It has been reported several times that the PreSonus AVB switch has a problem with reliably cascading RM/RML AI mixers. I have also seen the Master vanishing - ten minutes before an event was supposed to start. This was no fun at all...

Have a look at the Beta Firmware section of the Forum for more reports (you need to be locked in). If you happen to have a MOTU AVB switch, your problems will be gone right away. On the other hand the PreSonus AVB switch works well with only one mixer and 32 input channels.

It has also been reported that - when a PreSonus AVB switch is involved - the sample rates of the mixers can change. It happens that after cascading the mixers are using different sample rates.

May I ask which firmware and software versions are you using? And how are you cascading the mixers? Via computer or via iPad?

Backups... - this should work, but not via FireWire. The 'Universal Control 2.7 Milestone Release Notes' still say: "Restoring a Mixer Backup over Firewire is not possible. Must be done over Network connection."

Thanks for your quick answer.

I am not sure if this is related to the AVB switch. For some reason IP details of the RML was changed. It switched from Static to DHCP causing the cascade to break.
That thing is worrying me, because it can happen always, like now 2 hours before showtime.

Anyway, I will have a look at the beta section to see if I can find something there as well. Any other remarks or suggestions are welcome.

It is the PreSonus AVB switch, no doubt. If you check your set-up again, you will find the IP-addresses displayed "as usual", but this doesn't mean that the two mixers will be cascaded reliably. The Master can disappear any time, which means that no audio will be passed to the Slave. I have been checking this behavior several times and the only thing I can say is that cascading RM mixers is rock solid with the MOTU AVB switch.

Thanks for this.
If this is really the case that this is bad thing. The switch is sold as a supported model for AVB.

I am still puzzled however how the config of the RML could have been changed from static to DHCP and therefore received a new IP address. Any thoughts on that?

We also have a static ip-configuration of the two (cascaded) RM32's and the CS18.
We do use the MOTU AVB switch and in addition to that a router with DHCP activated for connected laptops and wireless devices.

Even last week it happened again (although with the BETA firmware, but it also occurred with the previous,'stable release' firmware installed):
the master got an address from the DHCP pool, different from the assigned static one.
The (maybe) weird thing was that cascading was (still) active, audio went through, but there was no way to control the mixers!
After a while, maybe after 10/15 minutes, without doing anything in the meanwhile, suddenly the CS18 became alive and made a (re)connection to this cascaded configuration.
Detail: The master RM still had a different ip address from the dhcp-pool instead of the static assigned address.
I can really assure you that it has nothing to do with a bad cable or netwerkconnection!

So, it certainly doesn't only occur with the Presonus AVB switch (but maybe we're talking about different technology errors or whatever, which seems to have the same results).

Right now I'm still testing the current beta firmware...

Thanks for your reply on this. Looks like we had a similar thing going on. I am bit reluctant to install the beta firmware on the devices.
According to a ticket I opened with Presonus it had to do with the autosave of the configuration. If you power everything down during the autosave things might get corrupted. So I am testing this now to make sure nothing is change on the board for at least 5 minutes before we switch everything out. Furthermore I have made also a static reservation in the DHCP scope for the equipment, so should the setting change from Static to DHCP the gear is still getting the same IP address. Hope that this will remediate the problem.

Maybe or maybe not has it something to do with the autosave function.....
A reboot of the RM32's brought it back to the original setting: the static ip-address appeared. I did not change anything to the configuration.

I also have had twice that 'lucky' coincident of powering off the RM32ai while apparently it was 'autosaving' the current settings. After switching on the RM32, all settings were gone !
That's something Presonus should also work on. This is unacceptable in a live environment.

If possible I close all the connections (UC Surface/CS18/QMix) to the mixers first, and just as you say, wait for a couple of minutes and then shutdown the mixers.

It's a good idea to define a static DHCP reservation in your router. It possibly masks this problem, but it might be a practical solution to this.

The "Autosave Problem" is specific for AI consoles, not for AI rack mixers. And anyway the intervall is only 45 seconds, so you will be "save" in one minute, not five... In general it's always good to wait a little bit before switching of the mixer. You might also want to save the scene or settings, so there is no need to hurry.

I placed a link to this thread in the Beta Forum with the hope that somebody of Presonus takes the time to read the posts here, which confirm several threads about bugs. But we as users need to deliver reproducable steps to prove that there is a problem. Otherwise the answer from Presonuns is "we can't reproduce the problem".

Besides the IP adresses problem the first post here describes the other problem: "So channel 1-32 went nuts and 33-64 was still ok." This is exactly what I have seen also, but you should check this again. My experience is that in this situation the Slave looses all audio on the outputs. It shows and processes incoming signals, but there is no audio output in Mains and auxes.