[f] AMB on Facebook

Will ο1 benefit this design?

omicron1: A low-jitter, low phase noise clock generator for upgrading digital audio devices

Will ο1 benefit this design?

Postby whomper » August 18th, 2018, 1:49 pm

Hi,

I have a ADC DIY module that I'll be connecting to the optical in of my audio interface. This will probably mean master clock is generated by the ADC, so I'm thinking of the clock quality.
Enclosed is a screen capture of the clock and convertors circuitry for review.

Could o1 deliver a better clock than what's already there?
You do not have the required permissions to view the files attached to this post.
κDC2496, γ1.5, γ2, γ3, α10 (α24 based), balanced β22, 2 x β24, M³, Grado ES2E, GS2000E, AGK K700, CBT36 Speakers
whomper
 
Posts: 264
Likes: 21 posts
Liked in: 34 posts
Joined: January 12th, 2012, 11:57 am
Country: Israel (il)

Re: Will o1 benefit this design?

Postby amb » August 20th, 2018, 9:43 pm

Maybe. You can probably feed a better clock into the MK2703B, but you cannot remove the MK2703B because it also handles PLL. That is, if you can find a high performance 27MHz oscillator for the ο1.
User avatar
amb
Site Admin
 
Posts: 10422
Likes: 176 posts
Liked in: 493 posts
Joined: March 29th, 2010, 9:49 pm
Location: Sunnyvale, CA
Country: United States (us)

Re: Will ο1 benefit this design?

Postby LV26 » September 17th, 2018, 12:45 pm

I think MK2703 PLL jitter kill o1 performance.
Everything should be made as simple as possible, but not simpler. // Albert Einstein
User avatar
LV26
 
Posts: 431
Likes: 55 posts
Liked in: 16 posts
Joined: May 7th, 2014, 1:46 am
Location: UKRAINE
Country: Ukraine (ua)

Re: Will o1 benefit this design?

Postby Balcmeg » December 28th, 2019, 8:06 pm

amb wrote:Maybe. You can probably feed a better clock into the MK2703B, but you cannot remove the MK2703B because it also handles PLL. That is, if you can find a high performance 27MHz oscillator for the ο1.


I know this is an old thread, but I stumbled upon it when I was looking for info and I thought my question relevant to this.

What is it one should look for in finding a good oscillator? Obviously operating voltage, logic levels, footprint and frequency - but what is important apart from that, and what values are considered "high performance"?

They all seem to be HCMOS - apart from Gollard wich is CMOS?
Frequency stability would be another factor - and all you recomme3nd for o1 are at +/-25ppm.

Anything else?
/Stefan
Balcmeg
 
Posts: 580
Likes: 25 posts
Liked in: 8 posts
Joined: September 6th, 2014, 6:54 pm
Location: Hong Kong
Country: Hong Kong (hk)

Re: Will ο1 benefit this design?

Postby amb » December 29th, 2019, 2:15 am

The things you kmentioned are all important of course, and fundamental to narrowing down your choices. Don't forget that the physical size and pin-out also limit your choices if you're trying to fit them on a particular PCB.

Aside from those, the single most important characteristic for high performance digital audio is jitter (or phase noise, which is basically the same thing expressed a different way). The lower the jitter or phase noise, the better. Unfortunately, many "general purpose" oscillators don't even specify this in their datasheets, except those specialty oscillators intended for digital audio or other applications where such characteristic is important. Also, even if you find these in the datasheets, they are not usually directly comparable between manufacturers (or even models) because the test conditions are different.

Lacking the proper (expensive) test equipment and environment to test them, there is no easy way for the typical DIYer to select oscillators in a scientific way. But if you find oscillators touted with low jitter/phase noise performance, then you're probably in the ball part. Other than that, search for other people's comments about their "favorite" oscillators" applied to digital audio circuitry.
User avatar
amb
Site Admin
 
Posts: 10422
Likes: 176 posts
Liked in: 493 posts
Joined: March 29th, 2010, 9:49 pm
Location: Sunnyvale, CA
Country: United States (us)

Re: Will ο1 benefit this design?

Postby Balcmeg » December 29th, 2019, 2:54 am

Thanks Ti,
Yes I had a feeling I would end up having to trust the internet instead of reading datasheets at the end of things.

I noticed that I am actually looking to build the same ADC as whomper posted in the first post here.
Either by replacing Q1 with an omikron1 27MHz - which seems to be hard to find oscillator of HiFi quality, but I am also hoping that it might be possible to replace the PLC and oscillator (IC6 and Q1) with omicron1 so the MK2703B can be eliminated.

From what I can see in the schematics and the instructions, the frequency out of IC6 is controlled by a combination of S0/S1 being either High or Low, so 4 combinations are possible: 8.192, 11.2896, 12.288, 24.576MMHz.
Out of these 12.288 and 24.576MHz are used for the same group of sampling frequencies (16, 24, 32, 48, 64, 96 and 192kHz), the complement to this is 11289.6MHz, which give the remaining 22.05, 44.1 and 88.2kHz sampling frequencies. The mentioned frequencies are the ones that can be selected by a rotary switch wich, in turn, control the output of a Programmable Logic Device wich send the signal to S1/S1. With luck only 2 clock frequencies are used from the IC6 and:

S1=0 : S2=1 -> 11289.6MHz
S1=1 : S2=0 -> 12288.0MHz
S1=1 : S2=1 -> 24576.0MHz

I would think that 11289.6MHz HAS to be used as one of the frequencies, and then the remaining would be either 12288.0MHz or 24576.0MHz. Hopefully, 24576.0MHz since more of the HiFi oscillators are available here - and this CLCK is where sample rates I am most interested in are (96 and 192kHz).
If it is 11289 and 24576, then I should be able to use S1 and connect to SEL on omicron1 and get the right clock signal. 11289 seems to be harder to find good oscillators for, but i don't see much need to have 44.1/88.2 etc out - so if that one has higher jitter I would not be devastated - only a bit annoyed not to be able to make it as good as I want.

Going back to the fact that I only am interested in 96 and 192kHz - If it will not work to use S1/S2 to select CLCK. then I would be OK to install a 1 Freq omikron1 (12288 or 24576) and leave it as a setup with fewer choices of sampling frequencies - I don't plan to have the selector switch accessible from the front anyway so it's not a big deal.

As I see it - if I set the switch to sampling freq 96kHz and have a fixed CKCK from omikron1 to 12288 or 24576kHz (with IC6 removed) - then there is no need to select sampling clock frequencies.

I have mailed the designer about this and hope he will get back about what frequencies are used - if not I can just build one and see what I get on the S1/S2 outputs and take it from there.

(I will also have a second look at the datasheets - but from my basic understanding, I think I can make it work)
/Stefan
Balcmeg
 
Posts: 580
Likes: 25 posts
Liked in: 8 posts
Joined: September 6th, 2014, 6:54 pm
Location: Hong Kong
Country: Hong Kong (hk)

Re: Will ο1 benefit this design?

Postby Balcmeg » December 31st, 2019, 8:11 pm

Happy new year everyone!

Ok, so I have been in touch with the designer and there is both bad and good news.
The good is that it would indeed be possible to replace the PLL with omikron1 and he agrees that this switch would a positive change.
The bad thing is that he is using 3 different clock frequencies, not 2 as I had hoped, and to get around this e reprogramming of Logic Device would be needed - something IM not prepared to do.

On the other hand - by using an omikron1 with 24576kHz oscillator (which is not that hard to find) I can get 48, 96 and 192kHz sampling frequencies from the ADC. more or less enough for me. since he uses a 16 positions HEX switch to control the settings, this would mean that frequencies are duplicated or in strange combinations in regard to copyright markings and copy protection, but nothing major (apart from the fat you need to remember your settings - and I don't plan to change them often - just set and forget after I found the one I prefer.

Moving on, if we add a 11.2896MHz oscillator as well we extend the available sampling frequencies to 22.05k, 44,1k, 88,2k as well as 44.1 with and without copyright/copy protection.
The only freq that is not covered would be 16k, 24k, 32k, 64k, and 48k copyright/non-copyright, etc.

Omikron1's SEL input would be controlled by S1 input of the PLL:
s1 = 1 for 24.576MHz
S1 = 0 for 11.2896MHz

I have been looking at the truth table to see if there is any way to hardwire any outputs and remove the unwanted positions that way, but there doesn't seem to be any.
Unless I get any other ideas, right now I plan to get a good 24.576MHz oscillator (Crystek) and use a more basic one for the11.2896MHz which I do not plan to use often - if at all. Its mostly a matter of testing if things work as planned.
/Stefan
Balcmeg
 
Posts: 580
Likes: 25 posts
Liked in: 8 posts
Joined: September 6th, 2014, 6:54 pm
Location: Hong Kong
Country: Hong Kong (hk)

Re: Will ο1 benefit this design?

Postby Balcmeg » December 31st, 2019, 8:55 pm

Ti,
For X1 (24.576MHz) I managed to get the Crystek ‎CCHD-957 - seems to be a pretty common frequency, but the X2 - 11.2896MHz is more difficult, even among the cheaper oscillators.
I would prefer to order from either Digikey or RS but can't find any there or at Mouser - am I looking at the wrong places (i.e. would you know about any offhand?) or should I just ignore the dual setting and go for the 48/96/192kHz samplings frequencies alone?
/Stefan
Balcmeg
 
Posts: 580
Likes: 25 posts
Liked in: 8 posts
Joined: September 6th, 2014, 6:54 pm
Location: Hong Kong
Country: Hong Kong (hk)

Re: Will ο1 benefit this design?

Postby amb » January 1st, 2020, 9:34 am

How about Digi-Key XC1641CT-ND or 1253-1436-1-ND? The latter has reasonably good specs (even though there is a "low phase noise" version of it which no one seems to carry).
User avatar
amb
Site Admin
 
Posts: 10422
Likes: 176 posts
Liked in: 493 posts
Joined: March 29th, 2010, 9:49 pm
Location: Sunnyvale, CA
Country: United States (us)

Re: Will ο1 benefit this design?

Postby Balcmeg » January 1st, 2020, 4:26 pm

amb wrote:How about Digi-Key XC1641CT-ND or 1253-1436-1-ND? The latter has reasonably good specs (even though there is a "low phase noise" version of it which no one seems to carry).

Thanks Ti,

Will try this one out.

Hopefully I will remember to come back here and update if o1 works in this setting.
/Stefan
Balcmeg
 
Posts: 580
Likes: 25 posts
Liked in: 8 posts
Joined: September 6th, 2014, 6:54 pm
Location: Hong Kong
Country: Hong Kong (hk)


Return to ο1 precision master clock

Who is online

Users browsing this forum: No registered users and 0 guests

Quick Links

AMB audio DIY resources
AMB audio shop