Thought I'd carry on here rather than continue to pollute the USB to S/PDIF thread.
As reported there, I've getting the audio breaking up when playing high bit rate audio through the Hiface 2 coax route into my dac. It's less bad playing the same audio through the USB interface on the DAC, but that's limited to 44.1k.
Power to the pi was suspected, so the PSU has been replaced with a chunkier one and I intend to power the hiface independently too when the cable turns up. I suspected the network speed so moved from wireless to wired, and that seemed to help, and the problems appeared to get worse when the home lan was heavily loaded with streaming video at the same time as I was playing audio.
Today though, by way of experiment, I decided to take the pi out of the equation and replace it with my lad's W7 laptop running foobar2000. I didn't bother with wired, relying on the wifi, but did use the hiface etc into the DAC. And it worked nicely with 24bit/96KHz (one track is still suspect and i thought I could detect an occasional glitch but on the pi it's completely broken up) despite me streaming 2 HD videos over the lan at the same time.
That suggests to me that the NAS and network are OK as are the downstream audio components, which only leaves the pi.
Wondering if it was a processor speed issue, I tried overclocking the pi. That made things worse.
Other than waiting to try injecting 5V into the hiface USB directly, I'm not sure what my next step should be.
Wow, you are having quite a ride! But it still doesn't rule out the PS as a culprit – may have been dodgy if it exploded!! Can you lay your hands on another Pi and PS, even as a loaner?
That'd be my next step – it would tend to rule out the Pi/PS for sure. After that I'd try a different USB/SPDIF interface or even try to borrow a 24-bit USB DAC to totally bypass the interface. I suspect the interface is the culprit if its powered by the Pi – did you say you had tried it in an outboard powered hub? Sorry, I'm being lazy here; should go back and read your previous posts but its late where I am…. Even plugged into a powered USB hub I would not rule the interface out.
Is your Pi an early, 256Mb unit? I have one of those and it does seem more sensitive to USB issues than the later 512Mb model.
Keep at it though, explosions and all…
I'll have to keep this short as I'm exhausted (not from pi)
Have 5 pi (long story), 1 x A, 3 x rev 1 B, 1 x rev 2 B. Currently using 512MB rev 2 B without polyfuses for raspyfi as that's as close as possible to Michelangeloz' development environment (note that 256MB rev 1 B with polyfuses worked fine with RC2, hiface and wireless)
Have been offered a 24bit/96KHz DAC to try but might take a few weeks to arrange.
Tried a Belkin powered hub, but whilst wifi dongle recognised by pi, Hiface wasn't.
USB to 1.1mm/3.5mm DC adapter has arrived so will try that with the power injector when other 2A PSU arrives to replace the one that went bang
Bed time…
Yes, the Hiface is now always connected directly to the pi, just letting Derek know that I'd tried it with a hub but it didn't work.
Hopefully tomorrow it will be directly connected to the pi with a power injector, but I still won't be using the hub
I assume you're powering the Pi via TP1 and TP2 to bypass the polyfuse. Can you measure the voltage at TP1 & 2 under load with the Hiface connected? It is indeed looking like a Pi issue. I'm also assuming you tried the Beta Raspyfi on the rev. 1 B? I have one running fine on the Beta software with no issues.
Only other thing I can think of is to confirm you've implemented the sec=ntlm addition to auto.nas?
Just going in to work, so this will be brief…
Rev 2 doesn't have a polyfuse so no need to bypass.
Voltage across TP1/TP2 under load, both with hi face and direct to DAC over USB = 3.22V. Interesting that I'm getting audio break up with and without the hi face today.
Hmmm. Shouldn't that be between 4.75 and 5.25?
I ran RC2 on the rev 1 complete with hi face but struggled to get the NAS mounted. Never quite worked out why, and as son as I got it working on the rev 2 I stopped looking.
Unfortunately I'm lending my meter to a colleague who is fitting a tow bar to his car so probably not be able to take any more readings for a couple of days.
Yes, the parameter above is in my auto.nas
It's definetely too low… As you said it should be from 4,75 to 5,25 … If this reading is correct, the PSU is not adequate… So it's an hardware issue…
D'oh! just double checked the diagram and whilst i was using TP2, in the low light I had misidentified TP1
Will need to re-check once I've got the meter back.
Haven't got a meter yet, but do have the rest of the power set up.
First thing to strike me is that the 5V injector cable apppears to be connected in parallel rather than isolated, so connecting/powering up the hiface caused the leds on the pi to start flickering; not enough power to boot, but some.
Connecting all up; 5V into the pi plus 5V into the hiface, the sound was fine and no problems.
Then disconnected the additional power to the hiface – no change
Then started loading up the network and streaming video from the NAS and the music starts breaking up. Stop all that and the music starts working again.
So, it would appear that adding additional power to the hiface makes no difference, but it could well be the network capacity or (quite possibly) speed of the NAS that's the problem.
Slowly making progress. I think.
You're up late!
Good. Progress, indeed. Guess you could always 'prove' it by running a USB HDD with some 24-bit music (and it's own power, of course) or try a USB stick on the powered hub – bypassing the LAN for all but the control apps. Unmount the NAS from auto.nas!!
There, now you are allowed to go to the pub….
I know I'm getting on a bit, but I don't think staying up to 2 in the afternoon is that late!
Yes, getting the music closer to the pi is the next step. I've already got music loaded on to both a portable hdd and memory stick, just need to mount those up and try them next time I'm home
Odd, was online when your time was more like 2am and saw the green 'Online' light illuminated so assumed you were back from the pub and fiddling with the Pi!! Must have been the 1's and 0's doing their thing….or my eyes!
Off to try different USB cables (including 5V linear injector cable) to see if I can hear a difference. And whether bypassing F3 alters the sound. And anything else a few G&T's suggest….
That'd be the PC typically being up and running 24x7 with the forum open in a tab on the browser. Having just had my revised monthly electricity payments through, you might see me apparently online a lot less often
Have fun with the cables
No meter still, however I've put some music on to a USB stick, and (once I'd realised that it was NTFS formatted, reformatted it and put the music on it again…) it works perfectly with the higher bit rate music.
So it would certainly appear that the problem I'm having with higher rate music is with the network and/or NAS, not the pi (nor the power supply, nor the HiFace 2 nor the DAC… phew)
Happy dance
I'm putting my complete music collection on my WD removable drive (once I've reformatted that as FAT32) so will have all of it available to play (hopefully) while I work out what to do next. to resolve the "getting the data to the pi" problem.
Well that's good news! Worth the odd pint to celebrate. So what's your game plan going forward?
Recently I indulged myself and bought a Topping T60 amp to see how close it came to some home-brewed Tripath amps driven by a Chinese tube-based preamp (surprisingly good given it cost US$70) and wanted to break it in for a week or so given it didn't sound so great out-the-box. I did not want to run it on my home LAN with the Synology NAS – just didn't want the NAS churning away for a week! So I bought a ZyXel MWR102 wifi router for the princely sum of US$17.99 and set up a completely separate network at home with the Pi connected via cable to a cheap switch. Used a single-drive Synology NAS I borrowed from a friend. Just a thought… Might solve those bandwidth issues.
Know what you mean about the electric bill. I cut mine by nearly 30% by switching the fridge off at midnight and back on at 5am every day. Did it via ZWave.
Still worth going the linear PS route even though you've ruled it out as a possible cause of your issues.
I played with USB cables and DACs over the last 2 nights or so and the most striking difference was the improvement in dynamics and clarity when going from a standard USB cable to one I had reduced to 6" and severed the +5V red cable coming from the Pi. I then used a LM 317-based PS I made on the power wires going to the DAC. Not sure if its the shorter length or the linear 5V (or both) but its definitely worthwhile.
Going to be interesting to see if there's any improvement with the silver Litz wire and the 'gold' USB A and B connectors I am patiently waiting on. Wouldn't have thought USB cabling would have made an audible difference but I guess if other cables do then why not USB? Surprising how some relatively cheap wire and connectors (US$30 the lot) might make a noticeable improvement.
Well that's good news! Worth the odd pint to celebrate. So what's your game plan going forward?
Indeed. I may indulge tonight
Game plan right at the moment is to put the pi back into its pibow case, tidy up the cables and use it as is with the NAS. The problems are only occurring with music at 2700Mb/s and above and at present I've only got 2-3 tracks at that speed to test the system. For the remainder of my music library, there's no problem.
Used a single-drive Synology NAS I borrowed from a friend. Just a thought… Might solve those bandwidth issues.
At some point I'm going to get another switch and co-locate the NAS with the pi, though I'm not sure that will help – my QNAP NAS might be well thought of, but it's got the 800MHz Marvel processor and the Samsung drives in it are definitely on the slow side. Co-location might be sufficient if there's nothing else going on but as soon as someone else in the house starts streaming etc, the problems might come back.
I've got a second 2TB drive that I use to backup the NAS, so I might try creating a pi-based NAS using that such that the music gets streamed from that to the pi and the main NAS is used for everything else. From what I've seen though, people are reporting 20Mb/s from an attached HDD to the pi but only 4-5Mb/s over the network when that same pi is sharing that HDD as a NAS. And that's not a lot of headroom if some of my tracks are 3-3.5Mb/s.
I guess where I'm going with this is that the right solution is to buy an up to date high powered NAS with fast disks and co-locate it with the pi, but they're not that cheap so that's not going to happen!
Know what you mean about the electric bill. I cut mine by nearly 30% by switching the fridge off at midnight and back on at 5am every day. Did it via ZWave.
As I've got an A+ rated fridge, "Economy 7" electricity (which makes electricity very cheap for 7 hours at night) and 4Kw of solar panels on the roof (which should make electricity very cheap during the day – when it's sunny) I really shouldn't have a problem.
Perhaps when I've completed my pi-NAS, solar power logging pi and my VPN network access pi then a pi project to monitor the electrical circuits to discover where the power is going is next
Still worth going the linear PS route even though you've ruled it out as a possible cause of your issues.
I'll add it to the to-do list, Though i think I need to do some more work with powre tools or I'll be sleeping in the shed
I played with USB cables and DACs over the last 2 nights or so and the most striking difference was the improvement in dynamics and clarity when going from a standard USB cable to one I had reduced to 6" and severed the +5V red cable coming from the Pi. I then used a LM 317-based PS I made on the power wires going to the DAC. Not sure if its the shorter length or the linear 5V (or both) but its definitely worthwhile.
As the injector cables were not too expensive, i may try snipping the +5V, but at the moment as the system is working nicely with the hiface plugged straight in to the pi I'm tempted to leave well alone – at least until I've got my meter back and measured the voltage drops with hiface on and off the pi psu.
. Wouldn't have thought USB cabling would have made an audible difference but I guess if other cables do then why not USB? Surprising how some relatively cheap wire and connectors (US$30 the lot) might make a noticeable improvement.
I think a lot of this depends on what the DAC is doing with the data. If it's buffering and reclocking there would probably be no difference, but if the DAC is relying on the input signal to clock the data through its circuits then the differences could be significant (though I'm not going to say better or worse)
Most Users Ever Online: 200
Currently Online:
6 Guest(s)
Currently Browsing this Page:
1 Guest(s)
Top Posters:
Gyp: 82
yoghurtlidlicker: 37
PaulM: 36
icai: 35
Derek: 29
Karsten: 27
tc-fi: 22
Ronaldo: 21
ENSen: 21
chriz: 20
Member Stats:
Guest Posters: 3619
Members: 569
Moderators: 0
Admins: 1
Forum Stats:
Groups: 4
Forums: 10
Topics: 469
Posts: 13163
Newest Members: polo2016, dongdong8, wlpwlp, Sager, ninestab123, lmx66, lmx, jlgg, chenyan, guoyanying
Administrators: admin (401)