advertisement


An epiphany as a result of using a USB / SPDIF converter

"Be interested to hear what that setup consists of. I'm running an RPi based setup personally and I'm pretty impressed with it so far.. The PSU side of it needs work and that's the next step of the project!"


Custom server with Linux OS. I can boot in Windows or Linux and Linux sounds way better. Manufacturers spend a lot of time trimming down OS's as much as possible for SQ, you can't get any more trimmed down than Linux.
Mahussive PS for the server PS - BPS's don't work for the main power - don't know the technicals but apparently computers power requirements work in a way that batteries can't do the job.
Internal SSD with it's own BPS - BPS's are mostly SOTM mBPS's with the exception of one Uptone LPS-1. The SOTM's generally sound much better.
External PS for motherboard.
Fibre optic out to Adnaco outboard USB stage with it's own BPS.
Audioquest Jitterbug's on ins and outs of Adnaco.
Chord Sarum Tuned Array USB cable to Regen ISO with it's own BPS
SOTM ASB/SPDIF converter (upgrade with Femto clock and it's own BPS
Audioquest Wild AES/EBU to Boulder Dac.

It's actually 11 boxes strictly speaking :eek:

I think that's everything :p

Edit:
Every step has brought a noticeable improvement in SQ, I'd sort them biggest to smallest difference like this:

Main PS for server
Femto USB/SPDIF converter
Chord Sarum USB cable
Adnaco USB stage
Regen ISO
BPS for USB/SPDIF converter
BPS for Regen
PS for motherboard
Audioquest Jitterbugs

Interesting - definitely a lot more boxes than I'm running! The BPS stuff is interesting, definitely something I'm having a look at, but the SOTM units are a little over my budget for a Raspberry Pi setup.. My current setup is as follows:

RaspberryPi 3B with standard RPi USB PSU
Allo Kali HAT reclocker separately powered by ifi iPower
Allo Piano 2.1 HAT DAC in dual mono mode
Moode audio 3.8 Linux distro

Sounds very good as it stands. I've got a galvanic isolator HAT board (designed by one of the bods on DIY Audio) on order and after that I'll be looking at the best way to power it...
 
To answer Julf's question as to why it would sound better - I have no idea! That's what I mean when I initially said even if you don't believe it, try it!

I can only assume that the USB interface in the netbook is a poor quality interface directly for whatever reason unbeknown to me. The difference when running a USB / SPDIF converter between computer and Arcam irDAC is profound! Not subtle at all! It makes me think that this async perfect single master clock connection is not all it's cracked up to be.

It also make me realise - get the interface integrity right first before comparing any DACs! The interface in my system is FAR more influential than the differences between DACs.
 
Fwiw, audio played directly from my NAS to my Dacmagic Plus via USB sounds great and is very enjoyable :)
 
"Be interested to hear what that setup consists of. I'm running an RPi based setup personally and I'm pretty impressed with it so far.. The PSU side of it needs work and that's the next step of the project!"


Custom server with Linux OS. I can boot in Windows or Linux and Linux sounds way better. Manufacturers spend a lot of time trimming down OS's as much as possible for SQ, you can't get any more trimmed down than Linux.
Mahussive PS for the server PS - BPS's don't work for the main power - don't know the technicals but apparently computers power requirements work in a way that batteries can't do the job.
Internal SSD with it's own BPS - BPS's are mostly SOTM mBPS's with the exception of one Uptone LPS-1. The SOTM's generally sound much better.
External PS for motherboard.
Fibre optic out to Adnaco outboard USB stage with it's own BPS.
Audioquest Jitterbug's on ins and outs of Adnaco.
Chord Sarum Tuned Array USB cable to Regen ISO with it's own BPS
SOTM ASB/SPDIF converter (upgrade with Femto clock and it's own BPS
Audioquest Wild AES/EBU to Boulder Dac.

It's actually 11 boxes strictly speaking :eek:

I think that's everything :p

Edit:
Every step has brought a noticeable improvement in SQ, I'd sort them biggest to smallest difference like this:

Main PS for server
Femto USB/SPDIF converter
Chord Sarum USB cable
Adnaco USB stage
Regen ISO
BPS for USB/SPDIF converter
BPS for Regen
PS for motherboard
Audioquest Jitterbugs

You could not make it up.
 
You could not make it up.

Have you heard his system and compared to yours?

I think it's great to have people like Thetiminator who are willing to go further than most of us would dream, or consider worth trying. When trying to improve performance you need to examine every aspect and try to iron out the weaknesses. That's the attitude team GB / Sky applied in cycling. Each improvement, no matter how small adds to the final outcome.

Look at the work that goes in to designing TT's and speakers. People often go to the extreme there without such criticism.
 
'Team Sky' or the BCF would implement a change then measure to confirm any improvement, not just believe there has been an improvement.
Keith
 
Fwiw, audio played directly from my NAS to my Dacmagic Plus via USB sounds great and is very enjoyable :)

I understand, perhaps your USB interface is better than mine from an old Netbook computer, or perhaps using a USB audio converter in your system will make it even more enjoyable?

All I can say is, after my experience it's worth a try even on systems that sound "just fine" with USB directly connected. For example, before considering upgrading your DAC in case there's more "under the hood" of your DAC ready to be exploited using a 'good quality' converter. Perhaps not but, but worth a shot I would think.
 
All I can say is, after my experience it's worth a try even on systems that sound "just fine" with USB directly connected. For example, before considering upgrading your DAC in case there's more "under the hood" of your DAC ready to be exploited using a 'good quality' converter. Perhaps not but, but worth a shot I would think.
An eminently sensible suggestion. In my case, the USB option sounded better, but there's no way you can tell what's best unless you try it.
 
Does lower jitter sound "better" to everyone? SPDIF is a horrible interface as there is no way to cleanly extract clock from it
 
Does lower jitter sound "better" to everyone? SPDIF is a horrible interface as there is no way to cleanly extract clock from it

Depends on the type of jitter. Sometimes the impact can be a slight 'blurring' or softening of the sound, which some people might like, finding a clock upgrade to be 'bright' or 'etched'...
 
Jitter is taken care of so well these days in transports and DACs (ASRC, PLL etc) so I think galvanic isolation is the first thing to tick off.

With my DAC I prefer TOSLINK to coax, I think it's due to the above reason.
 
Jitter is taken care of so well these days in transports and DACs (ASRC, PLL etc) so I think galvanic isolation is the first thing to tick off.

With my DAC I prefer TOSLINK to coax, I think it's due to the above reason.
I don't disagree with what you say. I would however add a note of caution. I see people using the term galvanic isolation as though it's a panacea for any remaining issues. GI sounds like an absolute term but the reality is different; it does not totally isolate and isolation varies I believe by frequency. We could do with the specs for the various GI methods used.
 
I don't disagree with what you say. I would however add a note of caution. I see people using the term galvanic isolation as though it's a panacea for any remaining issues. GI sounds like an absolute term but the reality is different; it does not totally isolate and isolation varies I believe by frequency. We could do with the specs for the various GI methods used.

Strictly taken, the term "galvanic isolation" refers to DC, and in that context it is an absolute. In normal usage, it also includes mains frequency AC.

In the case of optical S/PDIF it is an absolute even for very high frequency signals - *no* real-world electrical signal will pass through an optical fibre.
 
I don't disagree with what you say. I would however add a note of caution. I see people using the term galvanic isolation as though it's a panacea for any remaining issues. GI sounds like an absolute term but the reality is different; it does not totally isolate and isolation varies I believe by frequency. We could do with the specs for the various GI methods used.
Well toslink really does isolate galvanically. Aside from that, where isolation is partial, the specs would be interesting if they were meaningful. But that could only be the case if there is a clearly defined problem to solve ie a particular unwanted signal we were wanting to block out.
 
I think it's great to have people like Thetiminator who are willing to go further than most of us would dream, or consider worth trying. When trying to improve performance you need to examine every aspect and try to iron out the weaknesses.

It's just somewhat strange that someone would go through all that effort without making any effort to verify the results in any objective way. It is like a Formula 1 team that would never do any timing, performance or speed measurements, but only rely on the driver to report how fast he felt the car went.
 
Jitter is taken care of so well these days in transports and DACs (ASRC, PLL etc) so I think galvanic isolation is the first thing to tick off.

With my DAC I prefer TOSLINK to coax, I think it's due to the above reason.

ASRC / up-sampling is not bit perfect, and any form of sampling rate conversion does introduce additional artefacts. PLLs can suffer from jittery inputs & clock drift, which can be remedied through good analogue design principles (PLLs are analogue, not digital) and receiving extremely low jitter / clock drift signals.
 
Strictly taken, the term "galvanic isolation" refers to DC, and in that context it is an absolute. In normal usage, it also includes mains frequency AC.

In the case of optical S/PDIF it is an absolute even for very high frequency signals - *no* real-world electrical signal will pass through an optical fibre.
The way the term is often used it seems as if people expect all noise to be blocked by GI. Clearly all noise is not blocked - that's the point I was trying to make.
 


advertisement


Back
Top