Resistor insulation problem

[skjafar]

Two weeks ago I was working on an 800V power supply when it suddenly started burning input fuses. It would directly burn the fuse just as i connect it to power.

Of course the usual diagnostics to checking shorts, DMM checks and then general components checks, no good, then I went and checked all the components (not so many) still all were looking ok. This was when I was really getting frustrated, it was an old linear regulator, so what I did was connect an external DC power to the regulation circuit with current limit to see if I can check the problem live, the bloody thing was working ok, of course I was working on 20V and it originally has an 850V input, then one of my colleagues suggested that it might be insulation problems, I tried to check it without power I was not able to, then I took the gamble of having the board directly in my view then plugging it in, I saw a small spark, it was in a middle point between a transistor and a resistor, so I instantly thought it might be sparking between the leads, nope, after actually having to remove both the transistor and the resistor I found out that the insulation layer above the resistive film in the resistor has decayed in an area as big as a hair pin, and this has happened only on the bottom side of the resistor, with the ground plain underneath all the components 850V on the resistor was enough to create a spark there.

Simple problem of changing the resistor  took one whole day 

[David_AVD]

Sounds like a clearance issue.  I don't think I'd have 850V on a resistor that was pressed up against the 0V plane.

[G7PSK]

I would have put a sleeve onto the resistor as the insulation cover on the board would be damaged as well. The insulation on resistors is not much more than paint and should not be relied on at high voltages.

[skjafar]

I would have put a sleeve onto the resistor as the insulation cover on the board would be damaged as well. The insulation on resistors is not much more than paint and should not be relied on at high voltages.

What I did is raise the resistor off the board. Good idea though.

[Len]

You need to change your sig now. 

[skjafar]

You need to change your sig now.

It's My Signature. So I don't care what you think.

[Balaur]

You need to change your sig now.

It's My Signature. So I don't care what you think.

Len was joking. Based on this thread's subject, it looks like you also need to check resistors in faulty test gear.

[emc2]

You need to change your sig now.

It's My Signature. So I don't care what you think.

Len was joking. Based on this thread's subject, it looks like you also need to check resistors in faulty test gear.

what i was thinking

[skjafar]

You need to change your sig now.

It's My Signature. So I don't care what you think.

Len was joking. Based on this thread's subject, it looks like you also need to check resistors in faulty test gear.

If it's a joke then I apologize for coming too harsh.

[David_AVD]

Ahh... the difficulties of Internet humour....   

[larry42]

without a schematic it's hard to envisage it completely, but you know that resistors have a max. voltage rating that is independent of their power and resistance rating, right? Special resistors are required for 300V operation. I know, I ordered ones for 750V AC.

[David_AVD]

The working voltage (from end to end) is also a different parameter to the insulation (coating) rating too I believe.

[Len]

Ahh... the difficulties of Internet humour....   

Yes, I've added a smiley which I should have put there.

Sorry if I gave offense!

[SeanB]

The funny thing as well is that the common white ceramic boat style resistor actually is not an insulated resistor ( strange that, it is in a ceramic boat and held in with cement) but is actually rated as an exposed element device. This is because it can run so hot ( bright cherry red on overload no problem) that the ceramic gets conductive, not an issue with a sub 1k unit, but enough to cause problems with leakage current to adjacent conductive substances touching them. you have to go to the metal clad ones to get a proper case to substrate insulation rating.

[skjafar]

Yes, I've added a smiley which I should have put there.

Sorry if I gave offense!

No harm done mate, I am new in the forum, maybe other guys here know you more so they knew what you meant.

Now we can be friends 

[David]

I remember working on a similar problem with some resistors on a power supply. The insulation was changed by the manufacturer without telling anybody and they had "pulled it back" from the leads which caused arcing. Got to play with a flash tester which was quite fun though!