Hey @Doug_Sadler,
Thanks for the update. @mjw is quite correct in the above reply.
In the meantime, we were able to review a fresh diagnostic report from your Roon Server, and can see that your DAC is not the problem, and neither is a missing Linux driver.
The UD-507 is detected cleanly by the Pi every single time, as a fully UAC2-compliant ALSA device, TEAC Corporation UD-507 at usb-0000:01:00.0-1.4, high speed, driver USB-Audio, with the full PCM/DSD format list exposed (up to DSD512 / 384kHz PCM).
RoonServer also successfully sends setup → stream → start and tracks play. So the USB handshake is fine.
The failure is happening one layer up, on the network link between RoonServer and the RoPieee bridge, and it manifests as RAAT clock-sync loss. The recurring fatal sequence is:
[zoneplayer/raat] failed to sync sender clock to endpoint UD-507
[zoneplayer/raat] failed to sync clocks with any endpoints..giving up
[zone TEAC UD-507] Track Stopped Due to Error
and separately:
[zone TEAC UD-507] Track Stopped Due to LostEndpoint
[raatserver] [RaatServer ropi4eee @ 192.168.1.81:9200] lost client connection
[raatserver] ... client connection failed. Giving up
[raat] RAATServer discovered: RaatServer ropi4eee @ 192.168.1.81:9200 ← re-appears moments later
The Pi keeps vanishing from the network and then being re-discovered seconds to minutes later. That is exactly the symptom you described: “plays for a short while and then dies,” and “usually it stays dead.”
The two concrete root causes we can currently see:
1. The Pi’s IP address is changing (DHCP lease churn). The bridge appears on the network as both 192.168.1.81 and 192.168.1.93. When the lease flips mid-session, RoonServer’s existing RAAT connection to the old IP dies, that’s your “lost client connection / Giving up / re-discovered” loop. This is the single most consistent pattern in the logs.
2. The link to the Pi is dropping/flapping at the transport level, on top of the IP change. The clock-sync drift figures themselves are healthy when a session does hold, so the DAC and the audio pipeline are sound, the connection is simply being yanked out from under an otherwise-working stream.
Notably, the network path described in the thread — 16-port D-Link switch → 8-port D-Link switch → Pi4, is a classic spot for this. Cheap unmanaged switches with green/EEE (Energy Efficient Ethernet) power-saving, or a marginal cable on that second hop, produce exactly this intermittent-drop signature.
Let’s see if any of the following helps:
- Give the Pi a static/reserved IP (DHCP reservation in your router by the Pi's MAC). This directly kills root cause #1.
- Bypass the switch chain as a test: plug the Pi directly into the main 16-port switch (skip the 8-port), or temporarily run it straight to the router. If drops stop, the second switch or the cable to it is the culprit.
- Swap the Ethernet cable to the Pi and the uplink cable between the two switches.
- Disable EEE / green-ethernet on the switches if they're managed; if unmanaged, try a different/known-good switch.
- Only after the link is stable, if you still see issues — revisit Resync Delay.
We’ll be monitoring for your reply, thank you!