In Reply to: " frankly I disagree on several points" posted by Allen Wright on October 31, 2005 at 04:04:20:
Hi Allen
Thanks for getting back.>No arguments with what you say at all - the clock waveform etc is very important - but while the TENT oscillator module we use in the RCM may look like a regular sealed package TCXO/VCXO - internally it is not! It is the subject of a huge amount of research by Guido Tent and what's inside is his own design made for him by one of the commercial oscillator manufactures so the price is acceptable.
The Tent X0 (50ppm?) does indeed looks like regular half-dip, DIL-8, 11x11x8mm package, Whilst it is not my place to debate the levels of research that have gone into this package by Guido, if you research the internal construction of this type of half-dip DIL-8 TCXO whether commercial or custom made the level of phase noise (or jitter) is limited by the design and package.
With any commercial TCXO (or VCXO); square wave, TTL/CMOS output it is driven by an “inverter†IC with compensation capacitors. For a high spec TCXO, the configuration and the IC used dictates phase noise; the IC determines the phase noise and jitter.
The DIL-8, Half dip Package uses a “naked crystalâ€, this package generally has low Q = higher jitter.
On the basis that the XO you are using is a high spec type the inverter IC would have been selected for lower jitter, the crystal selected for and tweaked to achieve tight tolerance.
If you analyse the circuit of a TCXO (or VCXO) you have crystal, inverter IC and compensation capacitors. This is the same circuit type arrangement that is generally fitted as standard to CD/SACD players. You can improve the crystal and the IC but you still have a crystal connected to the input an output of an inverter dictating phase noise and jitter. Improving power supplies to the TCXO will help but you need to design a low phase noise oscillator, not “integrated in an ICâ€, where you can control various parameters if you want lowest jitter and optimum sound quality.
>>Take a look at LC Audio; Lars Clausen realised very early in his designs that you start with a pure sign wave oscillator to achieve lowest phase noise, and then convert it to square. With this method you can control each element; oscillator, power supplies, sine-square
>Sorry - this last sentence is not clearly understandable, by me at least.
To overcome the limitations of the basic or higher spec TXCO (VCXO) crystal + inverter arrangement you need to design an oscillator from the ground up. Clock designers such as the LC Audio and Audiocom realise this and use pure sine wave oscillators which have very low phase noise. By using a pure sine wave oscillator you can specify the type of oscillator, the parts type and quality, i.e. low 1/f noise transistors (essential for low phase noise), the crystal cut and quality, even down to the resistors and capacitors.
These methods do cost more, high quality parts are used and more of them, but the end result is a superior clock.> It needs to be more than *a few inches* to get such open unshielded layouts away from the digital noise in the players I work with. Ten inches/25cm is about the minimum and I believe this will cause any clock a problem. In our listening tests, going from the original 5inches we used in Level 2 & 4 to less than one inch made a HUGE difference - which is why we focussed on being able to do this correctly in Level 5. And we had both AudioCom & LC clocks in house for comparison.
Can you explain to me where you found the source for all this “digital noise†in the SCD-1?
How are these clocks “picking up†all of this digital noise?
Thanks,
lovekylie
This post is made possible by the generous support of people like you and our sponsors:
Follow Ups
- Re: " frankly I disagree on several points" - lovekylie 11:03:20 10/31/05 (8)
- "Can you explain to me... - Allen Wright 10:22:21 11/02/05 (7)
- LK, no response to this data? (nt) - Allen Wright 02:13:41 11/04/05 (6)
- I find it hard to envisage... - lovekylie 10:09:42 11/04/05 (4)
- In each case a re-installation should effect a cure" - Allen Wright 13:05:55 11/04/05 (3)
- Perhaps you are missing the point here... - lovekylie 14:16:02 11/04/05 (2)
- Re: Perhaps you are missing the point here... - Allen Wright 09:03:29 11/06/05 (1)
- Re: Perhaps you are missing the point here... - lovekylie 11:52:38 11/07/05 (0)
- Re: LK, no response to this data? (nt) - lovekylie 10:08:33 11/04/05 (0)