Make an ASCII drawing how the devices are  connected digitally via AES and ADAT.

All devices need to be clock synchronized.
Only one clock master is allowed.

Clock master is the device using it's internal clock and sending the clock signal out through any digital output (ADAT, SPDIF, MADI,...) or Word Clock, if present/used.

All other devices have to receive the clock from the master either through a digital input which is connected to the master or by word clock, if present/used.

It's possible to clock synch the setup either completely by using digital connections or by word clock alone or a mix of both.

From operational perspective its beneficial to use the main RME interface as clock master, then  the application/DAW can set the clock rate via the RME driver and in an ideal case for the whole environment.

This works best over 2ch connections like AES and SPDIF. With digital connections using SMUX (channel multiplexing), higher sample rates (double/quad speed) can not be distinguished. The protocol header has no marking for SMUX connections.

In your stup you might use the RauDAT as clock master by configuring in the driver settings clock source = internal.
Please make a drawing which shows whether and how all other devices (clock slaves) get the clock signal from the master.
They need a setting like clock source = ADAT or sometimes also called optical. Or AES or SPDIF...

In the driver settings of the clock master (RayDat) you can see the sych status if it has an inbound digital connection from a device acting as clock slave. A "lock" is not enough  status synch is desired / needed. Otherwise the device is not really in synch with the master.

Sorry gent, I was simply misled by your statement:
"I had a thought that perhaps the lexicon was q the raydat somehow???"

If you have a valid undestanding and concept / setup , then something like this should not be the case...

I would test, what triggers this reliably and then simplify the setup. Does this happen also with RayDat alone as CM and then add another device, retest and so on until the issue starts again.

seems to me the raydat is slave to an external clock that is not sending a sync signal (e.g. device turned off). if you then reboot the RME device it goes into clock mode 33khz

I bought it this year. I updated all the drivers/firmware? when I installed it on the new machine as well last week.

M

p.s. nice to see you still around Mathias we go back to the 90's and Nuendo 1 you had me soldering capacitors on a motherboard once to solve an issue

that was going to be my next thing to try.

it's in the only small PCIe slot on the board , the same size as the card tabs so I put it in there. I imagine it's the original spec and the longer ones are all PCie 4/5 x16

it's PIA though as I've got to disconnect/and reconnect all 8 ADAT cables!!!

I might dig into the BIOS and see how things are set regarding fast boot, PCI specs. As i said it seems like it's not getting initialised on boot when it happens. Maybe modern PC's boot too fact for 'old' PCI to cope with....

just a thought...

Obviously this is a bleeding edge cpu and MOBO . It worked flawlessly on the olf ASROCk x570 and 5950x machine though. However that was a different architecture. I may force the PCi slots to older PCI specs. I'm not using a GFX card so don't need PCie 5 specs.

M

Devices going to 32k is often the result of a clock loop (two devices trying to sync to each other). What happens with nothing connected to the card?

Update: This issue returned again yesterday....

I'm away from the studio now fro Xmas but will report back with any further changes when I try them.

Interesting that another poster is having a sample rate reset issue as well with a RAYDAT.

M