Help with 12v 1A output from 5v MCU Pin - Transistor getting hot with no load

Thread Starter

USAWME

Joined Mar 29, 2023
18
Hi all

I have a circuit i have made and this is 1 output stage for my pcb.

When running the output the transistor heats up quite quickly with no load attached at the mosfet.

See attached image for the schematic where Compressor_Trig_1 is the output from the Atmega pin, Both R17 and R18 have values of 10k and the diode is in flyback configuration.

I'm a bit stumped even in a simulator the currents are quite low 10mA or so even with the 1A load on the output of the mosfet ( Compressor_Out_1)

So am a bit lost as to what the issue is.

The board has 10 outputs and all do this so its not as if it is just a once off thing and also have 5 boards made and they all do it across all of them.

Thanks for the help.

Cheers,
Darcy
 

Attachments

MisterBill2

Joined Jan 23, 2018
18,937
It looks like probably the 2N3904 is not switched on and so the mosfet is full on and the part of the circuit not shown is drawing all the current the 12 volt source can supply. That wire goes on down to something we are not aware of.
 

Thread Starter

USAWME

Joined Mar 29, 2023
18
I don't know that chips analog and gpio setup settings but are the pin internal pull-ups off in the config? They normally are about 45 to 50K tied to VCC, so with a 10K pulldown that's in the range of 1V on a voltage divider.
No internal pullups are on just output switching high and low Normal state is low and a switched output is high

Something is wired wrong then. My first guess is the diode isinstalled backwards.
That was my initial thought but i have tested and its not backwards but even if it was wouldnt that cause the mosfet to get hotas would be dead short and not the transistor?

My first thought is the MOSFET and diode are installed backwards.

Or maybe just the diode if a drive signal is present to the NPN.
From what i can tell with my board and tracing the traces the diode is correct unlessthe markings on the diode are back the front

It looks like probably the 2N3904 is not switched on and so the mosfet is full on and the part of the circuit not shown is drawing all the current the 12 volt source can supply. That wire goes on down to something we are not aware of.
The wire goes to nothing currently it goes to a header pin that then goes to a plug so right with nothing connected that transistor heats up, If i connect the output to the header pin it all functions as normal and works fine switches on and off with gpio switching and functions other than the transistor heating up



Thanks for all the replies, i will investigate some more but as far as the circuit diagram there is nothing that stands out to be a bad design is there?

Thanks,
Darcy
 

MisterBill2

Joined Jan 23, 2018
18,937
I would call that having the mosfet fully on biased as a default is not a good design. That assures that the output will be on before the controls become active. The power-on default mode should not be output on.
 

Thread Starter

USAWME

Joined Mar 29, 2023
18
I would call that having the mosfet fully on biased as a default is not a good design. That assures that the output will be on before the controls become active. The power-on default mode should not be output on.
Thanks for the response, I'm not following though when you say default on? Currently when it starts up all outputs on the mcu are driven low and so is the mosfet the output is only on when the gpio pins are on so the 12v output only comes on when the gpio pins are driven high.
 

Alec_t

Joined Sep 17, 2013
14,370
Double-check that R18 actually is 10k and isn't bypassed by some rogue conduction path. Also, verify the pin-out of T9.
 

BobTPH

Joined Jun 5, 2013
9,127
BobTPH said:
Something is wired wrong then. My first guess is the diode isinstalled backwards.
That was my initial thought but i have tested and its not backwards but even if it was wouldnt that cause the mosfet to get hotas would be dead short and not the transistor?
In the first post you said:
When running the output the transistor heats up quite quickly with no load attached at the mosfet.
The MOSFET IS the output transistor.

So now you are saying it is the other transistor (2N3904) that is getting hot?
 
Last edited:

Thread Starter

USAWME

Joined Mar 29, 2023
18
If T9 gets hot and both R17 and R18 are 10k then something is wired wrong.
Can you show pictures of your circuit?
Heres some pictures:

20240424_083242.jpg
20240424_083431.jpg
PCB1.jpg

Double-check that R18 actually is 10k and isn't bypassed by some rogue conduction path. Also, verify the pin-out of T9.
Have confirmed they are both 10k with a multimeter


In the first post you said:

The MOSFET IS the output transistor.

So now you are saying it is the other transistor (2N3904) that is getting hot?
Yes the transistor gets hot not the mosfet so even with no output connected to the mosfet the transistor gets hot (2N3904) as long as the 5v output is turned on from the gpio pins of course when off its all cold and no voltage is present on the mosfet output as intended. When on it does work and 12v is present on the mosfet output but the transistor gets hot.

post pictures of your construction...
Have posted pics above in this post.

Cheers
 

panic mode

Joined Oct 10, 2011
2,795
which part exactly is overheating?
what is the supply voltage? drawing shows 12V but is it really 12V and is the polarity correct?

i see pictures... not the greatest... not sharp with good contrast. missing bottom side view... glare... not aligned... do i need to turn my head to look at this?

but there is also PCB layout and the 2N3904 and SQ2309ES pinouts match their respective datasheets.

i sort of see 3904 marked on T1 so this is probably ok.

1713914232836.png

i see sPTGD marked on what should be PMOS. according to this file, that PMOS part is made by Vishay and should be marked 8P.

looks like you have a different part there...

1713914312126.png
 

ronsimpson

Joined Oct 7, 2019
3,131
For the 3904 to be hot there must be current and voltage across it. Current is coming through the 10k resistor or through the MOSFET. I don't see any other sources. I would pull the MOSFET and see if the 3904 is cold AND/OR pull the resistor and test.
I feel that removing the MOSFET will tell us something.
1713917285268.png
 

ElectricSpidey

Joined Dec 2, 2017
2,829
Silly me, I must try to remember that not everyone refers to MOSFETs as "transistors".

And the clue that the OP was referring to the bjt is right in the first post. (10 mA in the sim)

So, if the 904s are overheating...that's a real mystery.
 

Thread Starter

USAWME

Joined Mar 29, 2023
18
which part exactly is overheating?
what is the supply voltage? drawing shows 12V but is it really 12V and is the polarity correct?

i see pictures... not the greatest... not sharp with good contrast. missing bottom side view... glare... not aligned... do i need to turn my head to look at this?

but there is also PCB layout and the 2N3904 and SQ2309ES pinouts match their respective datasheets.

i sort of see 3904 marked on T1 so this is probably ok.

View attachment 320588

i see sPTGD marked on what should be PMOS. according to this file, that PMOS part is made by Vishay and should be marked 8P.

looks like you have a different part there...

View attachment 320589
Sorry for the pictures was hard to get a picture of it. The marking is 8PTGD and acording to the pick and place for the pcb manufacturer its the same part p channel mosfet 60v etc.

Supply (marked 12v) is a 12v transformer on the bench where tis testing is but in the vehicle will be 12-15

For the 3904 to be hot there must be current and voltage across it. Current is coming through the 10k resistor or through the MOSFET. I don't see any other sources. I would pull the MOSFET and see if the 3904 is cold AND/OR pull the resistor and test.
I feel that removing the MOSFET will tell us something.
View attachment 320591
Thats what i am thinking somehow between 12v line returning fro the mosfet. SHouldnt any issues with the 12v going through R18 a thats 10k so talking mA there but if something is returning between mosfet and transistor then it would cause issues as whatever voltage is there would be connected to ground through the transistor.

Im not sure how to eliminate this if thats the case. I thought this design was correct wit what i was trying to acheive.

I will pull the mosfet and see what happens

im going to assume it will be cold as thats the only plausible answer that i can see.

Silly me, I must try to remember that not everyone refers to MOSFETs as "transistors".

And the clue that the OP was referring to the bjt is right in the first post. (10 mA in the sim)

So, if the 904s are overheating...that's a real mystery.
Sorry i should have been more specific
 

Thread Starter

USAWME

Joined Mar 29, 2023
18
You don't literally mean a 12V transformer, do you? A 12V DC power supply, surely!
Sorry yes a 12v dc power supply
It is a normal brick wall transformer though not anything fancy

But in car the same thing happens on a 12v dc battery so assume floating around 12.4v give or take
 
Top