Weird leakage current. bad PCB?

[Sleeper]

Hi all,
I'm building ruff records helios EQ. no problem.
And for makeup I'm building a basic API makeup amp.
It's the circuit seen in the solid state pultec, etc. and I've built it at least a dozen times without any problems.  what I'm seeing on this one is weird.
or maybe it's normal and I just never noticed it before.

So I'm testing the board BEFORE I put in the 2520 and for some reason
the feedback circuit between input- pin and the output pin is showing voltage.
That part of the circuit is isolated from ground and the output transformers by capacitors and there's no DC on either side of those, as expected.
Between -20mv and as low as -3Volts
If I short that part of the circuit it dumps the voltage, and I'm assuming that the higher negative voltages are the result of the capacitor charging up. 
???
I built it on perfboard and used epoxy to glue in the pins before I soldered in my parts.
There are no shorts and the circuit is correct. So weird.
I've ground down the epoxy to isolate that part of the circuit, cleaned the surface with denatured alcohol, you name it.  That lowered the leakage to 1/3rd its previous levels, but there is still voltage leaking along the surface of the board.
It's pretty clean voltage...  looked at it on my scope and it's a pretty flat DC line.

My question is this ??? what does this do to the performance of the amp? 
Will the leakage read as noise and distortion or is it just going to cause a simple offset? 
The voltage isn't escaping past the blocking capacitors...
or
in other words, should I rebuild these on a different kind of perfboard
(one that isn't contaminated or whatever it is that's making them so leaky)
or just ignore it.

Thanks for any advice
Kelly

 

[ruffrecords]

What sort of perf baord is it? Is it glass fibre based of SRBP?

Cheers

Ian

 

[Sleeper]

Hi Ian,
That's what I've been thinking.
The perf board is pale yellow, almost for sure fiberglass.
I got the perf boards at a surplus place here in the Burbank jungle.
Quality there is generally very high, loads of surplus aerospace parts, but I must have picked up something strange or infected with some kind of moisture. The boards didn't look like they were sitting around in a puddle of dirty water.

I thought it might be the epoxy, but I've carved out little moats around the pins and the parts, there is no longer any epoxy bridging one section to the next.

I'm leaning towards rebuilding the amps on new perfboard, particularly since the EQ goes into my mastering chain,
however 
I'm still curious what effect this DC leakage would have on the feedback loop of an amplifier.
OTOH it's more of a theoretical question since I've only got a pair of 2520's right now and I don't particularly feel like putting them at risk to find out.

Kelly

 

[okgb]

I once mistakenly used acid core solder and it made  all sorts of subtle bizzare connections,
fixed it by simply cleaning the boards with water , perhaps there is some contamination on your material ?

 

[Sleeper]

perhaps there is some contamination on your material ?

do ya think?

oh and by the way, that's the circuit, below the cat...

 

[Andy Peters]

oh and by the way, that's the circuit, below the cat...

My cats do that.

-a

 

[Sleeper]

My guy loves to sit in my lap while I solder and lie about on things I seem to be interested in.
I'm so gay for him.
In the future, all submissions of working drawings should include pictures of cats.

 

[JohnRoberts]

Hopefully this is not your problem, but I recently discovered that my balky starting lawnmower was experiencing high impedance (leakage) at the magneto kill winding from dirt inside my on/off switch. While only 10 meg or so, it was enough to make it hard to start,, I cleared it by scraping off the top layer of crud from the bakelite switch contact board.

JR

PS: I don't have a cat to photograph, perhaps I can photograph my neighbor's cat's dirty footprints on my car roof and back window. 

 

[Andy Peters]

In the future, all submissions of working drawings should include pictures of cats.

Tim Berners-Lee was asked why he invented the World Wide Web, and he responded, "To make it easier to share photographs of cats."

-a

 

[Sleeper]

I scraped and grinded and jabbed away with my dental tools, up until the point where I posted this thread. I yanked the transformers and I'm making progress on new boards... different substrate.

I'm leaving the leaky one's intact for testing, but I do suspect something moist and conductive must have gotten into the very fiber of the boards.  OR it's the solder pads on the perfboard making some kind of network... the suspect boards are by Vector.

Besides the fact that the voltage GOT there. 
My real interest at this point is what happens
Theoretically_
pretending that the op-amps behaved normally
The leakage would be like injecting bias voltage into the gain/feedback loop

I've built a few single sided amps, and these work fine, but the bias offset is injected to the in and out points of the amp, so the amp feedback works correctly.

In this case though, it seems like the feedback would create all sorts of waveform anomaly, because it's biased differently than the input.
Wouldn't it?
??
I've asked my cat, but really, he's no help...

 

[ruffrecords]

Hi Ian,
That's what I've been thinking.
The perf board is pale yellow, almost for sure fiberglass.
I got the perf boards at a surplus place here in the Burbank jungle.
Quality there is generally very high, loads of surplus aerospace parts, but I must have picked up something strange or infected with some kind of moisture. The boards didn't look like they were sitting around in a puddle of dirty water.

Moisture was my first thought. If these have been stored for a long time they could easily have taken up moisture. Might be worth trying baking them in an oven for an hour or so.

Cheers

Ian

 

[okgb]

Did ya say ?
if you tried cleaning the boards , deflux , isopropyl , water .
anti furball  medicine .

 

[Fdieck56]

Leaky electrolytic in lower feed back leg? Nah it’s probably the cat hair. BTW I once found an electric cap in my dogs poop. Short leads, removed from board so I didn’t worry.

 

[mkiijam]

Is this it?

"One of the most common application problems encountered is the failure to provide a dc return path for bias current in ac-coupled operational- or instrumentation-amplifier circuits."

https://www.analog.com/en/analog-di...oblems-when-designing-amplifier-circuits.html
I admit I'm not entirely sure what your issues is, but it reminded me of a Pultec-ish EQ I built, and like you, was using a DOA for the make up gain. I noticed DC build up and then found this article. When I gave the DOA a DC path to ground ( I think I chose like 100K or something) the DC issues went away.

It occurred to me that the filter section of the EQ is really just one big cap, so it fit into what the relevant section of the article was talking about.

BUT. maybe this isn't what you have.