Perhaps they used them in the Proteus 2000 racks also? those sell much cheaper used on ebay, often around 20 USD if someone yanked the ROM out of it. I don't own any to check for the chip though, any volunteers out there?
You don't happen to have a service manual for these do you?
My XL-7 exhibited a problem like this. The clicks were present on both the analog and S/PDIF outputs, and were apparently processed by the effects processor as reverb settings would affect the sound.
Example here: http://www.butoba.net/downloads/xl7click.wav
I never tried the sub outputs to see if the problem was only on the main outs.
All diagnostics (in the XL-7 diagnostics meny) passed without problems.
At the time I was in contact with EPR Electronics, who claimed they'd never heard of a problem like this. Their best option was to purchase a new replacement motherboard from them, but including postage, customs and tax it would have cost me about the same amount as another Command Station, so by chance I got hold of a PX-7 whose mainboard now populates my XL-7 without problems. Only gripe is that it says 'PX-7' in the display when it starts up.
Back to the problem: In my case it got more frequent when the machine was warm, i.e. after having been switched on for a while in a warm room (I noticed the problem in the summer, no air conditioning). I tried narrowing the problem down by running it with the case open and heating various parts of the board with a hot air gun (hair dryer would do too). From what I could make out, heating the Xlilnx FPGA chip near the MIDI connectors made things worse, so I figured I'd try to get a replacement and see if that made a difference. Unfortunately I have not been able to source the chip (Xilinx SPARTAN FPGA XCS05XL in a PLCC-84 package). There is a seller on eBay who apparently sells these chips; I purchased one but after having removed the original chip and replacing it with a socket, it turns out the replacement chip was either faulty or fake; I'm leaning towards the later, as the print on the replace chip is white whereas on the original one it is golden. There's a variant of the chip in a different packaging which seems possible to get hold of, but building an adaptor would be a nightmare so I'm not going down that route.
The fact that heating the FPGA caused the problem to get worse doesn't really mean the problem is actually in the FPGA, it could be another chip that's actually the culprit, but at least it's a place to start. I was considering the FX RAM chip, but heating that area didn't affect the frequency of the problem so I decided not to try replacing it even. I did get hold of a spare from some plug-in-RAM board though just in case.
My recommendation if it is in fact the same problem as I had is to replace the mainboard, either with a spare one, or get a whole unit. I would be happy to test a replacement FPGA on my original XL-7 board but it doesn't seem possible to get hold of one.
/Ricard