... Am I missing something else ?

Thanks
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf