Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

Measurement of Mosfet Losses

Status
Not open for further replies.
I

ionp

Member level 3
Member level 3
Joined
May 26, 2006
Messages
65
Helped
8
Reputation
16
Reaction score
4
Trophy points
1,288
Location
USA
Activity points
1,780
Hi.

I am trying to track down the cause of some MOSFETS heating more than previous "identical" devices, and am looking for a reliable way to measure losses in-circuit.

My circuit is something like an H-bridge, with 340VDC fed to the legs and 75A peaks through devices. The frequency is up to 400kHz and operation is into a resonant load and discontinuous.

Measuring device current without introducing too much inductance and measuring small on-voltage levels with limited common mode rejection are my main enemies.

If anyone could share technical papers containing relevant techniques to avoid pitfalls in these areas, I could save whats left of my hair from all the head scratching.

Thanks in advance!
 

For current measurement, small Rogowski probes like **broken link removed** or self elaborated designs are a good option. For the voltage measurement, high speed differential active probes and possibly special Vds clamp circuits for on-state voltage measurements are required. Regular passive probes with toroid cores as common mode choke can work in some cases, too.
 
  • Like
Reactions: ionp

    I

    ionp

    Points: 2
    Helpful Answer Positive Rating
FvM said:
For current measurement, small Rogowski probes like **broken link removed** or self elaborated designs are a good option. For the voltage measurement, high speed differential active probes and possibly special Vds clamp circuits for on-state voltage measurements are required. Regular passive probes with toroid cores as common mode choke can work in some cases, too.
Click to expand...

Thanks, FvM!
I was unaware of these, and will benefit from knowing about them now.

Anyone know of voltage probes that will do better than the Tektronix P5205 for common mode for this?
I still want to measure ON voltage across the mosfets in the midst of 600V peaks when off.
 

i actually did this once, for 600V bridge with 100A throughput. i used rogowski coil as FvM suggest (the green sense wire comes in all sizes). it only measures AC current, so keep that in mind, but it has huge bandwidth.

the voltage measurement is tricky.. especially if you want accuracy. the voltage is dynamic and in a high noise environment and floating with respect to ground. so.. tough. what I ended up doing was designing a gate driver board with a voltage measurement through fast op-amps with high noise rejection. I had a 5V processor on board (high side) which could communicate through fiber optics.. so that's how I got the data out. also there are some tricks such as blocking the sense until AFTER the voltage has switched and the device has fully saturated (i.e. don't try to measure when the noise is the highest). but this is not your typical solution, definitely a custom job.

perhaps your solution is more easily found? you want to "track down the cause of some MOSFETS heating more than previous "identical" devices"

post the link to the datasheet. the answer is probably there.

Mr.Cool
 

Thanks for the thoughts, Mr. Cool.
That sounds like considerable development time and I really hope it is beyond what I will need to do here!
Here is the data sheet for the devices : **broken link removed**
 

If you are having problems with new device, assuming it has lower Rds-ON spec, it probably means your driver is not capable of supplying the greater gate capacitance of new device. 265 nC of Qg is pretty hefty device. Compare that to previous part used.
 

I may have not been clear. The part number and manufacturer of the mosfet is unchanged.
 

ionp said:
I may have not been clear. The part number and manufacturer of the mosfet is unchanged.
Click to expand...

Hi ionp,

are you repairing the device? try to see using an oscilloscope if what happens with the waveform.
 

hmm... now i am confused.

you said:
I am trying to track down the cause of some MOSFETS heating more than previous "identical" devices
Click to expand...
which implies that the MOSFET has changed

and now you say:
The part number and manufacturer of the mosfet is unchanged.
Click to expand...
which implies the MOSFET has NOT changed

?
 

Sorry for the confusion!

The design has not changed.
About a month ago, new units built to the same specs started running hotter, as measured by a thermister glued to the top surface of one of the mosfets.
So, the heating appears on different complete units but the mosfets involved are of the same manufacturer and device type as are the drivers and all the other relavant electronics, compared to those built in the past.

The waveforms are close to identical, but I have not had a good way to measure in-circuit.
I have one of the probes recommended by FvM on order, which may clarify things when I can take data with it.


I appreciate the help and will try to be clearer.
 

Lot to lot, there can be an almost two to one tolerance to gate threshold which also translate to Vgs vs. Rds-ON.

If you have marginal Vgs drive this may account for the difference.
 

The driver outputs swing to 12.6V.
At least one power module exhibiting the problem had a higher voltage swing on the gates than the properly functioning reference unit.
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top