Author Topic: The importance (or not) of gate current measurement  (Read 1059 times)

0 Members and 1 Guest are viewing this topic.

Offline bdunham7

  • Super Contributor
  • ***
  • Posts: 2685
  • Country: us
Re: The importance (or not) of gate current measurement
« Reply #25 on: June 13, 2021, 08:41:29 pm »
Alas, no readable links!  Any summaries of the content anywhere?

I reread your original post and it appears I stopped reading after 'mosfets', so I'll amend my original answer--in the case of high power IGBTs, maybe.  I knew someone that worked at a company that developed an early high power resonant soft switching system using IGBTs over two decades ago, when IGBTs were new-ish.  I don't know if the final product had any closed loop controls for voltage or current, but I know that gate control was pretty critical and they blew up a fair amount of stuff in testing.  At that level, research in development of a new product using new methods, perhaps current would give you some information that voltage won't--especially if you don't yet fully understand the characteristics of the gate.

For regular MOSFETS, I still think the main goal would be getting the gate from the threshold voltage to the saturation voltage (or close to it) as quickly as possible without causing other problems--and that you can determine by measuring the voltage directly. 

A 3.5 digit 4.5 digit 5 digit 5.5 digit 6.5 digit 7.5 digit DMM is good enough for most people.
 

Online MasterTech

  • Super Contributor
  • ***
  • Posts: 1115
  • Country: 00
Re: The importance (or not) of gate current measurement
« Reply #26 on: June 14, 2021, 06:02:41 am »
Updated the post with links

I reread your original post and it appears I stopped reading after 'mosfets', so I'll amend my original answer--in the case of high power IGBTs, maybe.  I knew someone that worked at a company that developed an early high power resonant soft switching system using IGBTs over two decades ago, when IGBTs were new-ish.  I don't know if the final product had any closed loop controls for voltage or current, but I know that gate control was pretty critical and they blew up a fair amount of stuff in testing.  At that level, research in development of a new product using new methods, perhaps current would give you some information that voltage won't--especially if you don't yet fully understand the characteristics of the gate.

One of the articles talks about controlling drain current via gate charge in IGBTs
 

Offline bdunham7

  • Super Contributor
  • ***
  • Posts: 2685
  • Country: us
Re: The importance (or not) of gate current measurement
« Reply #27 on: June 16, 2021, 05:32:46 am »
One of the articles talks about controlling drain current via gate charge in IGBTs

I can only read the abstracts, but I get the basic ideas.  I'm not sure what you mean by the above, but my understanding of the reason for the resonant soft-switch system in the system I referred to is to make it so that the IGBT transitions occur at low voltage or current points, and that is because once you have optimized your gate drive and made sure that all the elements in the large parallel power IGBTs are fully on, you are left with the device-inherent transition times as the dominant factor.  Perhaps IGBTs have improved in this regard over the past few decades.

https://patentimages.storage.googleapis.com/ad/2a/db/1cd8d1377a464d/WO2017201209A1.pdf    (these were very high power AC to AC conversion systems, like 480VAC 3PH 100kVA)

As for the articles, the first looks like they are researching and characterizing SiC devices by gate current measurement.  I'd still wonder what the gate voltage looks like, but I suppose that's probably in there.  The second looks like they're optimizing the gate drive for lower gate drive losses "with little sacrifice in switching time" using active current control and feedback.  If you are designing a new gate driver IC, that makes sense.  The third looks the most interesting because they seem to be directly addressing the issue of uneven transitions among parallel elements with some current sensing and integration scheme.  I'm not sure whether those are discrete elements or they have some method for doing this with existing large commercial parallel element high-power devices. 

Anyway, yes at this level of analysis, current measurements make sense.  Just don't expect to do it on the cheap! 
A 3.5 digit 4.5 digit 5 digit 5.5 digit 6.5 digit 7.5 digit DMM is good enough for most people.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf