Unitiserve into nDAC ; Toslink vs Coaxial ?

Posted by: PBenny1066 on 17 September 2011

Has anybody compared Toslink vs coaxial from Unitiserve into nDAC ? Hope to make the comparison myself soon, if I can borrow a Toslink, and report findings, butnin then meantime am curious.

Also do Toslinks vary in sound quality or can i just get any old cable from Dixons for the sake of this comparison ?

Cheers, Paul
Posted on: 17 September 2011 by frankster_666

Hi Paul,

 

I already wrote it in another thread. 

1) Toslink is in principle inferior to coaxial.

2) Yes, there a good and bad toslink cables as there are good and bad coax cables (bnc plugs are better than cinch - in principle).

 

Please tell us your findings in due time.

 

Regards 

Frank

Posted on: 18 September 2011 by Guido Fawkes

With my Uniti, Toslink sounds better (not a horibble plastic cable from Dixon though, but a good quality Glass one). With my Sonos in to Naim DAC I still use an electrical coaxial cable, but I don't have a UnitiServe. I would suspect a good quality glass TosLink would sound better, as in principle, it isolates the DAC from noise. 

 

Either way a good quality cable (without paying silly prices) of either type should give good SQ. 

Posted on: 18 September 2011 by frankster_666

Guido, I don't say that Toslink can't be better than coax in certain surroundings. I think it is absolutely possible. It depends on the quality of cables and on the quantity of noise of the source (regarding isolation). A good glass cable is way ahead of a cheap plastic cable. Even between plastic cables there are notable differences. I didn't want to believe this until I tried (because it's all 0 and 1...). I guess the quality of the plug of a Toslink cable could also play a role (e.g. mini plug vs. regular plug). Nevertheless optical digital connection is widely regarded to be more prone to jitter than coax.

 

Coaxial digital cables have to be strictly 75 ohm which is not always the case. The biggest problem here seems to be the cinch plugs, so it is better to use BNC if possible.

 

But in the end it all depends on the individual case.

 

Frank

Posted on: 18 September 2011 by Guido Fawkes

> Guido, I don't say that Toslink can't be better than coax in certain surroundings.

 

Sorry Frank - didn't mean to imply you did. I agree with what you are saying and that is why I use a glass optical Toslink to feed my UQ and a Coax for my Naim DAC - I also agree with you about BNC: I am a member of the Bayonet Neill–Concelman [BNC] appreciation society: I hate RJ45 plugs with a passion (horrible things are the worst part of any Ethernet implementation - bring back 10BASE2). Cinch/RCA plugs are fine, but not as nice to use as BNC plugs. 

 

My computer/USB/S-PDIF is not galvanically isolated from my UQ by design (not mine the manufacture's) so introducing a high quality glass Toslink makes a noticeable improvement. It is really easy to hear. 

 

However, you are completely right that a different set-up may give different results. 

 

One of the best ways I have used for getting music off a computer is APT-X over Bluetooth as implemented by Chord. I wish others had tried this, but it seems to have gone the way of Betamax. 

 

My suspicion that the US would benefit from optical (only a good quality one with decent connectors) is that I regard the US as coming from the PC family, albeit with lots of optimisation by Naim, so I'm thinking - could be a bit noisy - but you have to try it to find out. 

 

What I was try to emphasise to the OP is that we are talking Glass Optical Toslink, not cheap plastic rubbish (I have got one of those and it is dreadful even though I cannot figure out why it should be - so I make no excuse for describing it the way I do). So poor plastic Toslink vs. Chord Co Coax is not the way to compare (the Chord will win hands down). 

 

I think we are saying the same thing, Frank. 

 

Al the best, Guy 

Posted on: 18 September 2011 by frankster_666

+1

Posted on: 18 September 2011 by Simon-in-Suffolk
Over the short lengths we are talking there will be minimal difference between polymer or glass fibres. However the accuracy of the termination, the degree to which it is polished is important. Imperfections here will cause reflections or ghosting. This has the affect of smearing the signal, which in turn with SPDIF signals may cause jitter in the clock recovery. The other issue with fibre is that it is susceptible to light scatter, which in turn will smear the pulse. Now the dispersion response can be affected by the cable geometry, and so if the cable is vibrating due to sound pressure in theory there will be feedback into the fibre affecting the shape of the pulse. In theory a damped fibre with a thick sheath should mitigate this.
Of course there is no such thing as 0 or 1s in electronics, that is purely an abstract modelling concept. In the real world the bits are carried as signals, and this signals will be modified by cables, fibres and electronics.
So in short ensure the connecting equipment has good optical transducers, you use a good quality  plastic or glass cable, that has a flat quality  polished surface for optimum optical light transfer and mitigation of reflections,  and is mechanically damped, and you should be fine.
Simon
Posted on: 18 September 2011 by AMA
Originally Posted by Simon-in-Suffolk:
Of course there is no such thing as 0 or 1s in electronics, that is purely an abstract modelling concept. In the real world the bits are carried as signals, and this signals will be modified by cables, fibres and electronics.

Surprisingly this simple fact is not accepted by a wide range of music lovers which leads them to believe that digital replay is always bit transparent and does not need a special care.

Tell them about the ferrite core on the ethernet cable and they will laugh...

 

What is funny is that professional audio manufacturers do not normally mislead the customers with prescribed stability 0s and 1s but people have a strong tendency to misinterpret some of ambiguous statements.

 

People much prefer to believe in something they WANT to believe in...

Posted on: 18 September 2011 by frankster_666

This is a very good point. The 0-and-1-concept seems to be a kind of over-simplification regarding audio applications. It is used too often as a killer phrase and discussion stopper. 

Posted on: 18 September 2011 by PBenny1066
Interesting replies, thanks, as ever there is more to this digital jiggery pokery than meets the eye. (Interesting - when I typed jiggery, the auto text modified it to "jittery" !).

So the design  trade offs seem to be noise etc from isolation issues on one hand (coax) vs jitter on the other (optical), assuming good quality cables in both cases.

I will try to coax (pun) a decent optical cable from my dealer on loan, and give Dixons a miss.

Cheers, Paul
Posted on: 18 September 2011 by ameden
Originally Posted by Simon-in-Suffolk:
Over the short lengths we are talking there will be minimal difference between polymer or glass fibres. However the accuracy of the termination, the degree to which it is polished is important. Imperfections here will cause reflections or ghosting. This has the affect of smearing the signal, which in turn with SPDIF signals may cause jitter in the clock recovery. The other issue with fibre is that it is susceptible to light scatter, which in turn will smear the pulse. Now the dispersion response can be affected by the cable geometry, and so if the cable is vibrating due to sound pressure in theory there will be feedback into the fibre affecting the shape of the pulse. In theory a damped fibre with a thick sheath should mitigate this.
Of course there is no such thing as 0 or 1s in electronics, that is purely an abstract modelling concept. In the real world the bits are carried as signals, and this signals will be modified by cables, fibres and electronics.
So in short ensure the connecting equipment has good optical transducers, you use a good quality  plastic or glass cable, that has a flat quality  polished surface for optimum optical light transfer and mitigation of reflections,  and is mechanically damped, and you should be fine.
Simon


My U-Serve is connected to the N-Dac by an Atlas Equator Toslink, the 2 BNC Coax inputs on the N-Dac are already taken by the CDX2 digi out and the U-Qute digi out...

I guess I could also try a BNC to RCA Coax for the U-Serve to N-Dac and see if it is better than the optical....any thoughts ?

 

Thanks

 

Anthony

 

 

Posted on: 18 September 2011 by Simon-in-Suffolk

Ameden, its worth a try. My advice is not use rca to BNC converter. If you have to use these connectors have the cable made up, but please ensure the RCA has 75 ohm characterisitc impedance. A few do but not many. Alternatively you could go DIY and use some RG59U or RG 6  75ohm Coax cable and put your own connectors on.. (Wiring these connectors onto Coax if you have never done before will need a little practice)

Simon

 

Posted on: 18 September 2011 by ameden
Originally Posted by Simon-in-Suffolk:

Ameden, its worth a try. My advice is not use rca to BNC converter. If you have to use these connectors have the cable made up, but please ensure the RCA has 75 ohm characterisitc impedance. A few do but not many. Alternatively you could go DIY and use some RG59U or RG 6  75ohm Coax cable and put your own connectors on.. (Wiring these connectors onto Coax if you have never done before will need a little practice)

Simon

 


Thanks Simon

 

the current BNC cables (Qute and CDX2) are 'Mark Grant 1000GHD', so will order a 'made up' BNC to RCA version and compare with the Toslink....

 

 

Thanks for yr advice

 

Anthony