Go Back   Home > Forums > Source & Line > Digital Source

Digital Source Digital Players and Recorders: CD , SACD , Tape, Memory Card, etc.

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 20th December 2012, 11:06 PM   #161
diyAudio Member
 
zenelectro's Avatar
 
Join Date: Sep 2009
Quote:
Originally Posted by couch View Post
Thanks for this, Joe. I just read the related material on your web site. Please correct me if I'm wrong, but it occurs to me that (following your "jitter is analogue" image) that jitter frequency would be (as with analogue waves) inversely proportional to 'wavelength'--in the time domain, presumably the number of clock cycles affected per 'cycle' of jitter (with jitter amplitude being the amount of phase shift per clock cycle). So very low frequency jitter (also called 'wander,' I believe) would, if this model holds, be that which shifts the signal across the largest number of clock cycles.
There are a number of ways of characterizing jitter. From lowest freq -> high

Long term drift = days / months / years
Allan deviation (or variance) = 0.1 sec to 1000 sec (or longer)
Phase noise plot = 1Hz to 1MHz from carrier.

If you are interested do some googling there are a number of pdf's and ppt
downloads to explain further. Too much to go into here.

Quote:

If this is true, then it makes perfect sense that, other things being equal, the lower the frequency of jitter, the more distortion of the signal it would cause. A 1-Hz jitter component would affect, for instance, 44.1 thousand samples of a Red Book CD signal. A two-Hz jitter component would affect 22.05 thousand, etc. The additive effect of several low-frequency jitter components (even allowing for statistical cancelling out of some the huge sum of phase shifts) could, even if jitter amplitude were held constant across the jitter frequency spectrum, result in a potential for dramatic timing errors in signal reconstruction at all audio frequencies. Whereas higher-frequency jitter, affecting comparatively few clock cycles, would be more likely to inflect harmonics/timbre higher in the audio band, and their additive effects would be smaller.

It would be interesting, if adequately sensitive test and clock-perturbing equipment exists, to do listening tests on the effects of injecting different very narrow frequency bands of jitter on the same audio signal (clocked by something like the Rubidium clock), followed by the testing the effects of "stacking" several of those narrow bands of jitter, particularly at relatively low frequencies.

In any case, such a model would seem to confirm both the efficacy of ultra-low noise power supply design and construction (to power the digital circuitry) using known techniques, as well as of the discovery and implementation of clocks with optimum inherent stability (thus synchronous oscillators--rather than PLL--based on SAW devices would be candidates?). Given successful execution in these two areas, then one could look at minimizing sources of random rather than periodic jitter, such as thermal noise, etc.

I'd welcome comments on all the above--I am no specialist in this field and I could be way off base, or this could all be old hat to those in the know, but if it is, I don't think its impact on audiophile gear design has really hit home yet.

Thanks--this my first post, so please set any flamethrowers to "low."
There's too much to go into here. Rubidium are not necessarily good for audio
they are for long term time keeping. Short term phase noise can be poor.
Do some searching on phase noise of rubidium clocks to get up to speed
on their performance. There are ones that do have extremely good
performance at audio freq but these are VHQ OCXO's phase locked to the
Rub std.

WRT SAW oscillators / VLN power supplies:

There are efforts here (Joe) of going to large extremes to try and minimize
ultra LF noise in clock power supply. Even though PS may take 30 min to
ramp up, this doesn't mean it will have super low LF noise. The ramp up time
can be simply sped up with a few diodes. The LF noise generally comes
down to semiconductor 1/f noise. I don't see any projected specs or
measurements.

SAW oscillators generally have worse low frequency phase noise than
crystal (BAW) oscillators. Below 10Hz offset I don't know. I'd like to see
an Allan Dev. plot for SAW oscillator.

There are other factors like G sensitivity. Here I think SAW might be very
good.

So far I don't see much correlation between oscillator jitter and sound
quality. In fact it almost appears opposite.

They may well sound good but objectively speaking we are flying in the
dark here.
  Reply With Quote
Old 21st December 2012, 08:49 AM   #162
diyAudio Member
 
Joe Rasmussen's Avatar
 
Join Date: Jan 2003
Location: Sydney, Australia
Send a message via MSN to Joe Rasmussen
Just received '105 today. Only taken the top cover off.

I imagine, for ease, that I will be able to dispense of the 7.1 board while I actually work on the player. It should run in stereo mode without that 7.1 board - when finished, put it back.

This should make it a lot easier to work on than '95.

Even though much taller, internally this box is still fairly brim full. For the money paid, this is a lot of player.

Cheers, Joe R.
__________________
The "Elsinore Project" DIY Speaker System
Custom Analogue Audio - we also support and promote non-profit DIY
  Reply With Quote
Old 21st December 2012, 10:00 AM   #163
diyAudio Member
 
Joe Rasmussen's Avatar
 
Join Date: Jan 2003
Location: Sydney, Australia
Send a message via MSN to Joe Rasmussen
Quote:
Originally Posted by Coris View Post
The solution is to lift up all the output pins of the DAC and do another connections.
Hi Coris

I would definitely not attempt that, and won't have to.

Remove the I/V chips, then use/join pads for Pin 2 and Pin 6 on all of them, and they will be paralleled. Then use your own designed I/V stage etc.

We have worked on a number of Yamaha CD-S1000 and CD-S2000 CD/SACD Players, and there you had to join up pads as well because the Burr-Brown current DACs were is monaural mode (one for Left and the other for right) where each phase had its own I/V IC. Worked just fine.

Cheers, Joe R.
__________________
The "Elsinore Project" DIY Speaker System
Custom Analogue Audio - we also support and promote non-profit DIY
  Reply With Quote
Old 21st December 2012, 10:09 AM   #164
diyAudio Member
 
Joe Rasmussen's Avatar
 
Join Date: Jan 2003
Location: Sydney, Australia
Send a message via MSN to Joe Rasmussen
Quote:
Originally Posted by Ric Schultz View Post
However, it is better to get rid of the summing opamp altogether and use a better IV converter then straight out through a coupling cap...way better. When you sum you are just adding the sound of another opamp and its passive components used (resistors, caps, solder joints). Less is better!
Hey Ric, that is right, and I did this first some years ago and used a coupling cap as the DC was not cancelled out. So what if I put a coupling cap in the signal path, on the output of the I/V (single-ended), it still sounded much better.

I recall the first one was the XA-9000ES Sony, when Sony went over to Burr-Brown DACs. Just use a decent cap.

Cheers, Joe R.
__________________
The "Elsinore Project" DIY Speaker System
Custom Analogue Audio - we also support and promote non-profit DIY
  Reply With Quote
Old 21st December 2012, 10:14 AM   #165
mvc is offline mvc  United Kingdom
diyAudio Member
 
Join Date: Jun 2004
Location: Cambridge
Default Volume control

I now have a bdp-105 at home, and I really like it. I use it to drive a pair of active monitors directly, which sounds great. The only thing I worry about is the fact that I find myself listening at volume levels mostly below 50, sometimes below 30. I have tried to find out more, but failed. Does anyone know how many dB per volume step the Sabre DAC attenuates?

And, more complicated, assuming that I have a good preamp (say noise and distortion are below -100dB), at what attenuation level will the digital volume control be worse than my
analoge preamp? ESS claim about 130dB s/n for the Sabre, so I would assume attenuating more than 30dB could make it worse than my preamp, all assuming that my preamp can retain 100dB s/n while attenuating more than 30dB, of course.

Obviously, I have tried both, and I think I prefer the direct connection, but it is still early days.

BTW, My preamp is a Django transformer preamp, using the S&B TX-102 transformers. (yes, 100dB is just a guess)
  Reply With Quote
Old 21st December 2012, 10:20 AM   #166
diyAudio Member
 
Joe Rasmussen's Avatar
 
Join Date: Jan 2003
Location: Sydney, Australia
Send a message via MSN to Joe Rasmussen
Quote:
Originally Posted by couch View Post
Thanks--this my first post, so please set any flamethrowers to "low."
If that was your first post, then that may well qualify as the BEST first post ever.

Bravo!

Your explanation of "wander" seems spot on. In fact I wish I had come up with that, and indeed I shall use "wander" as a simplification. May I?

We know that somebody on the West Coast in the US is also looking closely at this - so some serious research is definitely happening. And in time I expect more will come to light and a wider general understanding. You certainly seemed to have grasped some of the complexities very well.

Cheers, Joe R.
__________________
The "Elsinore Project" DIY Speaker System
Custom Analogue Audio - we also support and promote non-profit DIY
  Reply With Quote
Old 21st December 2012, 10:45 AM   #167
Coris is offline Coris  Norway
diyAudio Member
 
Join Date: May 2009
Location: Norway
Quote:
Originally Posted by Joe Rasmussen View Post
Just received '105 today. Only taken the top cover off.

I imagine, for ease, that I will be able to dispense of the 7.1 board while I actually work on the player. It should run in stereo mode without that 7.1 board - when finished, put it back.

This should make it a lot easier to work on than '95.

Even though much taller, internally this box is still fairly brim full. For the money paid, this is a lot of player.

Cheers, Joe R.

Cheers for your new Oppo player!

It is interesting to see that it will work without the multichannel board on its place (connected I mean...). The 95 model did not worked (booted up) without multichannel DAC, or without that chip clocked...
As I could see, the firmware chips for both DACs are planted on the stereo board...
Anyway, just try without that board. I`m a little bit sceptic that the this model too will boot up without multichannel board.

Thanks for infos above....

Last edited by Coris; 21st December 2012 at 10:49 AM.
  Reply With Quote
Old 21st December 2012, 11:24 AM   #168
diyAudio Member
 
Joe Rasmussen's Avatar
 
Join Date: Jan 2003
Location: Sydney, Australia
Send a message via MSN to Joe Rasmussen
Quote:
Originally Posted by Coris View Post
Anyway, just try without that board. I`m a little bit sceptic that the this model too will boot up without multichannel board.
Not my '105 - but the first one here and over the next few weeks, as times allow, will certainly be looking at things.

To try without 7.1 Board, all I need to do is not to remove it, but simply disconnect the analog power supply harness/connector and also pull out the ribbon. I don't need to physically remove it. But right now I haven't even turned the player on yet.

Cheers, Joe R.
__________________
The "Elsinore Project" DIY Speaker System
Custom Analogue Audio - we also support and promote non-profit DIY
  Reply With Quote
Old 22nd December 2012, 12:33 AM   #169
diyAudio Member
 
zenelectro's Avatar
 
Join Date: Sep 2009
Quote:
Originally Posted by Joe Rasmussen View Post
If that was your first post, then that may well qualify as the BEST first post ever.

Bravo!

Your explanation of "wander" seems spot on. In fact I wish I had come up with that, and indeed I shall use "wander" as a simplification. May I?

We know that somebody on the West Coast in the US is also looking closely at this - so some serious research is definitely happening. And in time I expect more will come to light and a wider general understanding. You certainly seemed to have grasped some of the complexities very well.

Cheers, Joe R.
Guys / Joe

I'm guessing 'wander' will come some where in the Allan Variance (Deviation)
and phase noise very close to carrier, say 1Hz offset. See attachment
below.

Joe, I'm guessing you are alluding to the fact that stability in the range
>10 sec is important? All your work seems to point in this direction. In this
case look at getting Allan Variance measurements for the oscillators and
look at phase noise 1Hz and below.

I would be surprised if stability in > 1000sec range is of critical importance,
interestingly, in the Rakon paper they suggest this range is not well
characterized.

At the end of the day, again, I still come back to the basic premise that we
have conflicting results here.

There is a huge effort placed on reducing LF noise in power supplies but all
the info I can turn up suggests SAW oscillators have worse low freq phase
noise and med / long term stability than BAW. Ref to attachment below.

I'm interested in any results or conclusions your people on west coast come
up with. However if they are to conclude anything meaningful, they will
have to get access to the specialised equipment required to measure these
areas of stability.
Attached Images
File Type: jpg SAW vs BAW.jpg (255.9 KB, 331 views)
File Type: jpg osc stability.jpg (171.2 KB, 314 views)
  Reply With Quote
Old 22nd December 2012, 04:35 AM   #170
diyAudio Member
 
Joe Rasmussen's Avatar
 
Join Date: Jan 2003
Location: Sydney, Australia
Send a message via MSN to Joe Rasmussen
Quote:
Originally Posted by zenelectro View Post
Joe, I'm guessing you are alluding to the fact that stability in the range
>10 sec is important? All your work seems to point in this direction. In this
case look at getting Allan Variance measurements for the oscillators and
look at phase noise 1Hz and below.

.... I'm interested in any results or conclusions your people on west coast come
up with. However if they are to conclude anything meaningful, they will
have to get access to the specialised equipment required to measure these
areas of stability.
Hi Terry

I don't have any conclusive sub-1Hz data, other than Paul Miller's example. I am not saying it is conclusive, but it does at least point to a kind of mechanism at work:

Click the image to open in full size.

Using rather expensive equipment (and you yourself indicate some exotic equipment may be needed, read expensive) he used 100X magnification of the Blue Box area:

Click the image to open in full size.

Same measurement, but the 'better' sounding player is Black and the less so Red. I can see Red cling right up to -70dB - and that's +60dB and maybe who knows how much worse? If the resolution was better, it could look truly frightening.

Even Miller himself did seem to know what he was seeing, and exactly where the noise comes from? It is power supply or...? Admittedly this is not an oscillator measurement, so I don't want to over-interpret it. But it is the first time somebody came across it and put a description on it: "Low rate, uncorrelated or noise-like jitter"

Also, that measurement starts at a much higher frequency, but that is neither here or there, it is that climb and would not also the windowing affect the interpretation of the frequency near the carrier? But it's the amplitude that is disturbing, as it climbs the carrier.

I am just struck by the fact that here is something not many people has commented on, and by some a lot smarter than me can figure out if this is a trend or characteristic that is repeated in a number of, for want of a better word, situations, oscillators included?

Quote:
However if they are to conclude anything meaningful, they will
have to get access to the specialised equipment required to measure these
areas of stability.
Yes, since if you extrapolate what your Rakon measurement down down to near 1 Hertz would indicate nothing, just mostly a straight line upwards and is there is potentially another trend below that at much lower frequency?

That could be Allan Variance/stability?

I suspect you have your doubts about that. I can understand that. But I am not so sure. To be honest, I don't know.

In my mind, I suppose that the window sub-1Hertz is then still, to a degree, an open book. That Miller measurement... it just gives me more and more food for thought?

ATTENTION Coris:

Of the 100MHz Oscillators you tried on the Sabre DAC, was one of them a Crystek 957 as used by the Buffalo guys?

What we desperately need to know whether anyone has compared it to a 100MHz SAW.

I have a friend who watches this thread, who has offered to send a spare 957 Oscillator to me, he likewise is going to run the comparison the opposite way, going from Crystek to SAW, and I will go from SAW to Crystek.

Will I be disappointed if the Crystek "BAW" sounds better than the "SAW"?

Most definitely not!

I will then definitely want to use it !!!

But IF the SAW sounds better, then in the light of above, there are questions to be asked and answers to be got.

Cheers, Joe R.
__________________
The "Elsinore Project" DIY Speaker System
Custom Analogue Audio - we also support and promote non-profit DIY

Last edited by Joe Rasmussen; 22nd December 2012 at 04:44 AM.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
OPPO 93 and OPPO 95 - A warning. sandyK Digital Source 3 26th January 2012 06:53 AM
Discussions about your favorite or DIYD/T-amp v-bro Class D 0 13th January 2009 08:37 AM
Class D discussions mskeete Class D 16 19th March 2004 06:27 AM
Discussions on Balenced Amplifers theChris Chip Amps 24 6th January 2004 02:40 PM
Do opamps discussions also have to take place here? Bricolo Chip Amps 3 27th April 2003 08:39 PM


New To Site? Need Help?

All times are GMT. The time now is 08:46 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright ©1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2