the circuit as shown should work. there is a problem which is not obvious. therefore one has to do some troubleshooting.
it could be issue with PCB or solder bridges under the parts. for example post #15 is spot on.

but... you are not exactly answering any of the questions...!

why is it hard to be specific and do some measurements so we have some real numbers to look at as opposed to assumed or expected values?

1. which transistor is overheating? BJT?
2. where is the picture of the bottom side of the PCB? are through hole parts assembled at PCB shop or after you received boards?
3. did you measure all the parts? if R18 is shorted, BJT will overheat but only when switched on. did you try to turn off input signal? if R17 is shorted, transistor will overheat but only when turned on.
4. what is the actual (measured) input voltage?
5. what is the actual (measured) output voltage?
6. what is the actual (measured) voltage between "+12V" rail and GND?
7. what is the voltage across R17 and R18?
8. did you try removing any of the parts?

a) removing BJT should allow you to test the output manually by touching gate of the mosfet with wire connected to GND. and you can measure current through that wire. this will be important clue.

b) removing MOSFET should allow you to test the BJT circuit and BJT should not be getting warm unless there is still a hidden short somewhere. but this is a MOSFET you would be removing so you need to follow ESD guidelines to not damage it (unless it already is damaged).

 

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...

I agree that you have the wrong part on the board. That is why in post #15 I wanted you to unsolder the "MOSFET" and see what happens. The part on the board is either a BJT or a Zener diode.

 

exactly, and a bit of probing with multimeter could quickly reveal the truth...

 

exactly, and a bit of probing with multimeter could quickly reveal the truth...

I have done some probing and i have quickly found an issue. Almost all of the resistors in the same part of the of the output circuits (R2, R4 and so on) (not all but in saying that not all heat up) dont have the 10k resistance anymore. Majority are around 0.3k now some around 1k and some around 0.4k

The brand new boards that haven't been connected to power are all at 9.9k - 10.1k ( as best my multimeter can measure)

So for majority of the output stages ( lets call this diagram an output stage) Most of the Resistors in position R2, R4, R6 and so on have dropped in resistance to 0.3-0.5k but on the brand new boards that haven't been powered on they all read 9.9 -10.1k thereabouts.

 

1) Unsolder a resistor. measure its resistance. Apply 12V to the resistor. Wait one day. Measure the resistor again.
or
2) Leave the resistor on the board. Measure resistor. Short out T1 or leave it on. Wait one day and then measure the resistor again.
I have a fear that M1 or T1 have become damaged and causes the meter reading to read 300 ohms. So removing the resistor from the board may tell us more.

 

1) Unsolder a resistor. measure its resistance. Apply 12V to the resistor. Wait one day. Measure the resistor again.
or
2) Leave the resistor on the board. Measure resistor. Short out T1 or leave it on. Wait one day and then measure the resistor again.
I have a fear that M1 or T1 have become damaged and causes the meter reading to read 300 ohms. So removing the resistor from the board may tell us more.

Thanks for the reply,

I'll pull a resistor and test it but there must be something wrong with the circuit causing this to happen which makes me want to abandon this design.

Any ideas on a new design? I'm open to all new components all i need is a 5v switching output to be able to switch 12v ( 5v on = 12v on) and has to handle around 2 amps of current ( normally only around 0.9 - 1.5a but would like a little head room obviosuly)

 

Just wondering if the solder mask becomes partly conducting when the board is powered up?
When a resistor is off-board, check the adjacent mask layer region with your Ohmmeter.