1

I'm using this depletion-mode MOSFET in a high-voltage power-management circuit. My 500V DC supply isn't high grade, and now that I've blown two of these MOSFETs (gate shorts to drain) I'm wondering if I can put any components on either side to protect it? I was getting weird results with the first one before realizing it was blown. I might have blown the second one (not sure at which point in my testing it failed) simply by turning the high-voltage supply on with it wired like this:

schematic

simulate this circuit – Schematic created using CircuitLab

For protection purposes: If it simplifies things I never want to see more than 200mA through the circuit. If no better ideas I guess I could put a fast-blow fuse in front of V+? But I'm not sure if that's the only thing that can blow this MOSFET. E.g., I'm beginning to wonder if having no load on the drain can cause problems.

feetwet
  • 2,322
  • 9
  • 31
  • 55
  • What is this schematics? I don't see any reason why this shoudn't blow off. – Marko Buršič May 10 '16 at 18:11
  • @MarkoBuršič - Explain why? Lack of load on drain? I couldn't find anything in the specs constraining drain voltage other than to be less than 1000V over source. And the power resistor on source should limit current to 5A even if the MOSFET isn't adding any resistance (which, according to its spec sheet, it should). **NB: The MOSFET is heat-sinked.** – feetwet May 10 '16 at 18:20
  • Please explain what you want to acheive, maybe I am wrong, but this circuit can just blow your MOSFET and nothing else. Perhaps you will get some information to modify this, if you specify the purpose. – Marko Buršič May 10 '16 at 18:24
  • Goal is to provide *constant current* to a varying-resistance load on the drain side. Which it does quite well at low voltage. It also does it at high voltage until something blows the MOSFET. The circuit has to be able to survive an effectively zero-resistance load (i.e., a short) on the drain to V+. If that would be expected to blow the MOSFET I would be interested to know why, but even more interested to know how I can protect and recover the circuit from that condition. – feetwet May 10 '16 at 18:38
  • Few notes: 0.2A*1000 Ohm=200V, A Vgs of -200V is just too much. Second, Vds =500-200=300V, P_dissipating=300*0,2=60W, which is also too much. Not sure about the schematics, if this circuit can really work as constant current source, don't know where you got it. – Marko Buršič May 10 '16 at 18:53
  • @MarkoBuršič - Yes, I have this heat-sinked to support 60W dissipation. My understanding is that *the whole purpose of the MOSFET is to add resistance so that Vgs doesn't exceed -5V*. Since the breakdown voltage is 1000V why would it fail to do that in this circumstance? – feetwet May 10 '16 at 19:06
  • I don't really see how the MOSFET get blown since the current is limited to a few mA when Vgs becomes reverse bias by a few volts. Perhaps it has to do with an in-rush current to the G-D junction when 500V suddenly comes on. Can try adding a resistor to the gate, perhaps with value of 1 to 10K ohms. And perhaps do it in conjunction of what Wojciech suggested to limit Vgs with zener diodes. – rioraxe May 11 '16 at 01:09
  • @rioraxe - Exactly my thinking (first part of your comment). And as I was turning this over in my head I actually thought, "In-rush Current Limiter ... is this one of those situations that calls for an ICL?" Regarding your second suggestion: Since no current is supposed to actually flow through the gate I may as well put a resistive barrier in the mega-ohm range on that contact, right? – feetwet May 11 '16 at 02:54
  • @feetwet If the gate resistor is in the megaohm range, the initial charge would drain slowly, leading to a significant positive Vgs. I am not sure what is the best value. – rioraxe May 11 '16 at 18:52

3 Answers3

8

The common way to protect a MOST gate is to use a Zener diode between gate and source. Your MOST has a max Vgs of 20V so add a ~15V zenned diode in reverse bias and you'll be fine. There are also devices called transils, which are specialized for over-voltage protection but do basically the same.

enter image description here

  • So if the MOST failed with a short between gate and drain then it must have been caused by excessive voltage difference between gate and source? – feetwet May 10 '16 at 18:15
  • If you get a G-D short it will kill the MOST and the Zener diodes with it. Maybe when connecting the transistor the gate caught some static or got charge through capacitive coupling and the voltage blew the gate oxide. Hard to tell what are you trying to do with this circuit, but do try the protection diodes first. They may point you to the source of the problem –  May 10 '16 at 18:32
  • I have only *observed* that the malfunctioning MOSTs have a G-D short. I.e., I take it as evidence that the MOST is blown. But I don't know *why* or *how* it happened. I just [added clarification of the circuit's purpose in question comment](http://electronics.stackexchange.com/questions/233394/how-can-i-protect-my-mosfet/233398#comment507327_233394). – feetwet May 10 '16 at 18:43
  • So I want a Zener or TVS to shunt current when Vgs exceeds ~15V. I.e., "clamping" voltage should be under 15V? And, presumably, all other voltage specs for such a diode will be under 15V so any such will do, and in either orientation? Even though this circuit could present a voltage difference in the hundreds of volts? Or is there some other specification I need to watch? – feetwet May 10 '16 at 22:08
5

This small Semtech TVS Diode Application Note called "Transient Protection of MOSFETs" provides detailed recommendations on employing TVS diodes to protect both the more sensitive gate and the entire MOSFET against transients, by connecting the diodes like this:

MOSFET Protection circuit

SamGibson
  • 17,231
  • 5
  • 37
  • 58
feetwet
  • 2,322
  • 9
  • 31
  • 55
  • @uhoh – I just updated the link, but the essence of the recommendations are summarized in the circuit diagram. – feetwet Mar 18 '18 at 03:22
  • Link is broken again. – maxschlepzig Feb 19 '22 at 20:43
  • 1
    @maxschlepzig - Hi, I couldn't find it any longer on the original website. However I found it on Digikey, so I updated the link to point there. – SamGibson Feb 19 '22 at 21:51
  • @SamGibson thank you. Ok, just in case, for future reference, the linked document's meta data: Semtech, SI96-13, Surging Ideas, TVS Diode Application Note, Revision 9/2000 – maxschlepzig Feb 19 '22 at 21:58
3

Just wanted to add, for future people viewing this, why the original circuit was failing. If the transistor turns on more than a little then a large portion of the supply voltage is across the resistor, causing Vgs to be a high negative voltage. In ideal conditions this would never happen but was likely occurring due to either pulse effects when turning on, or it began oscillating at some point. This could be avoided with a small capacitor from gate to source and a resistor, if this current limiter does not need to respond very quickly these can be oversized, making the selection easier. And still throw a zener in there just in case.

This circuit should both protect the gate and reduce the chance of oscillation:

schematic

simulate this circuit – Schematic created using CircuitLab

TWiz
  • 353
  • 1
  • 6
  • 2
    From my own experience with Zener diodes being used as protection against transients, don't use Zeners rated for less than 1W. I have checked the pulse response of 1/2W Zeners vs. 1W Zeners, and the 1/2W ones let the voltage overshoot happen, while the 1W ones cut it off. (I know it's an old question/answer, but I had to share this for others) – Edin Fifić Aug 29 '20 at 19:52