Conductive PCB substrate

oracacle

Senior Member
Just built up a board, basic tests showed things to be fine, until I tested the MOSFET output. no matter what they always had and output - and an input. After mush pocking around with a multi meter I found that the board substrate had a voltage by comparison to the 0v side of the battery. There is a resistance between the substrate and the copper too, its high from about 0.9 to 15M ohm.

The batteries supply 6v to the copper, which goes to the MOSFETs and 5v regulator. The board is showing about 3-4v

I have no doubt that I will have to make another board, which will give me a chance to make some minor changes. But this is still really quite frustrating.

Has anybody else come across something like this? It had me stumped for a couple of days.
 

neiltechspec

Senior Member
Is that RBP or GRP substrate ?

I have seen similar issues, with cheap imported RBP, was desperate for some PCB at the time. Ended up
throwing it away & waiting for my GRP delivery from CPC-Farnell.

Neil.
 

AllyCat

Senior Member
Hi,

Are you sure it's not just a surface effect (moisture), it's been quite humid (foggy) here in parts of the UK recently?

But the P in SRBP (or "Paxolin") does stand for "paper", which can absorb quite a lot of moisture. It does sound like "faulty" manufacture, but can't the FETs be "fixed" wih (say) 1 Mohm resistors between Gate and Source (generally good paractice in many applications anyway) ?

Cheers, Alan.
 

darb1972

Senior Member
but can't the FETs be "fixed" wih (say) 1 Mohm resistors between Gate and Source (generally good paractice in many applications anyway) ?

Cheers, Alan.
+1 for Alan's suggestion. Even with a "good" PCB, the slightest bit of dust in the right spot couple with a bit of moist air is possibly enough set off a FET.

Did you "brew your own" PCB or did you have it made? Maybe the etching process failed and left tiny traces of copper in the substrate. Do you have a microscope? If so, maybe have a close look at the etched areas.

Another thing I've come across in many designs that use FETs is PCB routing (complete removal of the PCB) between pins to further elinimate unwanted "stray" components.
 

oracacle

Senior Member
Its home brew board.
The FETs have a 10K between gate and drain. they are IRF9520 (n type FET). The circuit works perfectly on bread board.
If I remove the FETs there is still voltage on the gate and source, the tracks also become islands with them removed so should have not voltage at that point - they show about 3.5V.

I think it has paper between some layer of GRP

I am not sure what you mean Tex, its a single sided board if that helps
 

techElder

Well-known member
I am not sure what you mean Tex, its a single sided board if that helps
That does answer my question.

Multi-layer boards are like stacks of single layer boards. Your motherboard in your computer has many layers of traces. If you put a via through a multi-layer board, you have to be sure you clear an area on any internal layers so they won't touch the through hole.
 

AllyCat

Senior Member
Hi,

The FETs have a 10K between gate and drain. they are IRF9520 (n type FET).
The resistor should be between the Gate and Source to hold the FET off. And that FET is a P-channel device (which is why the Source is shown at the top of the symbol in its data sheet). :confused:

Cheers, Alan.
 

AllyCat

Senior Member
Hi,

At 15:19? OK, I know the forum clock is a little wonky, but it sounds as if you're worse than me. ;)

But to be fair, I believe there are a few night-workers on the forum

Cheers, Alan.
 

oracacle

Senior Member
I will try and get a pick up soon.
I found the board is FR2 - which IIRC does have some paper, wiki quotes as having low moisture resistance. living on Dartmoor means its always a bit on the damp side.
Second attempt did the same and I also found that the default pads on the MOSFETs are too thin to be reliably etched with no issue. to get over this I widened the pins placements and increased there size - this mean bending the pins to get the MOSFETs in, but if it gives a more reliable circuit than I am fine with that.
 

mortifyu

New Member
Hi Guys/Gals,

A VERY important point made earlier in this thread is the necessity of a resistor being present to turn the FET off.

This is because the GATE of a FET requires mere micro amps to turn it on.

Essentially a resistor from GATE to SOURCE in both cases of N-Channel or P-Channel FET's.

Often with 5V logic circuits I use a 10K resistor.

Think of it being that if you supplied the gate with a voltage to turn it on, then just removed the voltage, the gate is just left floating around in the breeze. The FET's internal gate capacitance is enough to keep it switched on in many cases hence using a resistor is the sure way to return it to OFF when the drive voltage is removed.


Hope this helps.



Regards,
Morty.
 
Last edited:

fernando_g

Senior Member
What sort of flux did you use? THIS is very important.
And most importantly how did you clean it?
Is it thru hole or SMT?

With the limited description, your problem seems to be surface ionic contamination.
Board leakage is then exacerbated by changes in humidity.
 
Top