DC buildup in Linkwitz Transform

[DF96]

Yes, either a dry joint or a crack in a PCB trace. If you are using modern lead-free solder then I understand that dry joints and good joints can look identical (I always use the old 'poisonous' solder). If you showed us the circuit we could probably tell you where to look for the problem.

[Simon B]

I'd look for an op amp input fed by a capacitor, that also has a resistor to ground, or to some other circuit node that direct current can flow out of.

If that resistor is not properly connected, the small bias current flowing out of the op-amp input charges up the capacitor and hey presto, the very fault you describe.

Probably, learn to solder properly, you've got a dry joint at the end of a resistor. Before you fix it, look at it, see how it looks different from good joints. As DF96 says (and I paraphrase a little) "use proper solder".

As DF96 also says, possibly a crack in a pcb trace. They can be very fine indeed and still be an open circuit. But my money's on a dry joint.

[sreten]

Hi,

Cracks in PCB traces can be obviated by tinning all of the PCB tracks
with solder, but doesn't work that well for high current tracks.

rgds, sreten.

[petarkb]

I couldn't find any obvious bad joints, but I "untangled" some blobs where lots of leads met. On the third one the buildup disappeared, and the output is DC-free!

Thanks a lot to everyone, and especially to Simon B. I went untangling the capacitor junctions first. The troubled capacitor was closer to the output, but was a part of the feedback loop. It's the rightmost C2 in the schematic attached.

There are 'before' and 'after' pictures, if anyone's interested.

DF96, sorry for misleading, but I didn't realize at the time. As you can see, the before picture doesn't have obvious 'bad' joints, just lots of evil blobby ones. Guess I underestimated their evilness.

My next project will be adding a high pass filter before the Linkwitz Transform. My speakers have Visaton W170S woofers, and the LT extends them to Q=0.7 and F3=30Hz. They don't do 30Hz like I expected. Distortion is bad, and cone movement is scary with anything except music. My room needs less deep bass anyway.

Going to use a 1st order high pass at 50Hz. If I mess that up, I'm gonna end up with a flat forehead, from a wall.

[petarkb]

My bad, the affected capacitor is C3, not C2.

[DF96]

As I said:

Originally Posted by DF96

A wiring error?

A short caused by too much solder in the wrong place is a common fault, often introduced during a 'repair'.