Use of a Ribidum Oscillator as refrence oscillator in a DAC - diyAudio
Go Back   Home > Forums > Source & Line > Digital Line Level

Digital Line Level DACs, Digital Crossovers, Equalizers, 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 5th January 2003, 02:27 AM   #1
diyAudio Member
 
Join Date: Sep 2001
Location: Salt Lake City
Default Use of a Ribidum Oscillator as refrence oscillator in a DAC

Has anyone ever done this? I heard that Sony made some special PCM units that used this type of oscillator. They are commonly available on E-Bay for a song and a dance.....
Mark
Attached Images
File Type: jpg ribedeum oscillator.jpg (49.1 KB, 1416 views)
  Reply With Quote
Old 9th January 2003, 01:33 PM   #2
alvaius is offline alvaius  Canada
diyAudio Member
 
Join Date: Feb 2002
Location: Toronto
Since no one is replying I will jump in. I guess the issue is that you don't have a lot of choice of frequency. What frequency range are these units generally operating at? Possibly you could use them to drive a DDS chip (Direct Digital Synthesis) to create a different frequency. You will still have system jitter issues, so I am not sure if their will be overall benefit. If you do buy one and use it, let us know how it goes, I am curious.

Alvaius
  Reply With Quote
Old 9th January 2003, 02:22 PM   #3
tiroth is offline tiroth  United States
diyAudio Member
 
Join Date: Dec 2001
Location: Pittsburgh, PA, USA
Yes, the issue of a useful frequency is a big problem.

Are there any DDS chips that have jitter tolerances worth looking at? I figure benchmark for a quartz oscillator is 1ps RMS and for an OXCO it should be better. Most of the synthesizers I've seen have been more in the 50ps range.

Something to chew on though: my DAC board has an asyncronous reclocker and a circuit to generate the audio clocks (MCK, SCK, LRCK) from the master oscillator. Since the ASRC is happy to clock the data out at any reasonable rate, there is no reason not to use a nonstandard rate, provided the DAC (AD1853) doesn't have a heart attack. I don't see why it should, but...maybe ask Analog?

Another route to go (that I am definitely going down) is ovenizing an integrated oscillator, or building a discrete oscillator and ovenizing the crystal.
  Reply With Quote
Old 9th January 2003, 03:54 PM   #4
diyAudio Member
 
nemestra's Avatar
 
Join Date: Jul 2002
Location: UK
Default PLL ideas

Hi Mark,

As alvaius has noted the output frequency is a problem. Most frequency standards output 1, 5 or 10 MHz. The jitter associated with a DDS chip is too high. A much better approach would be to design a frequency synthesiser based on a phase locked loop (PLL).

The output of the standard can be divided down to a much lower frequency e.g. 1 Hz and then used as a controlling input to a PLL. See http://www.rt66.com/~shera/ for an interesting article on the problems of PLL design. The standard reference work on PLLs is Phaselock Techniques by Floyd Gardner published by Wiley Interscience. More modern books are also useful.

This is not an easy task and the problem will remain of obtaining a high specification voltage controlled crystal oscillator (VCXO). For best stability this could be a temperature compensated TCVCXO or ovenized type OCVCXO. For the frequencies associated with CD replay these will have be custom ordered and will be expensive.

At least one crystal manufacturer does VCXOs as stardard or near-standard parts. Forssell and King describe a suitable 12.288 MHz VCXO in their Audio Electronics Five/1997 article 'Jitter Detector/Modulator Circuit Design'. See www.audioxpress.com for back issues. It was manufactured by Champion Technologies, distributed by Brookdale Electronic and the part number is K1525AE.

For improved stability the VCXO could be temperature controlled as tiroth has noted or indeed the whole PLL could be temperature controlled. Radio amateurs have been homebrewing OXCOs for years and back issues of the ARRL Handbook in the US and the Radio Communication Handbook in the UK will provide some ideas.

James
  Reply With Quote
Old 9th January 2003, 05:04 PM   #5
tiroth is offline tiroth  United States
diyAudio Member
 
Join Date: Dec 2001
Location: Pittsburgh, PA, USA
James,

Does a PLL design exist that can match the jitter performance of a simple quartz oscillator? Even Fox's JITO-2 PLL-based clock oscillators are over an order of magnitude worse than basic Valpey-Fisher quartz oscillators that are available for $1.30 each/100. It just seems like a lot of work to use a complicated circuit with a stable reference and then end up with something worse than your $1 crystal.

There are a number of companies that will custom-make OXCO with excellant jitter spec, but these are apt to be quite expensive.
  Reply With Quote
Old 9th January 2003, 05:18 PM   #6
diyAudio Member
 
nemestra's Avatar
 
Join Date: Jul 2002
Location: UK
Tiroth,

there are a large number of variables to trade off in any PLL design. It is possible to achieve 1ps RMS jitter figures from a suitably designed PLL.

To back up a bit, yes it is going to be possible to generate a clock for a CD transport or DAC using the rubidium oscillator as a master reference. Is is going to be easy - no. Its going to take blood sweat and tears and thats for someone who has done some oscillator or PLL design before. It going to require some time, some calculation and access to lots of fancy test equipment. Care will be required in layout, in power supply, in component selection etc.

A rubidium reference will give you long term frequency stability - if you periodically calibrate it against a known standard - cesium or passive/active maser. Is an absolute frequency accuracy of 1 part in 10^9 or 10^10 range important for CD replay - I don't think so. As many have highlighted the jitter / phase noise specs of the CD clock are much more important than the absolute frequency.

Is all of the above design and development work worth doing. Yes if designs like this interest you in themselves and you get a kick out of PLL design. At the end will you achieve better playback from CD than a good standard oscillator ( discrete, canned or ovenised ) ? Probably not. It depends on how much time you have available, I suspect.

James
  Reply With Quote
Old 9th January 2003, 08:56 PM   #7
diyAudio Member
 
Join Date: Jul 2001
Location: Orleans, France
The most important feature for a clock in digital audio is the lowest jitter. At present time, best oscillators are built with high quality quartz crystals, specially designed for this purpose. Rubidium oscillators are more stable over long periods, but exhibit higher jitter, and are not really suitable.

Regards, Pierre Lacombe.
  Reply With Quote
Old 10th January 2003, 12:52 AM   #8
diyAudio Member
 
Join Date: Sep 2001
Location: Salt Lake City
Thanks all for the replys!! It has given me alot to think about....
Mark
  Reply With Quote
Old 11th January 2003, 02:44 AM   #9
alvaius is offline alvaius  Canada
diyAudio Member
 
Join Date: Feb 2002
Location: Toronto
TCXOs and OCXOs are designed for long term stability, not short terms jitter performance. An OCXO will actually have higher jitter due to the added noise of the higher heat.

I have played around with cooling PLL circuits with good results in increasing jitter, but they will never be as good as a good crystal circuit.

Nemestra, are you sure that the DDS will be worse in terms of jitter than a PLL. That was true in the past, but with the new high speed products with 14 bit DACS, I am not sure. Just checked the AD9954 data sheet. It can achieve up to -125db phase noise. That sounds better to me than most PLLs, but I am going off memory. What are your thoughts on this?

FYI, for the rest reading, a good quartz oscillator can have phase noise down in the -150db or more range.

One avenue I want to explore is a SAW stabilized oscillator circuit. These can achieve phenominally low jitter performance and may not be overly expensive.

Alvaius
  Reply With Quote
Old 11th January 2003, 04:20 AM   #10
diyAudio Member
 
Join Date: Oct 2002
Location: Auckland NZ
Default oscillators

Hello Alvaius,
it is true that TCXO's are designed for long term stability but they can have very good phase noise performance. IMHO more than adaquate for clocks in audio equipment.

TCXO's are used in gps units which have very tight phase noise reqirements.

As the fundamental frequency increases the close in phase noise 1 Hz to 100 Hz increases this is because the Q of the quartz xtal decreases. You quoted -150 db at what frequency away from the carrier was this measured at.

Blair
  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
10 MHz oscillator possible to use ? Bernhard Digital Source 7 28th August 2007 09:07 PM
Look at this 42V/20A max PSU, and at this low THD oscillator. rephil Everything Else 3 6th August 2006 08:04 PM
oscillator dsavitsk Tubes / Valves 9 15th November 2005 06:07 PM


New To Site? Need Help?

All times are GMT. The time now is 01:27 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