D&R stereo compressor problem

[RuudNL]

Today I was given a D&R (Dutch Rubbish?) stereo compressor.
It is based on a THAT 4301 chip, but unfortunately it doesn't work.
The unit passes signal, with low distortion.
When I turn the threshold knob counterclockwise (-20 dB or lower), the unit starts 'compressing', even without an input signal.
I have the impression that this has something to do with the circuit between the RMS detector and the 'link' point.
The DC voltage on the wiper of the threshold pot can vary between -15 and + 15 V.
The threshold voltage is attenuated by the combination R3/R4 (R16/R17) and inverted by IC1d (IC2d).
But...there is also a positive bias voltage, coming from  D11/D13 and a negative voltage from R6/R18, making the whole situation a little confusing.  Anyway, diodes D12/D14 prevent any positive voltage getting to the following stage.
So: for some reason there is a negative control voltage after D12/D14 without any inut signal.
Of course it could be that the previous owner blew up the RMS detector. I will get new 4301's next week and see if this improves the situation.

Any thoughts in the meantime?

 

[JohnRoberts]

A leaky  electrolytic DC blocking cap could introduce DC that the compressor would read as signal.

JR

 

[RuudNL]

Just checked the input capacitors, but they are OK.

 

[RuudNL]

Problem solved!
After some careful studying of the components on the PCB, I noticed that they had placed a value of 20 K instead of 2K67 for R3 and R15.
So the compressor already left the factory with a serious problem!
I suppose this says something about their quality control...