XXHighEnd

http://www.positive-feedback.com/Issue14/spdif.htm

In my case I profit very much from using a 5m spdif cable and using no usb cable for my hi-face.
Greetings
Adrian

Or have a *real* 75 Ohm impedance ...
(which usually can't be achieved)

So, longer cables make the reflections (of wrong termination) arrive to late / die out underway. But I'd say the signal itself degrades with it and it's just a wrong solution. Oh, it can workout for the better, but arrange for the real 75 Ohm would again be better.

My 2c.
Peter

Aha! From the posts above, now I understand why ...

I read that jitter was susceptible mainly after USB-SPDIF (Hiface) conversion and that the data through USB to Hiface was robust. So thinking myself to be smart I bought a decent USB cable (female end) and connected the Hiface directly to the DAC with a male/male RCA adapter. I was expecting magic and I was very wrong.
The connection of the Hiface with SPDIF to the DAC was just superior. I couldn't understand it...

So I took it further... and connected the Hiface directly to my Laptop and DAC (no cables at all but with male/ male RCA adapter - good quality). The SPDIF was still way superior. So presently I am just sticking with the SPDIF cable.

Thanks for sharing, Flecko!
Raj 

I tried a b test.
a: Direct plugged to the dac with 5m usb and
b: with 5m usb and 5m spdif.
Unfortunatly my system was not good enough in this config to make a clear statement. The long usb cable decreased the sound of my outofthebox hiface very much.
But I can say for sure, that it is better to use a long spdif (50€ type) cable on the hiface than a long usb cable (10€ type).

I didn't wanted to say this is the best or only way. Just an inspiration for people who have a similar setup as I.
In my case, I have about 4m from pc to dac (spdis input). If you consider boundary conditions like that, you just have to get there somehow. And in this case it seems to be good.
It is questionable if one can manage such a tight design. Even the best cables have about 0,5% to 1% deviation from 75Ohm. And the designers of the electronics are aware of this problem (at least some of them) and try to go as close as possible to 75 Ohm with theire input and output design. There are also different approaches for that problem. Audiophilleo tries to avoid the refelction problem with a very fast output stage. Art Legato gives a 5m spdif cable to his usb/spdif converter. Both serious designs I think.

@Raj
I think the hi-face power supply also suffers because of the usb cable, especially if they are long. I tried different length from 0-4.8m and I would say, the longer the worse for usb-powered devices.

For my Hiface with BNC I tried a 10m cable with RCA's and adapters and it sounded more silky.
If you use the  RCA-fitted Hiface then 75 ohm's is impossible to achieve so then I guess a long cable is always a good option.

Another improvement is a 12- 20 dB attenuator as proposed a.o. by JKeny.
It is supposed to attenuate the reflected signals as well as the main signal but the main signal is still strong enough because the hiFace gives enough juice.
and the battery option sounds worthwile..

But you do know why, don't you ?

...

Because the HiFace has a way too high output level, so the reflections are also way too high, and now can easily be seen as falling/rising edges ...

So, Flecko's (or Steve's) general idea is still good, but the HiFace begs for it. It's just wrong to this respect.

Peter

I tried this tweak. I chose a 6db filter because of two reasons.
1: The shop just had that one, and it isn't actually a attenuator but a splitter which reduces the signal about 6db and
2: the other thought was, that a 6db attenuator already reduces the signal to 50%, 12db would be 25% and 18db 12,5% of the original signal and that seems to be a little much.

I can say it changes the sound and I would agree with Keny, it is smoother. But smoother is not necesarily better. But in this case it seems to be better because the stage becomes more 3D. I recognised no loss of bass but I am not sure because of the dynamics. Without that tweak it sounds a bit agressiv and this is so far, not clear to me if this is good, if you think about dynamics or bad if you think about distortion. Without the attenuator it sounds a little distorted to me. I will listen some days and then decide if I order a proper attenuator.
If this turns out to be a good thing, there is the question if this is a hiface specific error or just a common problem with spdif signals. It might be both.
Greetings
Adrian

After some more listening I can quite sure say, that the 6db attenuation will help the hiface giving better sound. It is more involving, softer but no loss of details, a little more musik in the room and more 3D. But I will not buy the attenuator. I decided to buy an ART Legato. It is a much better design and will not need the attenuation. As peter said, the attenuator helps the hiface because it has no perfect design. But with a better designed device, reflections can be much lower. Example (from Pat): If you have 20% reflection from your dac input and 20% from the source you will get 4% reflection to the input to the dac again. The 6 db attenuator reduces this to 2%. Without an attenuator and a propper source, in this case the legato, you have 0.13% reflection that get into the (still 20% reflecting) dac ->no attenuator needed. It is not said that the hiface is so bad, that it has 20% reflection but it will be far from the legato. So I ordered one.
Greetings
Adrian

You checked this site: http://www.audiophilleo.com/comparison.aspx  ?

Thanks but I checked it already and think from what I read that the Legato will perform better if you just looking at the main function, "converting" usb to spdif. The audiophillio 1 has nice features but nothing that I need. A direct comparison of the two devices would be interesting but it is not available in the net. At least I found nothing.
Greetings
Adrian

Just want to update this thread. I am ill (damn weather) and have time to play with my hifi-system. I planed to buy the legato but it was not available for al long time, things changed and I canceled the order. One interesting thing is, that upsampling was never really working for me and I had the best sound with 44.1/16bit. 88.2/32 sounded refined but also laked dynamic. It was one step forward and two back. One day I tryed the 24bit setting and felt it has the analog sound like 32bit but is as dynamic as 16bit. That was a nice improvement and I listened to 44.1/24bit. In the last weeks I tweaked the bnc input of my dac. I now use a " bnc feed through jack". I connect a bnc cable from the outside of the dac and instead of soldering twisted pair at the inside of the dac I can now connect BNC also from the inside! So I can go 75Ohm straight to the board. This was a quite hearable improvement. I tell this because I think it explains why today, as I am experimenting with xx, doubling has become an option for me again. Before this tweak I lost dynamic and gained a refined sound. I didn't liked this. Now it sounds refined but without losing dynamic. (Maybe the upsampled signal is more sensitiv to the transmmision over spdif?)  But if I get best sound with 88.2/24bit why buy a 44.2/16 bit only device? Doesn't make sense anymore and that is why I will try the audiophilleo2 instead. Should be able to improve over my standard hiface.
One thing lasts. 24bit ist still better than 32bit. If someone like to try this out and tell me what he hears, it would be interesting.
Greetings Adrian

Are you talking about 'DAC needs' setting in XXHE?

What I heard from those who can choose (which are very few !), they tell the same.
Theoretically it implies 3/4 of the "system load" in almost everything. So that can be a reason ...

Peter

Yes


Aha!
Thx for the quick answeres.

Very strange. Even ArcPred works now. It alwas had a better focus on the instrument but added a strangeness to the sound. Now it is just incredible how much realness it provides. I cleaned my system with tuneup and using a function of tuneup that is called turbo mode. This defenitly improves the sound. I would try vista but the cd I have doesn't install. Don't know why. But sound is great even with w7.

Flecko, with your Audio-GD Ref 7 DAC, are you bypassing the oversampling filter when you use Arc Prediction? I think you should...

Mani.

never thought about that...can I do this by just switching the jumpers? Do you know where I get the information how to do it? Great Idea!

I found the settings... GREAT.. I will play :)

It's just the little dip switch in the unit, isn't it?

Try to use QAP (i.e. 4x) upsampling if you can because DAP (2x) isn't going to be high enough without extra filtering. I think QAP is the filterless lower limit, with OAP (8x) or higher being the ideal.

In any event, keep the volume down as you're trying these things out because you may get some HF stuff coming through (depending on your system).

Let us know how it goes.

Mani.

The Reference 7 won't allow for QAP as it is limited to 24/96 AFAIK.

WOW! That was a great hint  :o :toomuch: :clapping: Highs are so much more present and real now...must listen a little longer there are different settings...1xoversampling and nos ...what is the differents?

OK, be very, very careful!

If Marcin is right, then you are now using DAP with no filter in the Ref7, right? This may be too low to go filterless - maybe Peter can give his advice.

I strongly advise against not engaging any filter at all either in XX (AP or AI) or in your DAC. You will have a lot of HF stuff coming through. It may sound quite euphonic at first, but it's not a good thing to do... Peter often quotes 30% THD, etc.

Mani.

here are the settings one can make
http://assets.head-fi.org/c/c1/c19b166b_DSP1+pin+settings.JPG

Yes.

there are two settings one can make.
dip 7 is bypass

and dip 4,5 one can set oversampling from 1X to 8x

The differents is not realy clear to me. is 1x oversampling still oversampling? It sounds not as bright as bypass.

Jep, QAP doesn't work.

ok thx ... i will go back to the 1x os settings.

This is totally crazy. I don't see any reason why Audio-GD have limited the non-USB inputs to 96KHz. Maybe because of the filter... but this can be bypassed!

Mani.

It would be great it that was possible :)

Ok. First results :)
no filter is like mani sayd no good idea. it has more present highs but after some time you know there is something wrong.
With the oversampling it is different.
Standart setting 8x.
The first new setting I tried was 1x os. The highs become much more dynamic and brighter (people sayed it should be the other way round..because earlyer roll off.)
drums sounded much more real. 8x sounds also nice a little bit layed back and rounder.
At the moment I listen two 4x, which seems to be a good compromise.
Maybe because 2usx4os=8x :)

Any Idea how to do?

I think you should direct that question to Kingwa from Audio-GD.

Adrian,

Maybe it is only me who doesn't understand or can follow ... maybe it is you ... I am not sure;

You should try No oversampling and No filtering in the DAC, and next use Double Arc Prediction. That is, this is what everybody is advising you (I will too). If that doesn't sound good to you, try the oversampling in the DAC.
It now looks as if you only tried the latter.

Btw, 96KHz "only" can have a good reason; it depends on the design. And indeed the Audio-GD does that (96KHz) AFAIK.

Peter

PS: And what also counts : Double Arc Prediction (without filter and oversampling in the DAC) officially is not enough. Whether you perceive this highly depends on the amps but also other "perceiving mechanisms" in the body. The harmonic distortion will start right above 24KHz for 48KHz base material or 22.050KHz for 44.1 base material. The peaks of it will be very close to the fundamentals. So, officially no good.
Also : without anything (filter / upsampling) those peaks will be right in the audio band. And on another hand : many people are not bothered by that.

Thx... now I get it. I was to excited to get a clear thought. Ok there is the oversampling and the filter. I can switch the filter on/off and adjust oversampling from 1x to 8x. Ok. What I did:
Changed oversampling rate from 1x to 8x with "filter=on" (best was 4x os)
switched the filter off and 8x oversampling=on (I thought the oversampling would be off if the filter is off)
I will try to disable both and go up with oversampling if it sounds bad...

Best sound is with DAP and 4xOs with filter=on. without filter it sounds grainy no matter if I use 8,4,2,1x Os.
After all. Today I could do a lot for the sound, very exciting. Must be my gift from Santa :innocent: Claus.
THX All

From a purely theoretical point of view, I would have thought that DAP-in-XX + 4xOS-in-DAC would have given you the best sound (8x oversampling in total). As Peter said, just DAP-in-XX with filter-off-in-DAC will give lots of HD near the real signal.

With filter-off-in-DAC, I would have thought that the oversampling switches (1x to 8x) would make no difference whatsoever. I mean, the filter is off, after all. And this is what you seem to be finding.

In any event, hopefully DAP + 4xOS is giving you good sound. And count yourself lucky that you have a true multi-bit DAC (i.e. no massive 64x or 128x oversampling) with which to try these things.

If the Ref7 could accept 176.4/192 signals through its BNC input, I think it would make it a very attractive proposition. I bet it'd work really nicely with a BNC-hiFace, XX set to QAP and the filter switched off in the DAC.

Mani.

I am back to 8xOs. There are "more" highs with 4xOs and it sounds more dynamic (agressiv?) but it is also fartiguing. There is something wrong, something that bothers my ears. 8xOs sounds finer and renders/seperates each instrument more. As always, it needs a little time to judge.

To be honest, at last, I kicked out ArcPred again. Don't know what was yesterday :) Maby to sick to hear. ArcPred has something but, there is still the awkwardness... However, now all is back to normal except that I kept the doubling.

That I don't use ArcPred doesn't make it a bad thing. It is not ment to work that way, it is for NOS-DACs. Justed in case "someone" misunderstands this.

Not me ! :) :)

The more you seriously report on this, the more I will learn from it !

Thanks,
Peter

I actually ment the million $ man ;)

I just saw this thread now & would like to add my bit to the discussion.

Firstly, long SPDIF cables work by delaying the signal reflections so that they no longer arrive back at the DAC at a critical moment. They do NOT work by reducing the signal or the reflections by any significant amount. The correct length of SPDIF cable has to be calculated - it is no simply a case of all cables over a certain length will work!

Secondly, Rf attenuators work by reducing the SPDIF signal AND by reducing the signal reflections by TWICE (because the reflection has to pass through the attenuator twice on the way to he DAC).

Yes, it is better to have a true 75 ohm impedance all along the SPIF signal path - this includes transmitter, output stage, connectors, cable, receiver input stage. If any of these are not truly 75ohm impedance you will have a signal reflection. So even if you have a well designed impedance controlled DAC, you can't control the reflections arising outside of the DAC.

I believe the RF attenuators are a useful way of reducing these reflections & hence the jitter resulting.

Yes, the Hiface has a high output SPDIF signal which could be considered a flaw but it also allows us to use a higher attenuation & thus a higher reduction of jitter.

BTW, the smoother, more analogue sound reported with using the Rf attenuators is, to my mind, a better, more realistic reproduction of what's on the recording. We often think more pronounced dynamics are better BUT I believe this is one of the mis-conceptions that listening to digital audio has given rise to. We think this is the detail of digital sound but in fact, I believe, it is the distortion of jitter that we are hearing & interpreting as "better dynamics"
 

That is true. But I think there is also another effect that comes with longer cables and that is very important. The reflected signal is not correlated to the signal if it arrives after a certain time. And that effect, you have simply after a certain length.

You can, by using good cables, connectors and a well designed source. You surely not get rid of reflections completely.

Yes, they can have a positiv influence but a well designed dac and source will be better. If you use 12db attenuator (and the dac still locks), the reflections are reduced about 12db. If you have a propper source, you can reduce it about 40db. That is a difference in signal strength of about factor 1000. That means the reflection is 1000 times higher with the attenuator compared to a propper source. It helps, but it is more like using crutches indstead of having healthy legs.

Jep :yes:

You mean it is no longer data correlated jitter?

Yes, exactly, you cannot get rid of all reflections completely & when you use a device with a faster rise time SPDIF signal (which will be more accurate) you will give rise to even more reflections. So why not use a simple & cheap way to deal with them - attenuators @ $12

The attenuator reduces the reflections by 24dB in the above example i.e a factor of about 1/16 whereas the signal is reduced by 1/8!
How many well designed transmission lines are there? From the article linked to:Also lots of people use RCA connectors - attenuators are the only answer to this impedance dilemma, I believe. 

It has still the same footprint but as it arrives some periods later it is not that correlatet to the data anymore as if the reflection comes in the same period of the signal from that it was reflected from. It has a more random character.

You are right but then it is still factor 100. I do not want to say that attenuators are not a good way to reduce reflections for a given system. It just should be done right from the beginning. Instead of using RCA one can install bnc very simple by replacing the in/output jack. But why doing all this if you can get propper designs? It will never become that good.

Yes, good point as data correlated jitter is the most offensive sonically but I don't know if this would be considered random jitter & sonically benign? BTW, a very short cable also answers this way of avoiding reflections arriving during the decision window!

Yes but in the case of the Hiface we can go upto about 20dB of attenuation so we have reached your best 1,000 times (40dB) reduction of reflections . That's why I consider the design flaw of the Hiface to be an advantage when using RF attenuators - all for $12, not bad, considering what well designed impedance guaranteed (1%) cable in the lengths suggested would cost (& this is only the cable end of things - a well designed transport & DAC is going to cost you of another wad of money. I'm all for doing it "right" but what is the expense of this?

Normaly the signal gets terminated by a serial resistor. Not the best way, as far as I know , but it is usually done like this. Of course you have to stay with 75ohm for the whole signal chain. I managed to go from the board of my dac to the hiface with 75ohm. The improvement sounded a little like having the attenuator in the signal chain.  It also  should lead to the same characteristics as it improves the same thing.

Ok, good point and good to know that it works. I had an 6db attenuator in my system. 20db should be even better. But still the hifacde has 45ps jitter figure according to Philip Gruebele. Do you use also better clocks in the hiface mod and can this jitter be reduced due to this?

Yes, not the best way, so you are immediately settling for less than he ideal - maybe a band-aid is needed :)! I doubt you got the same improvement from it as you would with 20dB of attenuation. Have you measured the characteristic impedance of the transmission line?

45pS of jitter on the output of the stock Hiface is damn good - what are you comparing to? I deal with the data correlated jitter in my mods to the Hiface by addressing the PS to the clocks among other things. I haven't had the jitter at the output tested but sonically I can hear a big improvement.

No, wish I could :) But I done it as good as possible. Before it was done really badly. I am sure the attenuator will improve the signal further.

It is not bad. Philip Gruebele (audiopilleo) claims for his device 8ps of jitter and Pat from Art even 4ps. And of course Peter with ultra low jitter of 0.5ps. But this are also just numberes and there is more stuff about that jitter problem like clockphase noise that is important. But here I must trust the experts and that they really know what they do. It needs a lot of experience and knowledge to make these things right.

That's my point - how do we know if the characteristic impedance of the device is correct? I's not simply a case of putting in a 75ohm R, as you know. You might just have substituted one set of reflections for another :)

Yes, these are just numbers & as a number of people say it's the type & spectrum of the jitter that is important. So 0.5pS may not be better than 8ps, who knows? Steve Nugent certainly claims that a 2ps clock sounds better than his 0.5ps one. I believe that most of the issues with clocks start with the clock PS - most don't seem to get this right. 

I think you can say, that if the signal is terminated by 75ohm+-1% resistor, the impedance that the cable sees is 75ohm+-1%. And the cable itself has again 75ohm+-1% if it is very good. And here you will have reflections. You never get lost of them but doing it like this, is way better than just use a rca connector with some random impedance and normal wires that also have just random impedance. I think I got your point when you say a attenuator is alway better than non (isn't it?). This should be in most situations a good thing. But on the other hand, the attenuator itself induces refelctions again. And IF your system would be really a great design, it could get worse.

Jep  :yes:  Clean power seems to be a golden rule in audio.

Depends on the receiver input circuit! The signal traces that come after that also need to be impedance controlled. A more linear stage is also better, anyway! Agreed Can you tell me where/how the attenuator induces reflections?

Jep  :yes:  Clean power seems to be a golden rule in audio.


[/quote]Even more critical in digital audio & even more crucial in clock PS - just read what Jocko (Pat) et al have to say about this.

Assuming that the signal path is very short it becomes less significant but yes.

They have themself an impedance that dissents from 75ohm, thus causing reflections. Also there are other ways how they can manipulate the signal, I would assume, like how constant they reduce the signal according to the frequencies you send through. There are different qualites of attenuator made for different frequencie ranges. There are also attenuators for 30$ or even 120$ made for higher frequencies and better quality.

have a look at the minicircuits PDF on fixed attenuators to answer/correct your above questions/statements which is what I recommend - $12 http://www.minicircuits.com/pages/pdfs/an70001.pdf

I see no contradiction to what I wrote. Can you name one?

You seemed to imply that these precision made attenuators (which have data sheets & graphs of their VSWR) introduced some impedance mismatch but I believe that this is a red-herring unless you can show a measured digital audio system where introducing these RF attenuators would have a detrimental effect!

Anyway, point is these are $12 & alleviate/solve a number of difficult issues. Best bang for the buck that I know of!

Even if they have exactly 75Ohm, you will have reflections, because your cable will not have exactly 75ohm! Impedance changes->Reflection
I have worked with an analyzer once and we had attenuators that were in the range of 50$ (I was surprized that they were so expensive). And as sensitive as audio is, it might be an improvement to use "better" ones. But maybe not. I think you have chosen good ones, no doubt.

Thank you, yes they are good ones.

In answer to your point about them having their own reflections - a quote from the PDF I linked to - the first lines read

And another which shows that there is no such thing as an ideal transmission line in the real world:
As I said, I think your statement is a red herring, however if you can nominate a system in which the attenuators would be detrimental, I would stand corrected.

By now ... somehow this comes to me as catching flies. :)
So, I thought to change the subject a bit ...


... who took that from somewhere else. Besides I won't believe it. And merely, what's the phase noise ?

So, John, what oscillators are in there ? (it may even be a synthetic thing, I just don't know).

Peter


PS: I just wanted to say something to deviate from the subject a bit, because otherwise you may end up fighting. Haha.

Don't worry Peter, no fighting from me!

Here's a Phase noise plot of the stock hiface (http://dc234.4shared.com/img/P6ltsA9D/s7/0.8058969533756806/Stock_hiface_Phase_Noise_plot.png)
taken by JosephK & posted on DIYA here http://www.diyaudio.com/forums/digital-source/168901-rf-attenuators-jitter-reducers-30.html#post2352984 (http://www.diyaudio.com/forums/digital-source/168901-rf-attenuators-jitter-reducers-30.html#post2352984) He says this about the above plot "The spdif bit rate at 44.1khz is 2.8224 MHz. Half of it is 1.4112 MHz, and that is the strongest frequency line if you look at an SPDIF data stream on a spectrum analyzer.
So probably this is the jitter caused by clocking out the data..

The tall line at the bottom of the spectrum is the ~80kHz power supply noise."

He goes on to say about some of my scope shots of this 80KHz PS noise in the stock Hiface & how they correlate with his plot above "Now, because of the correlation, I am quite sure that removing that power supply noise component from the clock, the recovered jitter would be much different.."

The clocks used are MEC clocks which are good performers - a 22.xxx & a 24.XXXMHz ones to deal with the two families of speeds!

Thanks John. But your picture doesn't work and the link doesn't either. :)

Doe sit work now?

Yes, both work now. Thanks !

BTW, seeing as we are talking about RF attenuators here are before/after scope shots of the stock Hiface SPDIF output
Without Attenuators:(http://dc181.4shared.com/img/kWM6gbim/s7/JosephK_Hiface_noatt.jpg)
With Attenuators:(http://dc252.4shared.com/img/Y8yhMnlu/s7/Joseph_K_Hiface_with_att.jpg)

The initial overshoot is because of the SPDIF output transformer & attenuators can't do anything for this! You can see a significant reduction in the reflections, however.

I do not want to say you are not correct. I think too, that a system where an attenuator has a negative impact is not easy to find. Maybe don't exist. But theoreticaly it is possible that the attenuator has a negativ influence because you can have a mismatching impedance between the attenuator and the cabel/dac/whatever. It might be realy small. One could calcuate that but I am to lazy for that now. We do not need an ideal system. Just a cable that has for example 76.ohm and a attenuator that has 75ohm. The impedance suddenly changes and this causes reflections. If the system is so good, that this reflections are higher than the reflection of the system itself, the attenuator will have a negative impact. But this is just theory and may be not the case for 99.99% of all hifi systems.

But without the attenuator the cable impedance of 76ohm will meet the receiver input circuit impedance of 75ohm & have exactly the same reflection. How does the added attenuator make this any worse? I don't understand?

Good point. Ok, than this just "works" if the attenuator has a worse impedance matching than the cable. This may never happen. You win :)

I still don't understand what you are saying but any reflections generated will be "doubly" attenuated as can be seen in the scope shots

Take a more or less ideal system. Cable has 75ohm and the attenuator has 75.5ohm. If the attenuator is not plugged in, you have close to zero reflections. if you now put in the attenuator, the signal gets reflected when the impedance jumps from 75ohm to 75.5.ohm. Then, when the attenuator meets the input of the dac (75ohm), impedance jumps from 75.5Ohm to 75ohm, again reflections. The reflections get damped but before there was even a lower amount of reflections. Now the signal is worse.

Sure, you can imagine all sorts of scenarios but how representative of the real-world are they? As stated already in that pdf I quoted, there is no such thing as an "ideal" transmission line! Are you saying the minicircuits attenuators are 75.5 ohms? I'm sure you can find a "cr*p" attenuator somewhere on ebay or in Jaycar but this isn't really what we are talking about! Just stick to minicircuits ones with VSWR close to 1.0!

Let's leave this it's getting nowhere!

Fine. I think there was some misunderstanding. I never wanted to question what was wirtten in this document or your findings or the quality of the attenuator. I just was looking for a situation (in this case completly theoretical, no reference to real world as it turns out ), where the attenuator might be bad.
I am going to sleep now.
See you!

Pardon my joining the discussion at this late moment, but we have just became aware of it.

First, the article at the beginning of this thread is from 2004. I do not know the author, but I believe at one point, he has stated that he no longer uses longer cables, as he now builds gear with a faster rise time.

Indeed, as rise time goes down, you can use a shorter cable. We used to supply a 20' cable, with a DAC we made (around 20 years ago), and as most of the sources were very poor, we adapted a long cable,  to minimize the effects of reflections. As has been pointed out, longer cables do not reduce reflections: they merely delay it to a point where they can do less harm.

How much harm can they do? It depends on a lot of variables. Impedance of both the source and receiver, as well as source rise time, are the main factors. I seriously doubt many folks here have the ability to measure those, so let me make a blanket statement, that can be used as a reference point.

Rise times can range from 3-4 nSec, on the high side (typical for HC logic chips, found in most stuff), to under 1 pSec, for "hot rod" stuff, such as you might find from small, niche manufacturers.

Return loss can be as bad as 8-10 dB, on the high side (yes, we have measured stuff this bad), to better than 40 db, on the low side. Stuff this low is not the norm. While we do not maintain an up-to-date chart, of every piece of gear, let's settle on 20 dB, as a rough value. (May be optimistic.)

So, let's start with rise time. Assuming you have a cable with a 70% velocity, it takes around 1.5 nSec, for the signal to travel 1'. So, 4 nSec rise time...........that is around the equivalent of 2.75' of cable. Or, around 1.4', for a round trip. (Remember, reflections have to make one full, round trip more than the incident pulse.) This means the trailing edge of the pulse will arrive at the same time as the first reflection, of the incident part of the signal. IF you have a 1.4' cable.

Not a good scenario.

If you desire that first reflection to arrive at a point where it can't muck stuff up, you make the cable a lot longer. So, there you have the genesis of why some of us use long cables.

(Actually, 20 years ago, before there were a lot of niche companies, making transports, the ones produced by the multi-national conglomerates were slower than this. They were only concerned about meeting regulatory mandates. So, by slowing down the signal, it made compliance much easier. Long cables were almost mandatory.) Obviously, if you have a faster rise time, you can use a shorter cable.

OK, someone is going to say long cables have more attenuation and will spread the pulse more, and, blah, blah, blah. 

Yes, they do. How much? A lot less than one might think. We have equipment that will not only allow us to measure how much the pulse is slowed down, but also how much it adds to the jitter. Let's just say this is overblown. The changes are negligible, for the lengths we are talking about. So, let's drop that part of the debate. We have measured some really slow cables, and for these lengths, it is not enough to lose sleep over.

So, if you a product made by a big company, like Logitech (which makes some nice stuff, at a good price), you may want to use something longer than 1 m. If you are using something from guys like Audiophileo (did I spell that right?), or Wavelength, their stuff is a lot faster, so you can get by with something short.

Ok..........but just how much do reflections harm us?

You have to know the return loss, of both ends, to come close to knowing.

Let's stick with that 20 dB number. (It makes the math easy, even if it is optimistic.)

20 db is 10%. 10%, on both ends. Which means you have 0.1 x 0.1, and end up with 0.01. Or -40 dB. This means the first reflection is down, -40 dB, relative to the signal.

Sounds like not enough to worry about, right?

As an engineer, I would concur. However, extensive listening tests say "Not so fast, bub."

Let me provide an anecdotal explanation, you know, the kind that will get laughed at, on certain DIY forums, that might shed some light.

A buddy convinced me to re-work the input, of his DAC. Reluctantly, I gave in. I forget the final results, but I am sure the input impedance was better than -20 dB. I know his source was better than -40 dB. So, we have 0.01 x 0.1, and get 0.001. That ought to be low enough.

Right? Wrong.

"Should I put an attenuator on the input of my DAC."
Save the $12."
"But should I...............? Can I?"
"Well, I measured that you can add over 20 dB, to your DAC, and it will not lose lock. Your money; do as you see fit."

Later.............

"I bought a 10 dB pad, and it made an improvement!!"

I'm glad for him.

So much for science.

Didn't you just get through telling us your measurements show that extra cable length won't hurt?"

Yes.

"And then you lay this story on us?"

Yep.

"Well, what should we do try?"

Try one, both or neither. Your choice. Just don't make up stuff about why or why not it works.

As I mentioned at the beginning: almost none of you know the answers as to how good (or bad) your gear measures. Don't assume, because it is most likely worse than you want to know. Just assume it is bad, and take steps to alleviate it. If your stuff is good, and your DAC doesn't cr*p out, the odds are you aren't hurting anything, other than your wallet.

ar-t, Pat or Jocko,
A nice informative post but can I ask you this as I know you have the expertise?
Is it not the case that the reflection only has consequence when it arrives during the transition point of the DAC which is somewhere in the middle of the rise-time?
What if the length of the cable is made too long & the reflections hits the next pulse's decision point? What would be typical lengths for this to occur?

Why is a shorter cable not also giving the same effect i.e. the reflection will not be coincident on the decision point & will have bounced back & forth enough times before the decision point is reached that it will have died down to insignificance by then? 

OK.............jitter...............a real long subject.

Jitter is a tough parameter to characterize. To engineers, like myself, jitter has its roots in phase noise. Which, is nothing more than a form of modulation. But, just like any other kind of modulation, unless you know the amplitude and frequency content, it is just a number. Numbers, without context, are just numbers. They look good on a spec sheet, but that is about all they are.

"Don't you give jitter specs for the stuff your company makes?" Do you want us to believe them?"

Uh...........yeah............numbers that look good on a spec sheet. (OK, someone challenged us to a numbers fight, and they lost. But don't tell them that is the only reason it is there.)

But, let's look at how those numbers come about.

Part of the problem is how you measure it, and how it gives the results.

The over-priced gizmo that we use, has an internal 100 MHz clock, as its standard. They spec it as being better than 3 pSec.

Sounds good, right?

Uh, no!

We can calculate the phase noise of that clock (ok, we have to make some assumptions), and converting that value to 11.2896 MHz (44.1 kHz x 256), it isn't that hot. The reference clock would measure around -70 dBc, at 10 Hz offset. (This means the noise, at 10 Hz from the frequency is only down -70 dB.) For a 11.2896 MHz clock, the same jitter number would give us a value of around -90 dBc. (We have some other gizmo, that tells us our clock is in that ball park. But, that is not the point.)

So, the first thing you have to understand is that you also have to know the frequency of the signal that you are measuring the jitter of.

If I told you our clock measured 700 pSec of jitter, you would say it is horrible. But, if you convert that 11.2896 MHz, down to 44.1 kHz (IOW, word clock), that is exactly what you get!

So, when some self-appointed expert tells you that anything under 2 nSec is not audible, and he fails to mention that he is using 44.1 kHz as the frequency that is being manipulated, to determine jitter sensitivity, set him straight. 2 nSec works out to be -80 dBc, for a 11 MHz clock. Which, is a decent clock.

Sort of.

"Whaddya mean 'sort of'. Explain."

When you are measuring phase noise, or any other form of modulation, you have to know the frequency spectrum. The freeware we use, to convert phase noise to jitter, only goes down to 10 Hz. Which is the lower limit on a lot of test equipment.

Actually, some equipment doesn't even go that low. A lot of oscilloscopes, that have built-in jitter measurin' abilities, only go down to 12 kHz!!!!!!!!!

Why? Because that is what a lot of telecom stuff is spec'ed for.

If you look at the companies that make oscillators/clocks, you will see, in the fine print, if you look really close, because they print it really small, it will say that jitter frequency is > 1 kHz.

Well, jitter measured only down to 1 kHz............anything will look good.

Let's take the 100 MHz clock, in our expensive gizmo, and limit the low frequency limit to only 1 kHz.

25 fSec. Yeah, 0.025 pSec.

Ridiculous, right?

So much for specs.................

The other thing that you have to know, and it is probably the most important.............is whether the jitter is Gaussian or deterministic.

Gaussian: just white noise. Exists everywhere, and can not be avoided. Not too horrible. Except............maybe............well, some of us feel the really close in stuff, you know, below the 10 Hz limit, on most measurements, is more harmful than the higher frequency stuff. But, let's forget about that. At least for now.

Deterministic: bad stuff. Two things can make up this sort of jitter. One is any form of tone, you know........something that is not random. Like 60 Hz, getting into the signal.

OK, this may not be good, but there is one form that is definitely worse.

The other is tones that are data-correlated. IOW, its frequency content has a relationship to the signal who's jitter you are trying to measure.

Yeah, that is the bad stuff. And guess what?? That is the kind you get, when you are listening to SPDIF!

Yep, without going into a lot of theory, the process of recovering the clock from a SPDIF signal gives you lots of data-correlated jitter. Definitely not bueno.

If you have trouble believing that, here is what you can do:

(Only for crazed DIYers, who can replace any chips they may blow up, by inadvertently shorting out some pins.)

Get a battery-powered listening device, of some sort. (Yes, a headphone type would be good.) Find the place on your SPDIF RX chip, that has the PLL filter. Listen to it.

Then play some music, and listen some more. If you can. It may make you ill.

OK, for those of you who don't have those abilities, you will hear a highly-distorted version of what you are listening to.

Yep.............not bueno. But, that is what you will get.

So, is that the only way to get data-correlated jitter? No. Take for an example a line of fine equipment, made by some big company (which, I may or may not have mentioned in my previous post), their gear has a fair amount of data-correlated jitter.

Why?

Well...........let's just say sticking the clock in the same chip as the SPDIF driver may be a good way to save a fraction of a penny, but does bad stuff, for jitter.

Actually, they are not the only ones! Yeah, we have measured other gear, that has data-correlated jitter. Similar manner, just different method. (Anyone who tells you "Hey, we use this brand of SPDIF TX chip, because it has a built-in reclocker, and that negates the use of an external one"................yeah, guess what...........wrong!)

"Well, what do we do?"

If you insist on using SPDIF, know that you will always have a certain amount of deterministic jitter. Even with reclockers and ASRCs. Highly reduced, but not totally eliminated. (Some of it gets converted into other nasty stuff, but let's not go there.)

But, more importantly, don't get hung up on jitter numbers. As you can see, there are all sorts of ways to measure and characterize them. None of them are good at telling you what will come out the end of your DAC. Guys like me use them, in the design process, to help in the evaluation stage. But even we don't all agree on how to measure jitter, or which methods yield useful data. OK, maybe some of us tout our GPT is better than the next guy's GPT. Unless you have a good knowledge of how it is measured, don't get too excited about it.

Having said all of that.................

A certain audiophile mag likes to show plots, of the jitter, of various pieces of gear. I would not get too hung up on them. But............if one piece of gear you are thinking about has a really nasty looking plot, with all sorts of stuff, sticking up, all over the place..................yeah, you may want to look at buying something else. The rest of them...................yep, you may have to listen to them, to know.

OK, I'm typed out, and you are all probably as red-eyed as I am, having both written and read this. You can send hate mail, to our CSB! (She is out of the hospital, but not back to working with us. Translation: no one is there to read it.) I may or may not check back. If I don't, take it as we are busy.

Happy listening, guys................

Any light you can throw won my above questions - I would be really interested in your answer?

Ah, that's a pity I was hoping to get some answers to these questions as I have had them for a long while & ar-t (aka Pat diGiacommo aka JockoHomo) has the expertise to answer them. Maybe he will in time?  I may even have asked him some of these questions before?

Hi Pat,
thanks for the explanation! I have learned something today  :yes:
Greetings
Adrian

Hi Flecko,
I see you let Jocko know about this thread :)
Pity he didn't answer the other Qs -maybe later?
Did you try the 20dB attenuators on your Hiface?

No I didn't. :) But I was happy that he got aware of it.

Yep, would be interesting.

Not yet. I am still seraching for a distributor in germany. Minicirciuts seems not to ship to germany for such small order. I asked a lokal electronic dealer and the attenuators are about 30€. It seems that I have to buy it there.

Minicircuits have international reps in Germany http://minicircuits.com/pages/irep.html#ge - try them

Thank you! I will contact them.

Yeah, minicircuits from UK will send an attenuator of 10 euro to the netherlands with 20 euro shipping cost.
But they are very friendly and helpfull.

The american guys were friendly too, but after I told them how much of each typ I want, there was no reply. But there was a great snow storm and a lot of staff were missing. Don't know if they have only forgotten to answere.

GerardA,
Are you using one on your Hiface?

Not yet, just ordered it.
Hope it will arrive this week and doesn't dissapoint me!
BTW, the VAT/BTW of 20 % has to be added, so total of € 36.......????

Ah, rip-off UK is nearly as bad as rip-off Ireland. Nearly cheaper to buy from the US  - $12 + shipping.

Post your impressions when you get a chance, please!

I will!

Anyone interested in jitter, might want to peruse this thread, over at AC:

http://www.audiocircle.com/index.php?topic=90220 (http://www.audiocircle.com/index.php?topic=90220)

All for now..................gotta get back to work!

Well John, thanks!!!

The attenuator arrived today, really quick.
It's 15 dB. And now for the sound, it is totally not what I expected!
Although I started playing albums I did not hear for a long time and can not set myself to much AB-ing yet it
is clear from the start. The sound is so different that it's hard to describe what the real difference is.
For the moment it's more dynamic and loud, especially transient wake you up/ gets you excited/ grabs your attention.
Now I'm not listening to digital anymore but just electronics.
And the weird mix of Led Zep 2, reminds me of Led Zep 3 on vinyl.
No more of that strangeness of digital.
Well maybe more later.
A must have for every hiFace-owner!

A little more:
Playing loud all evening, enjoying so much, every aspect is better so no long story!

Maybe, later you'll see a tear from my eyes!  ;) Time to stop!

Great GerardA, thanks for reporting back. It really changes the sound, doesn't it? More analogue like, more solid 3D sound stage, more texture to the bass, easier to listen to louder, no fatigue, etc. All the sonic characteristics of lower jitter. For a small investment, it's a great return, no?

BTW, these can be use don other SPDIF cables, not just on the Hiface. 20dB might be too much or not, in this case!

All of this and more!

BTW, I was thinking of RFI induced by the high SPDIF-signal in the DAC on the analog output being reduced too, but I gues it must me mainly the jitter doing this.
Will your mods to the hiFace add much on top of this?  :yes: :no:

Yes there may well be other benefits besides the reduction of reflections - Pat(ar-t) knows all about these devices :)
About another 90%, no kidding - ask Boleary here or look at the end-user/press reviews - they all consider the attenuators a minor uograde compared to the modified Hiface

GerardA, the Jkenny MK2 will make your current Hiface settup--with attenuator-- sound like its shrouded in mist.....you gotta hear the MK2 to believe it.

I want to give an update to this topic. I changed my setup. Now my pc is closer to my DAC. I tried different spdif cables between my Audio-GD DI and the Ref7. The outcome was, the best sounding cable was a 0.8m long digital cable from Van den Hul (digicoupler). Also a 1.2m long noname china digital cable sounded better than my 5m sommer digital cable. Interesting was, that coupling the DI direct to the DAC didn't sounded es good as with the digicoupler. I tested the the digital cables with my spdif test setup too. For that, a modified Pioneer PDS-707 is used. It has an digital out and input that can be switched. If I shortcut the input and the output with a digital cable, I can switch between the cable and no cable. So, the sound of a cable can be determined very good. The digicoupler sounds almost like no cable. Very neutral. I could well imagine, that there is no significantly better spdif cable available.
I think it is very much like Pat said before. You got to try yourself how different length of digital cable sounds in your system.