I’ve been struggling with my power supply for some time now and I’m hoping to get some sensible advice here (Yes, I know this might be a sensitive topic)…
My setup is a R-Pi Zero W with Schiit Modi mkI, connected to a <1000€ audio system. It reads the flac library from a NAS (also r-pi powered).
Sometimes after a restart of the system, the Schiit Modi is not recognised as output device anymore (also not listed in “lsusb”) and after several reboots and reconnects it might work again.
I used to have a Sony Xperia psu (5V/1.5A) as tested by RaspyFi and I measured the voltage going down to 4.7V at times. So I replaced it for a OnePlus 5V/2A charger that I had lying around. Despite measuring 5.03V on the Pi with this psu, the Schiit is still not always recognised after a restart.
Previously I had the R-Pi Zero, Schiit Modi, a usb hdd and an R-Pi model B connected to one usb hub, powered by a Meanwell 5V/3A psu, and never had this issue. My guess therefore is a power issue.
On paper, the Schiit Modi draws less than 120 mA and the R-Pi Zero has no current limit on the usb port, so this should not be the problem. Even with a decent 1.5A supply there should be plenty of headroom.
A colleague (electronics guy) pointed me towards a linear power supply right away, but these would be +/- €100, and I’m finding mixed opinions on whether it makes a difference on my not-so-high-end equipment. Unless someone can convince me of a linear power supply, I’m planning to save money and getting of these simple switching ones: Allo 5V/3A, 5.2V/2A low-noise PSU, or an original R-Pi 5V/2.5A.
Would it really matter which of those I will pick, if I’m not going for a linear power supply? They’re all designed for 5.2V on the Pi…
And more importantly, can there be another source of my problem? As stated, my 5V/2A charger supplies 5.03V to the R-Pi yet the Schiit is not always recognised. I’m not measuring with a scope so the actual voltage might fluctuate more of course. I would have to measure the voltage on the Schiit, does anyone know how to do this?
But could it be for example that there is a small delay in the boot of the Schiit, since it’s powered through the R-Pi, so that it is not initialised properly? In that case I would be better off with a powered usb hub.
So I tested again this morning. With a brand new Samsung S7 charger (and cable) I’m measuring 4.9 Volts, no connection to the Schiit. Reconnected the OnePlus psu, again measuring 5.01 V and no connection to the Schiit.
After booting without the DAC and rebooting with the DAC connected again, it was recognised and working again.
To me this sounds very strange… Any ideas what this could be?
When starting ,the PI draws more current than when IDLE, this might explain why your DAC does not get properly initialized (and it works after you plug it).
If you want my suggestion: get a PI3 (noticeable improvement on usability on Volumio) and the official Raspberry PI power supply. That’s all you need.
Side note: I was in your same situation with a DAC (particularly power hungry) and I solved it with this: volumio.org/product/ifi-idefender3-0/
So you plug the DAC to the idefender, the idefender to the PI ,and you feed it with a separate MicroUSB PSU. If you’re later wanting to upgrade PSU, that will be a good starting point
For now, it should be solved with the official R-Pi power supply you think? Any comment on the special “low noise” one I linked?
I was also considering the usb hub pHat, but that won’t be necessary then.
For the record: I did not hot-plug my DAC. I booted the Pi without dac did a proper shutdown, connected the dac and then booted again. After that it was recognised again.
What makes the big improvement between the R-Pi 3 and Zero for Volumio? Obviously it has more power, but thusfar I have not experienced problems with the Zero W, playing flac-files via wireless.
Yesterday I discussed this issue again with the electronics-guy at work. He also suspects a power dip during booting of the Pi. The Schiit cannot be properly initialised by this, and thus not recognised by the Pi.
His suggestion however was not to get a bigger power supply, but to split the usb cable and power the Schiit separately. Basically a 1-port hub.
His reasoning is that with one power supply it’s not guaranteed that the power supply does not dip during boot, even with the official ones. Apparently a downside of switching psu’s is voltage fluctuations at rapidly changing load. The other advantage of separating data and power is that at a later point in time, I can always power the dac with a linear power supply should I feel the need for it.
Important remark he made with this: keep the ground between the Pi and Schiit connected, they need a common ground for proper data transfer. In case of disturbances (ground loops), a 47 or 100 Ohm resistor can be placed between the Pi’s and Schiit’s ground.
I’ll Frankenstein this cable tomorrow at work and keep you updated.
Can I suggest that you peruse Archimago’s Musings (archimago.blogspot.co.uk/2017/01 … b.html?m=1) on the subject? He’s very clear on the subject of measuring actuals and is quite rude about the perceived need for linear PSUs. You need not agree with him but it’s very good value…
Sounds indeed as a power related issue. Powering the Pi and DAC independent is likely to solve the issue and can be beneficial to the SQ as well.
Going for a linear power supply to feed the DAC can be argued ( I do it ) but in your case maybe an Ifi nano iusb (ifi-audio.com/portfolio-view/nano-iusb3-0/) can be a suggestion. Next to having separate power for the DAC, you also get cleaned power AND reclocked USB signal into the Schiit.
Only down-side, it is as expensive (200+ Euros) as the rest of your (digital part) of your setup (Pi+DAC+PSU’s)
BTW: Am I right about assuming that your have the normal Modi (2) and not the Uber or Multibit as these 2 already have their own powersupply. So you need a solution to feed the power to the DAC via USB.
Maybe consider trading the Schitt Modi for the Uber version? Might be cheaper way than the Ifi nano iusb. But doesn’t give you USB signal reclocking
@chrisrm9208: interesting read, I’ll check it out!
@PatrickB
Indeed, I have the Schiit Modi, mkI even (the original).
I thought splitting the power would solve the issue, but I noticed interesting behaviour yesterday. I powered down the setup before leaving home for a week. Upon return, the Schiit would not be recognised again, regardless of reboots, disconnecting etc. After a factory reset, it was recognised no problems again.
It almost feels it’s not exclusively power related. If I hot-plug the Schiit, the (short) voltage drop is too large and the Pi reboots, indicating a power issue. But if it is a true power issue, one would assume that the Schiit is recognised or not recognised at random upon reboots, which is not the case. It is recognised after a certain boot sequence as described before.
Anyway, I’ll find out next week. I’ve soldered my usb splitter cable, just waiting for the micro-USB connector to arrive so I can plug directly into the Pi.
At one point I might upgrade the DAC too, but by then I think I will skip USB altogether and move to coax.
I understand that it is important to feed clean power to the Boss DAC, because its circuit will then not be affected by ripple / EMI / etc, and will operate free from line noise.
But is it necessary to regulate the voltage also? I mean, is the DAC sensitive to even small fluctuations (say 5V ±10%) in the line voltage? IMHO the regulator may add its own bit of noise to the outgoing power.
) but in your case maybe an Ifi nano iusb (