Chinese MOSFETS/IGBTS

[Fumeaux]

EDIT: Just to clear things up before anyone complains: I'm just talking about the 10 piece 10$ aliexpress mosfets, don't want to offend anyone

Hello Guys!

Due to recent events, I was wondering what your thoughts on chinese mosfets are.

In short summary: An IRFP460 from a normal source costs around 3€ (for me). The alternative are "IRFP460"s from china which you can get for 1€ each (10 for 10€).

The thing is that I have noticed that the chinese mosfets die alot more often and at lower loads (in contrast to the secifications). The same goes for my FGA60N65SMD, that died twice just today, which they shouldn't have. (halfbridge, 30V, 1A draw, 130kHz)

I just ordered real IGBTs and was wondering if these fake mosfets and igbts can be used for anything else or are those completly useless? At least I didn't lose too much money.

[profdc9]

You might be able to use them as low current/low voltage devices, but it's too easy to take a small die and put it in a big package.  Big dies cost more but are the ones that have the high voltage and/or current ratings.  Often you can find substitutions of fakes where the substitute superficially behaves like the real part, for example, LM358 dual op-amp for TL082, or one pin compatible logic family for another.  Sometimes, the substitutions are rejects of the real part that don't meet specifications, but might be sufficient for hobbyist purposes.  Occasionally, the parts are other parts relabeled which is easy to find.  But things like high power MOSFETS/IGBTS are almost invariably fakes because its too easy to substitute small die parts for large die parts.

You might try lcsc.com which might save you some money over digikey/mouser.  I think they have more to lose by counterfeiting because they are trying to establish a reputation as a parts supplier.

Dan

[Hydron]

I have had some luck with Aliexpress IGBTs here:
https://highvoltageforum.net/index.php?topic=193.0

Note that I did quite a bit of testing and checking before ordering more than one batch - I'm sure there are lots of nasty surprises out there!

You may also be able to apply some of the techniques I used in that post to your FGA60N65SMD, including the die size check - the die-only datasheet can be found here: https://shop.micross.com/pdf/fairchild/igbt/FGA60N65SM_IGBT_DIE.pdf

[Thunderstruck]

I wanted to ask the same question, for example IXGN60N60C2D1 IGBT’s are listed as obsolete from Digi-Key and Mouser, but there is plenty available on eBay and very cheap too.
A lot of stuff on eBay is counterfit, so I suspected that the same would apply in this case.
I also looked at some large caps, good price, but after a bit of searching I found that they are 100% fake - identified as fake by the lettering, units used for capacitance, etc.

Usually, if it sounds too good to be true, it is too good to be true.

For example, look at this thing, AUD $14, judging by the marks on the metal parts, it is made by a blacksmith 

[flyrod]

...wondering if these fake mosfets and igbts can be used for anything else or are those completly useless? At least I didn't lose too much money.

I've been re-ordering some chineze FETs off ebay, halfway just for the cheap entertainment value.  Most vendors will just refund your money when you point out that they're committing fraud, so all it costs is time in waiting for the shipping.  I'm at about 8% success rate in getting parts that match the specs you would expect with the way the case is labeled.  Most are fake, but I've seen a pretty good variety.  Some are simply trash, some are actually useable if you don't really need the spec it should have.  My advice: look for vendors with free or no-BS returns, expect fake parts, expect to wait for shipping, test whatever they send you.

These FETs all had the same label on their cases:

[profdc9]

I wonder if there's some simple tests that could be built into a device for identifying MOSFETs and IGBTs, kind of like the GM328 component tester that you can get on ebay cheaply.

For example, without really high voltages/currents being applied, one can look at:

1.  Gate capacitance with a grounded source/emitter
2.  Threshold voltage
3.  On/off time could be done by sending a periodic square wave to the gate, and then putting the drain voltage and gate voltage into a logic gate.  Using an AND logic gate measures the on-time which is measured using the duty cycle of the time that both the gate and drain are high logic level, and the off time is measured by a NOR gate, which measures using the duty cycle the time the gate and drain are both low logic level.
4.  By using a current sense resistor, the phase lag between the collector or drain current and the emitter-collector or source-drain voltage can be measured so the switching losses can be estimated, again using the duty cycle measured using a logic gate.
5.  Collector-emitter saturation voltage for an IGBT or Rds for a MOSFET.
6.  Body diode or reverse diode forward voltage.

To avoid damaging the device and to keep the costs of such a device down, I am avoiding anything that directly measures the maximum voltages or currents a device can handle.

A cheap device to identify fakes could help alleviate the problem.

[Mads Barnkob]

I wonder if there's some simple tests that could be built into a device for identifying MOSFETs and IGBTs, kind of like the GM328 component tester that you can get on ebay cheaply.

For example, without really high voltages/currents being applied, one can look at:

1.  Gate capacitance with a grounded source/emitter
2.  Threshold voltage
3.  On/off time could be done by sending a periodic square wave to the gate, and then putting the drain voltage and gate voltage into a logic gate.  Using an AND logic gate measures the on-time which is measured using the duty cycle of the time that both the gate and drain are high logic level, and the off time is measured by a NOR gate, which measures using the duty cycle the time the gate and drain are both low logic level.
4.  By using a current sense resistor, the phase lag between the collector or drain current and the emitter-collector or source-drain voltage can be measured so the switching losses can be estimated, again using the duty cycle measured using a logic gate.
5.  Collector-emitter saturation voltage for an IGBT or Rds for a MOSFET.
6.  Body diode or reverse diode forward voltage.

To avoid damaging the device and to keep the costs of such a device down, I am avoiding anything that directly measures the maximum voltages or currents a device can handle.

A cheap device to identify fakes could help alleviate the problem.

How about a "precision" scale? Properly not as cheap, but if you can sort fakes from real with a bigger/heavier die in seconds, it sure would speed up the process. I am not sure, but I assume there must be a weight difference between package plastic and silicium/bonding to heat sink.

[flyrod]

The differences in the copper backs will outweigh the differences in the silicon.  Have you guys seen these cheap little testers?:



From here:

https://www.ebay.com/itm/172557508334

That is a quick, cheap, non-destructive way to weed out the fake FETs.  The Rds number is usually off by a fair amount, so to further test I put 10A through one and measure the voltage drop.

[profdc9]

Yes that is the GM328.  I have one of those as well.  They're excellent little diagnostic devices.  I was thinking of something like this, but more specialized to power MOSFETs and IGBTs.

Dan

The differences in the copper backs will outweigh the differences in the silicon.  Have you guys seen these cheap little testers?:



From here:

https://www.ebay.com/itm/172557508334

That is a quick, cheap, non-destructive way to weed out the fake FETs.  The Rds number is usually off by a fair amount, so to further test I put 10A through one and measure the voltage drop.