Re: A bit of off-topic joviality..
Standards conversion has come a long way...
From full boards cramped of material to a few chips, and quality has improved a lot, too.
NTSC converted to PAL had a lowsy quality.
I did, but didn't know you were involved.
That's coc..
dhaen said:It is too
Standards conversion has come a long way...
From full boards cramped of material to a few chips, and quality has improved a lot, too.
NTSC converted to PAL had a lowsy quality.
dhaen said:
I did, but didn't know you were involved.
That's coc..
Guido,
forgive my ignorance, but I'm not sure what your graph is telling me. It appears to show rather a lot of noise within 20kHz of the fundamental, so doesn't this get converted into distortion by the DAC? Or does the division by 256 from 11.2896MHz down to 44.1kHz mean that your clock's noise is also divided in frequency? And that sort of musing makes me wonder whether that's the reason why multi-bit convertors without any over-sampling are becoming fashiionable again - they reduce the effect of clock jitter?
Perhaps you could show the spectrum of the typical cheap and cheerful clock in a CD player to allow comparison?
forgive my ignorance, but I'm not sure what your graph is telling me. It appears to show rather a lot of noise within 20kHz of the fundamental, so doesn't this get converted into distortion by the DAC? Or does the division by 256 from 11.2896MHz down to 44.1kHz mean that your clock's noise is also divided in frequency? And that sort of musing makes me wonder whether that's the reason why multi-bit convertors without any over-sampling are becoming fashiionable again - they reduce the effect of clock jitter?
Perhaps you could show the spectrum of the typical cheap and cheerful clock in a CD player to allow comparison?
EC8010 said:
Hi,
The units are shown in the graph, so it isn't "that" difficult.
It shows the density of the jitter spectrum, from 0Hz to 115Hz. On the vertical axis 0dB equals 74 ps.
NOS DACs are favourable because they are clocked at lower rates which makes the on-chip mess a little less, the absence of analogue filtering, and the abscense of the jitter factory, also known as digital filter.
regards
EC8010 said:
Hi
I had to rescale the picture due to size limits. Jitter cannot be removed, it can only be attenuated (even fifo's have crosstalk, a good example is the Chord DAC).
By the way, we promoted reclocking in DACs over 12 years ago
http://www.tentlabs.com/Products/DIYDAC/DIYDAC.html
cheers
You have a picture of yourselfLars Clausen said:
Never seen one before.
Jitter Factories?
Hi all,
Is it true that (oversampling) digital filters generate a lot of jitter, something like jitter factories?
Hardware-wise, I guess digital filters are mere synchronous logic units designed for lots of arithmetic operations and data buffering and shifting. Then, it would be hard to generate excessively large amount of jitter compared to other synchronous logic circuits, because all outgoing data would be buffered into shift registers and synchronously clocked out by the system clock.
Of course, a digital filter would have tons of logic gates inside. So if all of them are turning on and off simultaneously, they will obviously pollute power rails and ground a lot. And this rail noise will make shift registers to jitter quite a little when clocking data out. But I am not sure whether or not this jitter amount is really substantially bigger than, say, that of simple flip-flop 7474.
OK, it's all my mere speculation. If someone has a hard proof like a measurement showing a big jitter, that's the truth.
Anyway, if digital filters are really jitter factories, it wouldn't be a nice idea to reclock a DAC that contains a digital filter??
- SKK
Hi all,
Is it true that (oversampling) digital filters generate a lot of jitter, something like jitter factories?
Hardware-wise, I guess digital filters are mere synchronous logic units designed for lots of arithmetic operations and data buffering and shifting. Then, it would be hard to generate excessively large amount of jitter compared to other synchronous logic circuits, because all outgoing data would be buffered into shift registers and synchronously clocked out by the system clock.
Of course, a digital filter would have tons of logic gates inside. So if all of them are turning on and off simultaneously, they will obviously pollute power rails and ground a lot. And this rail noise will make shift registers to jitter quite a little when clocking data out. But I am not sure whether or not this jitter amount is really substantially bigger than, say, that of simple flip-flop 7474.
OK, it's all my mere speculation. If someone has a hard proof like a measurement showing a big jitter, that's the truth.
Anyway, if digital filters are really jitter factories, it wouldn't be a nice idea to reclock a DAC that contains a digital filter??
- SKK
KBK said:
Since currrent state of the art crystal oscillators are less than 1 or 2 dB worse than theory predicts how much better can it be?
If you want a reference oscillator I wouldn't suggest an aftermarket audio oscillator unless it has published performance data. The most meaningful would be a phase noise versus frequency plot in dBc/Hz.
Reference oscillators are available for not much more money that the most expensive audio aftermarket types from companies such as Wenzel.
www.wenzel.com
try these low noise types for example.
http://www.wenzel.com/catalog.html#HF Oscillators
I have had some personal contact with the people at Wenzel and I can say that they do what they do because they like it, and take pride in a VERY low noise (esp. phase noise (-180dB)) product. They are a family-run designed/made in US company.
If you want to pay for it, they will build you exactly the clock you need.
(I tried to get a job with them once but interviewed on 2 hrs sleep with a 2 month old baby in the house...
)
If you want to pay for it, they will build you exactly the clock you need.
(I tried to get a job with them once but interviewed on 2 hrs sleep with a 2 month old baby in the house...
)
CraigBuckingham said:
Hi Craig,
I tried to order some oscillators years ago. That company wanted two bank warranties and the price was high as CD frequencies are not standard. The oscillator had to be made to special order.
So I gave up after reviewing many price quotes of the oscillator mongers and builded my own which is more fun too. My schematic has been downloaded more than 3600 times here:
http://www.diyaudio.com/forums/showthread.php?postid=199928#post199928
Wenzel oscillator
Hello Elso
All the oscillators wenzel make are custom they have no off the shelf product.
A custom made clock made by these guys will cost you in 1 offs min$300 US , which is better than any clock anybody in these forums can make by a long way. They also have on there website technical stuff to help you build a clock which is equal to the best off the shelf clocks you can buy with phase noise of -170dbc and 10kHz offsets. They hold the record for making the lowest phase noise clocks at any frequency offsets fullstop. They are the best.
Here is the website info.
http://www.wenzel.com/documents/circuits.html
Regards
Arthur Rappos
Regards
Arthur
Elso Kwak said:
Hi Craig,
I tried to order some oscillators years ago. That company wanted two bank warranties and the price was high as CD frequencies are not standard. The oscillator had to be made to special order.
So I gave up after reviewing many price quotes of the oscillator mongers and builded my own which is more fun too. My schematic has been downloaded more than 3600 times here:
http://www.diyaudio.com/forums/showthread.php?postid=199928#post199928
Hello Elso
All the oscillators wenzel make are custom they have no off the shelf product.
A custom made clock made by these guys will cost you in 1 offs min$300 US , which is better than any clock anybody in these forums can make by a long way. They also have on there website technical stuff to help you build a clock which is equal to the best off the shelf clocks you can buy with phase noise of -170dbc and 10kHz offsets. They hold the record for making the lowest phase noise clocks at any frequency offsets fullstop. They are the best.
Here is the website info.
http://www.wenzel.com/documents/circuits.html
Regards
Arthur Rappos
Regards
Arthur
Arthur, I would just like to clarify one point you made.
Wenzel do not have circuits on their website which will enable someone to realise a low phase noise design like their own.
To echo what you said about being the best period, I remember reading an article sometime ago regarding his recording breaking oscillators. There were a couple of other, less well known to me, designer's figures on the comparsion graph. They were only a couple to a few dB off Wenzel's best figures.
Wish I could remember where that article was.
Regards, Craig.
Wenzel do not have circuits on their website which will enable someone to realise a low phase noise design like their own.
To echo what you said about being the best period, I remember reading an article sometime ago regarding his recording breaking oscillators. There were a couple of other, less well known to me, designer's figures on the comparsion graph. They were only a couple to a few dB off Wenzel's best figures.
Wish I could remember where that article was.
Regards, Craig.
Ok Arthur,
My misunderstanding of what you were stating. You did say "off the shelf" and I included Wenzel in that statement. I had a feeling that Wenzel do "standard" frequency oscillators "off the shelf".
However, I understand now that when you said that you had it in mind that Wenzel do not supply "off the shelf" oscillators, only custom. All of which I have not confirmed with Wenzel.
However, I would like to clarify your previous post where you say "its phase noise specs are limited only by the crystal you use with it". In this I understand you are referring to the oscillator circuit on Wenzels website at http://www.wenzel.com/pdffiles/xtalosc.pdf
Whilst this circuit may be good it's phase noise is not limited by the crystal that is used with it.
Regards, Craig.
My misunderstanding of what you were stating. You did say "off the shelf" and I included Wenzel in that statement. I had a feeling that Wenzel do "standard" frequency oscillators "off the shelf".
However, I understand now that when you said that you had it in mind that Wenzel do not supply "off the shelf" oscillators, only custom. All of which I have not confirmed with Wenzel.
However, I would like to clarify your previous post where you say "its phase noise specs are limited only by the crystal you use with it". In this I understand you are referring to the oscillator circuit on Wenzels website at http://www.wenzel.com/pdffiles/xtalosc.pdf
Whilst this circuit may be good it's phase noise is not limited by the crystal that is used with it.
Regards, Craig.
CraigBuckingham said:
Hello Craig
You are right about the circuit I am talking about .
The last statement you make should read more precisely a t the close in offsets below 200 Hz lets say the limiting factor of performace is the crystal and not this circuit.
Wenzel are a custom oscillator supplier , no off the shelf product when it comes to clocks, its made to order.
Regards
Arthur
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.