How to test power MOSFET / IGBT or is there a cheap tester ?

[DC1MC]

Hello everybody, after suffering today from a lack of power MOSFETs and having to cancel a nice repair ,  I've decided to stock on some. Unfortunately fleabay is ripe with fakes and I've even got a bad order from a mega-distributor (to their total forgiveness they looked into the issue, thanked me and refunded immediately).

Anyways, if some kind of puny component may cause some pain, one blown power MOSFET or IGBT can cause major pain/disaster, and kill some unobtanium parts. It could even be an unlucky lot original, not necessarily a fake.
So I need a way or device to test these components close to their nominal parameters, that means lots of volts (up to 500V )  and amps (up to 30A). In the data sheets I can see that the test circuit for non-inductive loads is pretty simple, just some resistive load and pulses to the gate. It's not a difficult thing to do, except for the f--ing high voltages and currents  !!!

So my question is: is there some kind of component tester for these things that doesn't cost as much as a car, or some kind of an working/proved project ?

 Any advice is most welcome.

 Cheers,
 DC1MC

[genghisnico13]

if thats something you do often I would probably go the DIY route. An Arduino with a MOSFET driver and an off the shelf HV power supply, and/or a low voltage and high current one, and a couple of places to hook up a scope or a DMM and you have the basics covered. This would be the most cost effective solution and seems simple enough to be able to make it reasonably fast.

[DC1MC]

Hi and thanks for the answers so far, here are some open points:

...of the shelf HV power supply...

Where can someone find such a beast that offers a bit of control and has an affordable price, so far I've only seen some kind of monsters for tube amplifiers or HAM transmitters and test instrument type things that costs as well as a car.

...use a curve tracer...

I'll gladly use one, but I've found that they either cost 1000s or  10s of 1000s and even so are mostly for bipolars, is there some kind of affordable curve tracer, a scope add-on will be great.

@blueskull - considering your profile, could you recommend a simplified practical design just stress the devices and gives a pass/don't pass characterization, I don't try here to do scientific measurements or check a semiconductor technology, I just want to make sure that I have a reliable component that will not blow and destroy more than its value.

 Thanks again for all the input,
 DC1MC

[capt bullshot]

I've never tested these for their parameters, just go / no-go testing, which can be done with a simple multimeter (using the diode test function).
For parameters, I rely on the the datasheet and a trusted source, not ebay or the like.

[Hydron]

The last time I bought dodgy semiconductors off Aliexpress (some FGY75N60SMD IGBTs) I was lucky enough to find the datasheet for the die-only version (https://shop.micross.com/pdf/fairchild/igbt/FGY75N60SM_IGBT_DIE.pdf) so I could sacrifice one to the belt sander and physically measure the die size.

Die size matched perfectly, the markings looked legitimate (in the same style as know legitimate Fairchild parts, with no sign of ground-down re-marked package) and measuring leakage current at rated voltage with a Megger also gave a value well within specs. At that point I put a couple in-circuit and gate drive waveforms (dependent on gate charge value, and thus the part being correct) were as expected, and they didn't go pop at 150A peak - happy result!

I've attached a pic of the die, an intact part and a known legitimate (Mouser sourced) Fairchild FGH40N60SMD part for reference. The engraving style is a 100% match IMHO.

[Mechatrommer]

transistor tester (https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/) has helped me much troubleshooting circuits. once i verified the mosfet still in spec, no shorted Vds or Vgs then i can continue put it back in circuit and replace the damage units for stress or real life test. if you want to test on mains, then the circuit should be simple full bridge rectifier and your switching circuit. want to be safe wrap it in bubble wrapping ten layers and solder 10m wire away to connect to the mosfet. wrap again mosfet in 10 layers bubble wrapping or non combustible materials

[DC1MC]

I've never tested these for their parameters, just go / no-go testing, which can be done with a simple multimeter (using the diode test function).
For parameters, I rely on the the datasheet and a trusted source, not ebay or the like.

How should we consider Conrad  ?

[capt bullshot]

I've never tested these for their parameters, just go / no-go testing, which can be done with a simple multimeter (using the diode test function).
For parameters, I rely on the the datasheet and a trusted source, not ebay or the like.

How should we consider Conrad  ?

I'd consider them trusted, though it's been a long time ago when I ordered something (most items are just quite expensive). If Reichelt has the part from a well known manufacturer (not the generic ones, they come from anywhere), I'd consider this trusted also. Mouser does also deliver to end users.

[genghisnico13]

Where can someone find such a beast that offers a bit of control and has an affordable price, so far I've only seen some kind of monsters for tube amplifiers or HAM transmitters and test instrument type things that costs as well as a car.

if you want High Voltage, Power and Precision, yes its gonna be expensive, or you could make a unregulated linear supply using a variac to control the input voltage (you should add an Isolation transformer or other ratio transformer to increase the maximum voltage and provide isolation).

I was thinking along the lines of High Voltage testing with low currents, with something cheap like: https://www.ebay.com/itm/DC-DC-Boost-Converter-8-32V-12V-to-45V-390V-High-Voltage-ZVS-Board-Module-NEW/332082164474?_trkparms=aid%3D555018%26algo%3DPL.SIM%26ao%3D2%26asc%3D44040%26meid%3D72e2455396454375948cd4f0432b5f8e%26pid%3D100005%26rk%3D3%26rkt%3D6%26sd%3D112515806601%26itm%3D332082164474&_trksid=p2047675.c100005.m1851.

I don't know exactly how good are the fake parts and how easily would be to spot them.

[Keysight DanielBogdanoff]

Like @blueskull said, a homegrown double pulse tester seems to be in industry go-to setup. Especially for wide bandgap stuff.

[DC1MC]

Come on guys, don't play with a poor (power electronics) newbie  , just point me to a working design that is not producing dead MOSFETS.
The Chinese power supply looks nice, I didn't quite understand how you control it, maybe by varying the input voltage ?

 Cheers,
 DC1MC

[Mechatrommer]

since the masters say so...
https://www.researchgate.net/profile/U_Dreher/post/why_the_width_of_two_pulses_is_different_in_double_pulse_test_for_IGBT/attachment/59d644eb79197b80779a01e1/AS:450858979729409@1484504565705/download/34_Double_Pulse_Test_Principle.pdf
http://powermodules.skyrock.com/3032905982-Double-Pulse-Testing-to-Switching-Features-of-Power-Module.html