Hi,
if you are interested I can still add one or more to the order.
Please, let me know.
Andrea
...oooops... but not SC-cut, AT-cut only.
GB update
The order was placed and the payment was done.
I expect about 8-9 weeks for delivery from Laptech. As soon as I will receive the crystals I ship them to all the members.
If someone was interested, I still have the following crystals in excess:
AT-Cut 5.6448 MHz fundamental 6 pcs
AT-Cut 6.1440 MHz fundamental 1 pcs
AT-Cut 11.2896 MHz fundamental 1 pcs
AT-Cut 12.2880 MHz fundamental 4 pcs
AT-Cut 16.9344 MHz fundamental 1 pcs
AT-Cut 22.5792 MHz fundamental 2 pcs
AT-Cut 24.5760 MHz fundamental 1 pcs
AT-Cut 25.0000 MHz fundamental 3 pcs
AT-Cut 33.8688 MHz 3rd overtone 3 pcs
AT-Cut 45.1584 MHz 3rd overtone 2 pcs
AT-Cut 49.1520 MHz 3rd overtone 2 pcs
The order was placed and the payment was done.
I expect about 8-9 weeks for delivery from Laptech. As soon as I will receive the crystals I ship them to all the members.
If someone was interested, I still have the following crystals in excess:
AT-Cut 5.6448 MHz fundamental 6 pcs
AT-Cut 6.1440 MHz fundamental 1 pcs
AT-Cut 11.2896 MHz fundamental 1 pcs
AT-Cut 12.2880 MHz fundamental 4 pcs
AT-Cut 16.9344 MHz fundamental 1 pcs
AT-Cut 22.5792 MHz fundamental 2 pcs
AT-Cut 24.5760 MHz fundamental 1 pcs
AT-Cut 25.0000 MHz fundamental 3 pcs
AT-Cut 33.8688 MHz 3rd overtone 3 pcs
AT-Cut 45.1584 MHz 3rd overtone 2 pcs
AT-Cut 49.1520 MHz 3rd overtone 2 pcs
....
Wow - so now the crystal are underways ... splendid, Andrea!
Thanks once more for the work and effort you put into this. To me it's actually quite special to have the opportunity to get what I hope will be quite unusual oscillators - and eventually be able to use and hear what this means to the sound.
Cheers to you
Jesper
....
Wow - so now the crystal are underways ... splendid, Andrea!
Thanks once more for the work and effort you put into this. To me it's actually quite special to have the opportunity to get what I hope will be quite unusual oscillators - and eventually be able to use and hear what this means to the sound.
Cheers to you
Jesper
24 MHz Crystal
Before you get too far please be sure of what you hope to accomplish upgrading crystals on motherboards and USB interfaces. In both cases jitter or phase noise is not a big issue. Absolute frequency accuracy is not really important either, the tolerances are broad enough that inexpensive crystals work fine. Some applications will use a packaged oscillator, others will use a crystal tied to a gate oscillator. In the latter case swapping a crystal without making the necessary circuit adjustments will certainly not work.
Don't expect either to have any sonic benefit, or any significant or detectable impact on performance. All the clocks connected to them will be completely asynchronous to any audio clock. The one case where there was an issue I encountered was a Via single board computer. However the audio clock/sample rate was 3% low so they could derive it from an internal clock. It did not help anything and no dac I had could lock to it.
Also for both this and for any audio application using a high speed driver (74AC or potato) you can get lower phase noise at the expense of turning your project into a powerful transmitter with a clean reference (NIST already does that with the WWV transmitters), not what you want in the presence of sensitive analog circuits. The cleanest clock drive is a clipped sine wave but that's not exactly want the CMOS circuit want to see.
This stuff is never easy.
Before you get too far please be sure of what you hope to accomplish upgrading crystals on motherboards and USB interfaces. In both cases jitter or phase noise is not a big issue. Absolute frequency accuracy is not really important either, the tolerances are broad enough that inexpensive crystals work fine. Some applications will use a packaged oscillator, others will use a crystal tied to a gate oscillator. In the latter case swapping a crystal without making the necessary circuit adjustments will certainly not work.
Don't expect either to have any sonic benefit, or any significant or detectable impact on performance. All the clocks connected to them will be completely asynchronous to any audio clock. The one case where there was an issue I encountered was a Via single board computer. However the audio clock/sample rate was 3% low so they could derive it from an internal clock. It did not help anything and no dac I had could lock to it.
Also for both this and for any audio application using a high speed driver (74AC or potato) you can get lower phase noise at the expense of turning your project into a powerful transmitter with a clean reference (NIST already does that with the WWV transmitters), not what you want in the presence of sensitive analog circuits. The cleanest clock drive is a clipped sine wave but that's not exactly want the CMOS circuit want to see.
This stuff is never easy.
24 MHz Crystal
Thanks for the advices Demian. I'm totally not expert on this matter, but it's perfectly clear to me that these clocks are all asynchronous respect audio signal.
But I don't know if that's enough not to get any improvement. Also I adopt Ian's FIFO system that only uses data and replace all clocks. But, again... as above...
I became interested in these modifications as other people reported much positive results. Here's thread on the - great - Phasure audio player forum:
Mobo main clock replacement...
At the end effort and cost is not that high - unless one would have to trash his motherboard at the end
Also there are 24.000MHz OCXOs, obviously expensive, commercially available for these purposes.
Thanks for the advices Demian. I'm totally not expert on this matter, but it's perfectly clear to me that these clocks are all asynchronous respect audio signal.
But I don't know if that's enough not to get any improvement. Also I adopt Ian's FIFO system that only uses data and replace all clocks. But, again... as above...
I became interested in these modifications as other people reported much positive results. Here's thread on the - great - Phasure audio player forum:
Mobo main clock replacement...
At the end effort and cost is not that high - unless one would have to trash his motherboard at the end
Also there are 24.000MHz OCXOs, obviously expensive, commercially available for these purposes.
Last edited:
There are so many questions to be resolved before any thought of the clock jitter being an aspect.
1) the mobo has a clock modulator to get through FCC testing. It doesn't actually reduce the RFI just spreads it out so the receiver doesn't see it as much. Essentially a jitter modulator/creator. Its everything you don't want for audio.
2) increased EMI from faster edges is often responsible for a "more incisive high end" which is really an artifact.
If you start with a Soekris or Alix board and follow the efforts to make it a low jitter NTP clock source https://www.febo.com/pages/soekris/ you may get positive results. i don't think a conventional motherboard would ever get there. And this would not be a quick project. I do have a couple of the Soekris net4501's i have retired. However for audio its not so simple.
A properly controlled experiment would be really difficult. I think Peter's last remark in the thread mirrors my feelings, neutral but skeptical.
Personally I suggest what I did in the Auraliti L1000- make the whole player in an EMI tight enclosure with the various subsystems well isolated from each other and the EMI well contained so it won't get out.
1) the mobo has a clock modulator to get through FCC testing. It doesn't actually reduce the RFI just spreads it out so the receiver doesn't see it as much. Essentially a jitter modulator/creator. Its everything you don't want for audio.
2) increased EMI from faster edges is often responsible for a "more incisive high end" which is really an artifact.
If you start with a Soekris or Alix board and follow the efforts to make it a low jitter NTP clock source https://www.febo.com/pages/soekris/ you may get positive results. i don't think a conventional motherboard would ever get there. And this would not be a quick project. I do have a couple of the Soekris net4501's i have retired. However for audio its not so simple.
A properly controlled experiment would be really difficult. I think Peter's last remark in the thread mirrors my feelings, neutral but skeptical.
Personally I suggest what I did in the Auraliti L1000- make the whole player in an EMI tight enclosure with the various subsystems well isolated from each other and the EMI well contained so it won't get out.
That Dutch guy is me! Herbert. Just read the articles on my website to understand the details of the oscillator Andrea still does not understand!
Read: https://www.by-rutgers.nl/PDFiles/Reproducible Low Noise Oscillators.pdf
Herbert.
Read: https://www.by-rutgers.nl/PDFiles/Reproducible Low Noise Oscillators.pdf
Herbert.
The best clock oscillator
The Dutch guy andrea_mori is talking about is called Herbert and you could find all the details of the oscillator in question on my website at:
https://www.by-rutgers.nl/PDFiles/Reproducible Low Noise Oscillators.pdf
The Dutch guy andrea_mori is talking about is called Herbert and you could find all the details of the oscillator in question on my website at:
https://www.by-rutgers.nl/PDFiles/Reproducible Low Noise Oscillators.pdf
Using SC and not caring about the exact frequency has some aspects of hybris.
But far out noise is only a circuit problem and close-in noise mostly a matter of Q
and cleanliness.
Below 1Hz offset we are approaching the drift area, and here things really get bad.
Note that the temp. of the oven must match the crystals sweet spot to 0.1K. That makes
it hard to repair a commercial OCXO (replacing thermistors) unless there is a sticker
inside that tells you the temp. (that happens!)
Other than a little bit more allowable power, SC has nothing more to offer than AT if you
don't care about the frequency. If you don't care about drift, even the AT dissipation
limit may be relaxed slightly.
regards, Gerhard
I absolutely follow the close in phase noise and temp etc. Issue. The one question I do have comes from notes elsewhere that an SC crystal will have higher Q. Is it so? If so is it the cut or the fabrication? Finally how does the Q affect the phase noise / jitter in the 10 Hz to 200 KHZ band? Above that band won't affect audio dac/adc and below that is more influenced by external stuff and probably a lot less audible.
Sent from my SGH-M919 using Tapatalk
Sent from my SGH-M919 using Tapatalk
Hi Herbert,
welcome to the "dutch guy" who designed the Clapp oscillator.
@1audio,
I think the higher Q of the SC-Cut comes from the cut rather than the fabrication. If you compare the Laptech 11.2896 MHz, the AT has a Q around 150K, while the SC has a Q of 850K or more.
@Gerhard,
Does the Q affect the close to the carrier phase noise? The Driscoll oscillator allows very high loaded Q, so the higher the Q of the Crystal, the higher the loaded Q of the circuit. Does this matter? Does the close in phase noise depend mainly from the Q of the Crystal (or of the circuit)?
welcome to the "dutch guy" who designed the Clapp oscillator.
@1audio,
I think the higher Q of the SC-Cut comes from the cut rather than the fabrication. If you compare the Laptech 11.2896 MHz, the AT has a Q around 150K, while the SC has a Q of 850K or more.
@Gerhard,
Does the Q affect the close to the carrier phase noise? The Driscoll oscillator allows very high loaded Q, so the higher the Q of the Crystal, the higher the loaded Q of the circuit. Does this matter? Does the close in phase noise depend mainly from the Q of the Crystal (or of the circuit)?
Clapp and AT/SC
Hi Gentlemen,
in my opinion the AT-cut is in a Clapp oscillator a better choice than the SC-cut because you could put much more power into an AT-cut Xtal. The Rm of an SC-cut could be 10x higher then with the AT-cut so the circle flow (Xtal, C', C'' in my Clapp) could be 10x higher with AT which means 20 dB on the S/N ratio.
Do NOT use circuits in which active components are part of the circle flow as in Driscoll !!
Mind that in my Clapp oscillator the diode BAT81 or BAT83 is crucial for the performance!!!
Herbert.
Hi Gentlemen,
in my opinion the AT-cut is in a Clapp oscillator a better choice than the SC-cut because you could put much more power into an AT-cut Xtal. The Rm of an SC-cut could be 10x higher then with the AT-cut so the circle flow (Xtal, C', C'' in my Clapp) could be 10x higher with AT which means 20 dB on the S/N ratio.
Do NOT use circuits in which active components are part of the circle flow as in Driscoll !!
Mind that in my Clapp oscillator the diode BAT81 or BAT83 is crucial for the performance!!!
Herbert.
- Status
- Not open for further replies.