[SOLVED] P-Channel mosfet voltage loss between D-S

[gmc]

I'm busy with a project to switch 5V (0.5A) using a p-channel mosfet (NDP6020)  and it's kind of working although I'm getting a unexpected voltage drop across the D-S of the mosfet.

https://www.onsemi.com/pdf/datasheet/ndp6020p-d.pdf

The mosfet is being driven by a NPN transistor and I'm using a 3.3V atmega88 microcontroller to switch the mosfet on/off

With the output of the Atmega88 at 3.3V this turns on Q2 which puts the voltage of gate of the mosfet  low enough to turn it on.

I would expect 4.98V at the drain. I though it might be as there was no load on the drain but have added in a 1k resistor from drain to GND and still seeing 4.63V

Clearly there is some issue with the circuit or my understanding of mosfet is wrong.

I've taken voltage readings and added them to the schematic.

Where am I going wrong?

[Phil1977]

Did I read it right you loaded the FET with 1Ohm? Then around 5A I_DS are expected, and a voltage drop of 350mV with a RDS_on of 50mOhm + some mOhm of PCB/solder/voltage source resistance seems not unplausible.

Can you measure the I_DS?

[gmc]

Sorry, correction. loaded with a 1k resistor. 

[Phil1977]

Did you incidentally swap S and D? Is the switching working in general?

[gmc]

I've rechecked and orientation is correct, but it's not switching off either

Tried swapping the mosfet - same issue.

Have updated the schematic with voltages when the output the uC is LOW.

[HwAoRrDk]

Are you absolutely, positively sure you haven't got the MOSFET the wrong way round? A voltage drop of 0.35V sounds like what you might see for the V_F of the body diode passing current. Especially if it "won't turn off".

[gmc]

100% sure.  Rechecked footprint and soldered in the right way.

Unless I've got some dodgy mosfets but i have tested them with a component tester and the pinouts match.

[Phil1977]

Do you have a transistor tester?

Many MOSFETs from dubious sources are fake, sometimes they just relabel anything in a TO220-housing.

[gmc]

Transistor tester passes and matches the footprint.

[wraper]

NDP6020 is N-channel and you tester shows that too https://www.farnell.com/datasheets/59594.pdf, NDP6020P is P-channel. Also this is an obsolete part which you should not buy for anything other than repairs, still I would get something different for that as you'll likely get fakes from places where they are available.

[gmc]

Aaggghhh. How many mistakes can I make on this project

It's supposed to be a P-channel but during my search for a mosfet I select N-channel instead.


Will look into making some hardware hack to use a N-channel (Or probably better just buy the correct mosfet)

[wraper]

Will look into making some hardware hack to use a N-channel (Or probably better just buy the correct mosfet)

N-channel won't work without about 8V or higher supply to drive its gate unless you put it on the low side.

[gmc]

Scratch the above.  Another doh moment.

I bought a NDP6020P (So at least I'm not losing my mind) which as you said should be  a p-channel mosfet.

It's testing as a N-channel so I've been sold a fake as you suggested

[wraper]

What's actually written on the part? I see listings on aliexpress use both part numbers in the title. It's very rare that suffix means different channel type, so easy to get confused even for experienced engineer.

[gmc]

NDP6020P

[wraper]

Then must be fake, counterfeiters looked in the wrong datasheet 

[gmc]

Thanks for the help. At least that explains it.  Will source a more modern part while from a better source now.

[Phil1977]

Then must be fake, counterfeiters looked in the wrong datasheet 

Or the P-type is 1cent more expensive... This way the component will at least not work from the beginning instead of failing during usage.

People who organize this should get circumcised with the solder iron.