help with strange mosfet behaviour

Hi,
I have checked I have biased mosfet correctly.
Like this-

IRF540 has 12v supply, so NPN transistor is used as inverted PWM drive from arduino output with 4k7 ohm to base and 470 ohm between 12v and collector to drive mosfet gate.

I am checking each stage of a PWM control of a glass grinding motor.
Before I connect the Arduino, I hardwired the NPN driver transistor so mosfet gate has 12v via 470 Ohm resistor.
The motor runs as expected. However, the double-check with the DVM was strange.
As in the diagram above, mosfet gate has 12v relative to source.

But Rds ON is more like 0.3 Ohm than 44milli Ohm.

Tried another identical mosfet, then another from a different batch.
Then I just connected the mosfet to a lamp as the load and a resistor to 12v.

They all give the same large voltage drop and in the case of the 50w lamp load, obviously excessive heating.

I can't figure where I have misunderstood the biasing or the datasheet.
I should be operating well into the saturated region with gate at 12v?

Problem solved.

These are counterfeit junk and ebay is flooded with them, so be warned!
Look closely at the code numbers.
http://www.ebay.co.uk/itm/271608493619?_trksid=p2057872.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT

Apparently that was a very busy day at the little fake factory.
Millions with the same date code! You will actually receive items identical to the picture.

Now the problem is, I also received 5 of these identical fakes at 5 times the ebay price from a reputable UK supplier.
The genuine factories also turn out the fakes, so the paperwork looks correct to the UK importer.

They seem to check out ok with a DVM, but are way off spec. and will overheat instantly.

Do you have a 4-wire multimeter? You can't measure resistances that small, that accurately, without one.

I assume you measured the voltage across the mosfet with this known load/current.
Did you check the datasheet of the mosfet with these values.
e.g. a hot mosfet has twice the on-resistance
A 100volt mosfet might not be a good choice for a 12volt supply.

If it's a problem, use several fets in parallel for a lower Rsd-on.
Leo..

Morgan and Leo, thanks for your responses and suggestions.

There are easy ways to measure low ohmic values using Ohms law and a couple of DVMs for current and voltage.

The point is that these IRF540 mosfets on ebay are not what they say they are. Also available via UK re-sellers of bulk imports of the same items.
My faulty devices all perform the same, about 0.3 Ohm, very high Rds ON.
They could be another mosfet die, in a fake case (heatpad is different to a genuine IRF540, leads are also longer).

If anyone looks up this post, they need to know this may be the cause of unexplained overheating.
The IRF540 is a very common device.
It is cheap and robust, and suitable for 12v in its genuine International Rectifier manufactured form!
It is also still in a big plastic package with leads.

There are enough posted and responded issues to do with using logic level mosfets with Arduino without needing dodgy devices thrown into the mix.

We don't know what level of experience you have, and if you know what effect a current meter can have on a circuit.
e.g. a common DMM, set to 20A, could have a resistance of 0.3ohms (because of the leads).
That could introduce a volt drop of 2volts with a 6amp load.
Best to measure the voltage across the mosfet when it's fully "on", and not the current through the circuit.
But you might be right about fake parts.

Did you try two or three of those mosfets in parallel?
Leo..

The point is that these IRF540 mosfets on ebay are not what they say they are.

Have you compared a "genuine" IRF540 to a "fake" in the same circuit?

Or, are you comparing how you expect the IRF540 to perform to how the ones you have are actually performing?

Hi
This document may help.
Its Vishay document on how you measure the parameters.
Also

Tom… :slight_smile:

an957.pdf (228 KB)

OK, thanks to all you guys helping out with this.

Yes, I'm aware of your relevant points.

To keep it brief, I didn't go into all the details of testing.

I had a couple genuine IRF540 which came from Farnell as comparison.
I had put them into an inverter repair, but whipped them out for a check.

Rds on those measured 0.033 Ohm in exactly the same circuit, same conditions.

I've managed a very prompt refund on one batch of the dodgey ebay ones, awaiting another seller to respond (UK).

The main thing is that on this forum are many who could do without this kind of issue - so hence my follow up on solving the problem.

You guys are always so helpful, so its a bit of selfish altruism.

Mark

Now the problem is, I also received 5 of these identical fakes at 5 times the ebay price from a reputable UK supplier.

Have you considered contacting International Rectifier? (Especially about the "reputable UK supplier".)

They are in the best position to reduce / eliminate the problem for everyone including yourself. You may even be able to parlay your experience into a fistful of samples.

Friend of mine works for a company that has a policy of X-raying high value chip deliveries
before accepting goods in, to check there is actually a silicon die inside them (I presume
the bond-wires are what actually show up).

They're not looking for illegal immigrants on the X-ray?

Then there is the case of the hard drive that was just a USB flash drive and a few lumps of steel hot-glued inside the case so it felt heavy. Presumably the flash drive was hacked to over-report free space.

Perhaps you should contact your local customs office? They are usually keen to catch fraudsters.

Russell.

Simplystupid - Thank you for posting this and you're quite correct that the fake factory has been very busy!

It's not just the IRF540's that have counterfeited, I have a batch of counterfeit IRFZ44N's that have the exact same lot and date code as the 540's shown in the eBay photo. Got them from the Chinese supplier Banggood. I did not check the price against a authorized distributor, this would have shown that they had to be fake. DigiKey: $1.95, Banggood $0.38 which just proves the old saying correct once again: if it sounds too good to be true, it usually is...