FWIW, I believe this was introduced in J.
Here is an image of an audio track PC76 enabled in I.

and the same project on J

Was J the version that introduced upsampling?

 
I don't have a spectrum for it, but I can repro this at -123
 

The confirmation of an extremely low-level signal would explain why stacking 200 channels wouldn't be enough. If the signal is coming in at -129, and is being amplified by 20 dB, that brings it to -109. Even turning up all the level controls, inputs, outputs, etc. would like not bring it to the point where it would register on the -90 dB scale.
 
It would be interesting if those who can reproduce this see the level increase if channels are cloned.
 

mettelus
Was J the version that introduced upsampling?

 
Close...it introduced upsampling on playback, and patch points. 
 
I'll pursue this with Noel, the fact that it doesn't happen in I and happens in J will probably narrow things down enough to find out what the issue is.

scook
FWIW, I believe this was introduced in J.
Here is an image of an audio track PC76 enabled in I.

and the same project on J

Steve, you are one amazing forum host. :-)

Thanks for looking into this, everybody - at least it confirms that my computer isn't just being uniquely targeted with cosmic rays or something like that. :)

One last thing. The image above for SONAR Ipswich only looks that way because no audio had passed through the track. After setting the input to something other than NONE and setting it back, the image looks like J. SONAR I acts the same X3e. I suspect this noise is intended. Overloud does this in a couple of other plug-ins too. The only thing that has changed in SONAR J and newer is when the noise shows up.

In an oscilloscope it shows as a slightly distorted square wave, hence all the harmonics. Looks much like I'd expect simulated AC hum to look like. But I can't see why this would be intentional behaviour if the tone is not consistent across sample rates. Bug or inaccuracy in the upsampling, perhaps?

Kylotan
Thanks for looking into this, everybody - at least it confirms that my computer isn't just being uniquely targeted with cosmic rays or something like that. :)

Well that's a separate issue... as are poltergeists.
 

scook
One last thing. The image above for SONAR Ipswich only looks that way because no audio had passed through the track. After setting the input to something other than NONE and setting it back, the image looks like J. SONAR I acts the same X3e. I suspect this noise is intended. Overloud does this in a couple of other plug-ins too. The only thing that has changed in SONAR J and newer is when the noise shows up.

Damn, you're good.
 
I looked at the help file for the PC-76 to see if it said what that meter screw adjustment did, and the second sentence said "This module is meticulously modelled after one of the most renowned and famous solid-state (F.E.T.) compressors that is used in professional studios and recordings." Based on what I've seen from other manufacturers, that usually includes the noise. As to why it would change with sample rates, well, it IS audio...

Here's mine.  PC76 on the track or the Master Bus, span on Master bus.  Disappears if I switch off the PC76.
 
Looks more like the result of a comb filter than harmonics.  Nulls at 100,200,300,400, etc.
 

 
 
EDIT :  Saturation module has an odd ~2 second beat to its background noise.