Author Topic: Help with Schottky datasheet thermal specs  (Read 817 times)

0 Members and 1 Guest are viewing this topic.

Offline PeabodyTopic starter

  • Super Contributor
  • ***
  • Posts: 2248
  • Country: us
Help with Schottky datasheet thermal specs
« on: May 12, 2020, 06:45:28 pm »
I would like to use a Schottky diode in a circuit where reverse leakage current could be important because the leakage current would flow into a mosfet gate and cause mischief.  And I understand that leakage increases exponentially with the diode's temperature.  So the question is - at the maximum current that will be flowing through the diode just before it becomes reverse-biased, how hot will the diode be.  If I know that, I think I can ballpark what the leakage current would be at that moment.

I'm looking at the datasheet for the 1N5820-22 diodes

https://www.onsemi.com/pub/Collateral/1N5820-D.PDF

and trying to understand the thermal data.  Looking at Figure 6, it says that at 1 amp DC current, power dissipation is 0.4W.  I assume that means the forward voltage would be 0.4V.  I think Note 4 and Note 5 may have the information I want, but I'm at loss to understand them.  The closest I can get is that the highest value in Note 5 is 63 degrees C per watt.  Does that mean that my 0.4W dissipation would raise the temperature from 25 to 50.2 degrees (25 + (.4 x 63))?  Am I anywhere near right?

Well, if anyone can explain this, or point me to an explanation, that would be great.
 

Offline Vovk_Z

  • Super Contributor
  • ***
  • Posts: 1478
  • Country: ua
Re: Help with Schottky datasheet thermal specs
« Reply #1 on: May 12, 2020, 07:56:21 pm »
I haven't seen the datasheet now but as I know at maximum forward current (3A for 1N5820) cristal temperature can be about 100-125 C or higher (with +20..+25 C ambient).

Your calculation is right for DC. If current is pulsating then diode will be hotter (as in fig.6).
 

Offline Etesla

  • Regular Contributor
  • *
  • Posts: 154
  • Country: us
Re: Help with Schottky datasheet thermal specs
« Reply #2 on: May 12, 2020, 08:04:07 pm »
Could you just put a high value resistor from the gate to the source of the mosfet to drain that leaking current away from the gate?
 

Offline magic

  • Super Contributor
  • ***
  • Posts: 7454
  • Country: pl
Re: Help with Schottky datasheet thermal specs
« Reply #3 on: May 12, 2020, 08:07:49 pm »
I think Note 4 and Note 5 may have the information I want, but I'm at loss to understand them.  The closest I can get is that the highest value in Note 5 is 63 degrees C per watt.  Does that mean that my 0.4W dissipation would raise the temperature from 25 to 50.2 degrees (25 + (.4 x 63))?  Am I anywhere near right?
Yes, that's how it works, assuming that the 1A current flows long enough to heat up the device from 25°C to 50°C.
Observe that 63°C/W applies only to mounting method 2 with maximum lead length. Mounting method 1 with clipped leads is a bit more efficient and mounting method 3 shows that further improvement is possible by adding a copper plane around the diode and mounting it in a particular orientation.
Beware that air near your PCB may get hotter than ambient.

Note 4 shows how to model heat flow using electric equivalents: heat generation is replaced with a current source and thermal resistance is replaced with resistance, producing voltage drops which represent temperature gradients.
« Last Edit: May 12, 2020, 08:10:06 pm by magic »
 

Offline PeabodyTopic starter

  • Super Contributor
  • ***
  • Posts: 2248
  • Country: us
Re: Help with Schottky datasheet thermal specs
« Reply #4 on: May 12, 2020, 10:28:33 pm »
Could you just put a high value resistor from the gate to the source of the mosfet to drain that leaking current away from the gate?

Yes, but it's a P-channel mosfet, so the resistor is to ground so the mosfet will turn on when the Schottky path is turned off.  Leakage would raise the gate voltage, which could prevent the mosfet from turning on fully.  But yes, the ultimate question is the value of that resistor to make sure leakage doesn't mess things up.  I mean, I could just make it 1K, but I hope something a little less wasteful will work.

I'll attach the circuit which currently calls for a regular rectifier diode.  For a variety of reasons, it needs to be a Schottky.  Leakage may actually not be a problem at all.  If the mosfet doesn't turn on, then there isn't going to be much voltage on the cathode to leak through in the first place.  Anyway, I just need the regulator to switch over to battery power from USB power with no interruption, so the Arduino being powered won't reset.  It's supposed to act like a UPS.
 

Offline PeabodyTopic starter

  • Super Contributor
  • ***
  • Posts: 2248
  • Country: us
Re: Help with Schottky datasheet thermal specs
« Reply #5 on: May 12, 2020, 10:40:54 pm »

Yes, that's how it works, assuming that the 1A current flows long enough to heat up the device from 25°C to 50°C.


Yes, I found a Great Scott video where he actually puts 1 amp through a 1N5819 and measures its temperature at 50 degrees.  That's with the diode not mounted to anything except alligator clips.  So I'm in the ballpark.

Quote
Observe that 63°C/W applies only to mounting method 2 with maximum lead length. Mounting method 1 with clipped leads is a bit more efficient and mounting method 3 shows that further improvement is possible by adding a copper plane around the diode and mounting it in a particular orientation.
Beware that air near your PCB may get hotter than ambient.

This is part of a modification to an existing circuit.  So there's no mounting as such.  The diode leads will just be soldered to existing pins and the diode body will be lying on the PC board.  So thermally it's not the best of situations.  But it looks like leakage will not be s big problem in the ballpark of 50 degrees C.  Julian Ilett has a video where he aims his reflow gun at a Schottky diode and measures leakage of over 200 mA.  I just wanted to be sure I wasn't looking at that kind of situation.
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Help with Schottky datasheet thermal specs
« Reply #6 on: May 13, 2020, 12:30:40 am »
Why use a huge-ass power diode for a small gate signal?

If it's supposed to bypass (when the USB side is active I guess?), two could be used, one for the gate and one for the bulk power path.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline PeabodyTopic starter

  • Super Contributor
  • ***
  • Posts: 2248
  • Country: us
Re: Help with Schottky datasheet thermal specs
« Reply #7 on: May 13, 2020, 04:04:45 am »
Yes, current from the USB power input will power the entire load through the Schottky diode.  That's why it's hefty.  I'll have to think about a second diode for the gate.  I'm not sure it would change anything as far as leakage is concerned.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf