I’m planning changes to my audio setup prior to moving house. As part of this, I’m hoping to minimise the number of ropieee’s deployed (mostly for cost and practicality reasons) and would like to connect 3 USB DACs to the same ropieee. I’m concious that this could theoretically saturate either the Rpi4 ethernet (Roon treating them as 3 separate endpoints that all hit the same network interface) and/or the Rpi4’s ability to drive 3x USB’s at once. For cost and simplicity reasons I’m not talking about complicated DAC’s, these are not my main listening area and will just be a USB->RCA cable into the receiver.
I had even 4 DACs attached and this worked! But after a year or so that PI4 died, now I use 2x PI4 and two DACs attached to each; flawlessly working since!
Thanks @Dr_Klaus_Schmitt, that’s great news. I’m hoping the fact that the Pi later died is a bit of bad luck rather than that the use-case broke it… so far I’ve never had any of them die going back to Rpi1 (although obviously it’s not doing ropieee).
My Rpi4’s (with heatsinks) tend to hover around 48 degrees as soon as the weather warms up, but they don’t get much hotter than that under normal conditions (my Rpi3 gets substantially hotter, mostly due to poor heatsink options). Of course I haven’t tried numerous DACs at once, will have to monitor that.
Wow, 4 DACs, I’ve never run that many, I only have 3 on one RPi4! I haven’t watched the temperatures, but I’ve never had any issues with my passive case that’s basically all heatsink Amazon.com
Also, you shouldn’t be anywhere close to saturating a wired gigabit connection (1Gb/s or 1000Mb/s) or the USB3 port (5Gb/s or 5000Mb/s, or USB2 at 480Mb/s each port). Stereo CD quality uncompressed audio is only 1.4Mb/s or if you’re running high bitrate streams, stereo 192K/24b uncompressed audio is 9.2Mb/s. If you’re using DSD: stereo DSD64 (SACD) is only 5.6Mb/s while DSD512 is 45.2Mb/s. So, even running 4 streams of DSD512, you’d be using less than 20% of the network.
@Jim_F I’ve used 2 at most so far, but with the heatsink case I’ve been using I’ve found ambient temperature has a bigger impact on core temp than whether it’s playing music or not. Perhaps with 3 or 4 that would no longer be the case.
@Marian Yes, good point. My planned USB->RCA cable would be bus-powered, but it’s also a very dumb cable (and not a great quality DAC) so I’d hope the power draw would be small. My use-case is somewhat unusual… I’d be connecting the same pi to the same Denon AVR, via 3 separate “DAC cables”. The purpose of this mostly so that I can use Roon to see each AVR zone as a completely separate endpoint (and combined with Deep Harmony extension it can power on/off the AVR zone). If the USB “DAC cables” sound terrible then I could upgrade the DACs at a later date without changing anything else.
Got it. So the DACs are USB-to-RCA dongles and powered by USB. This is what they have on the Pi’s hardware page. If you suspect it’s not enough, you can do what they suggest and use a powered USB hub:
Maximum Power Output
As with all computers, the USB ports on the Raspberry Pi supply a limited amount of power. Often problems with USB devices are caused by power issues. To rule out insufficient power as the cause of the problem, connect your USB devices to the Raspberry Pi using a powered hub.
Model
Max power output of USB ports
Raspberry Pi Zero, 1
500mA per port1
Raspberry Pi 2, 3, 4
1200mA total across all ports
Raspberry Pi 5
600mA if using a 3A supply, 1600mA if using a 5A supply
For the original Raspberry Pi 1 Model B the limit is 100mA per port.
For what its worth, I’ve been testing with 2 USB DACs and there was no change in cpu temp versus 1 (or versus idle).
That said, my idea of using the cheap USB->RCA cables is a flop… I didn’t expect them to be great, but I also didn’t expect them to sound so bad I don’t want to use them
USB->RCA “cables”? So that means those have a built in (cheap) DAC - how else would that work. Of course these are not playing in the same league as a decent DAC (which surprisingly could be had for little $$, like the TOPPING D10S…)
Even the cheap Topping DACs are more than decent, so maybe those built-in DACs are decent enough. @James_Fitzell, what model are you using? I’d be curious to measure those.
To be honest I hadn’t expected to really notice the difference… the signal path was identical except for the DAC and I figured I’d not hear that level of variation.
Was very noticeably worse compared to my DACMagic100.
I got one of those for about $9 on Amazon and measured it with my Cosmos ADC. THD+N is 0.042% (67.5dB) and stop band attenuation around -70dB. Yeah, not exactly state of the art.
I haven’t watched the temperatures, but I’ve never had any issues with my passive case that’s basically all heatsink 
