Low-level noise from N72

[mGeorgeStuv]

Hi, I've built a couple N72s. I've done the Oscope bias procedure with them and tested them with both a dynamic and condenser mic. Both pass signal, sound good. Did a pink noise to output test and the spectrum looks flat.

WIth one of them I am getting intermittent low-level noise which I can hear on the output at higher gains. I would describe it as the kind of a cold solder joint or intermittent contact noise - kind of scratchy or rumbly. It is there with a mic, or when I unplug the mic or short 2-3 together on the input jack.

The noise gets louder with higher gains and disappears with the trim pot (R19) turned all the way down, which suggested to me it was not in the output (Q4-Q6) stage. I have a bench supply which I have used for testing other modules, rigged with the +/- 30 provided to a molex connector, but not the phantom power. I tried this supply with no difference, suggesting to me it's not the phantom power circuit. Then removed J5 and J6 which I guess does the same thing (doh!) and no difference.

If I add a clip jumper between pins 7 and 10 (shorting the secondary) of the input transformer the noise goes away. I checked the input transformer wiring and jumpers, reflowed solder on the transformer connections, C1, the input jack, and the input HF cap network with no difference.

I was thinking of redoing the input transformer installation. I'm a bit worried about desoldering the big tabs - anyone done this? I have a Hakko desoldering iron with a couple different size tips, but it seems like it would be hard to get it to the point of being able to pry those loose without damaging the pads.

Also has anyone had bad transformers that acted this way? Or know of a different problem that I should be checking for?

Thanks

[mGeorgeStuv]

Continuing to troubleshoot ... removed C1, no difference. Pulled wires for pins 2 and 5 on transformer (primary) up from board to isolate transformer from jack, etc. No difference. Repeated my experiment by shorting pins 7 and 10 (secondary) of transformer. Noise is very attenuated again. To make sure reflowed solder on input stage (Q1-Q3), attenuator switch and resistor network. Remove jumper. Noise seems better. Need to compare to other N72 again. So maybe was input stage or resistor network solder joints. Any thoughts still appreciated. Thanks.