Double Pulse Test question

[senmeis]

Hi,

I’m reading this article to see how to Test GaN, SiC MOSFETs or IGBT Devices: https://www.infineon.com/dgdl/Infin...N.pdf?fileId=5546d46271bf4f920171ee81ad6c4a1f

Is this Double Pulse Test useful for all the use cases? For example in Doherty PA two power transistors (or MOSFETs) are not connected as high and low side pair. Does DPT make sense for this architecture?

 

[FvM]

Double pulse test is evaluating saturated switching behaviour. Not relevant for RF.

 

[senmeis]

Thanks. Do you mean DPT is only applicable for LF circuits such as control circuits?

 

[FvM]

For switching stages in power electronics

 

[D.A.(Tony)Stewart]

It makes sense to use DPT methods for measuring margins to SOA failure specs with fast high V*I transitions in class D to F amplifiers.

https://www.rohde-schwarz.taipei/data/activity/file/1688113270495955616.pdf

 

[FvM]

To apply double pulse test, device Vds and Id must be measured in time domain. I guess that's more or less impossible for RF amplifiers.

 

[senmeis]

Thanks.

 

[D.A.(Tony)Stewart]

Double pulse test is evaluating saturated switching behaviour. Not relevant for RF.

Digital switches can be used for RF and DPT with the right equipment in RF-switched topology is possible.

https://cdn.teledynelecroy.com/files/pdf/equipment-list-dpt.pdf

RF Power Amplifiers: An Overview

Learn about RF power amplifiers, amplifier classes, advanced topologies, and recent technological advancements in wireless communication systems

resources.pcb.cadence.com

 

[FvM]

The current probes discussed in LeCroy note expose maximal 200 MHz bandwidth (Pearson 8585C) respectively 100 MHz (fastest PEM Rokowski coil), no chance to perform double pulse test for RF operation frequencies.

Even if you consider Tek 1 GHz CT1, it's too large to probe drain current of a RF transistor.

 

[mtwieg]

Short answer is no, double pulse testing is not applicable to RF circuits.

Single and double pulse test is for investigating issues specifically related to hard switching (diode reverse recovery, overshoot, snubber networks, etc). It's meant to allow the peak power level to be very high while keeping average power level extremely low, meaning you can see problems clearly before they have a chance to do actual damage.

The long answer....

A similar sort of testing principle can be applied to RF power circuits. For example, one could apply the DC bias and RF power to an RFPA in very brief pulses while testing in order to troubleshoot while reducing the chances of actual device damage. I did such testing on some LDMOS PAs to ensure the bias was thermally stable. But aside from the general principle, the details of how measurements are performed will be entirely different at "real" RF frequencies (>10MHz roughly).