Many many problems (several revisions), but I cannot go into them, because fortunately I was not part of the hardware design team. Definitely nothing I would recommend as part time project.

The digipotis were actually one of the hardest problems for this project. Additional to the many pot values and different tapers, you need to also cover a wide variety of voltages that can even be centered around zero (it is an audio signal after all), so you need to be able to handle negative voltage. DigiPots also have a capacitance, so when you have to replace high values like 1M-Ohm you will wind up with a low pass filter. We built a bunch of modules for common pot values and do the taper and uncommon values in software. If both sides of a poti are used, we will need to use two digipotis to simulate them.

Noise was a hard problem as well. In the end what helped us a lot was to keep the FX device in its own case and even solder the digi potis right in place where the real potis had been before. You can see that here: https://im.static-thomann.de/pics/images/stompenberg/backsta...

The relais modules are on the outside of the case, but on the left you will see a flatband cable (digital) run inside of the case, where the digipoti modules will sit. So all analog signals never leave the original case and the metal housing shields all kinds of external interference.

Our main concerns were a) the conversion process would be much more complicated b) hardware would be more expensive c) the motors would not live as long as the digi-potis