Clock experiments
 

Over the last year, I've been using a special rubidium clock as master clock for my DCS system. The clock is a prototype from one of the leading digital timing companies in which they synch rubidium with an ocxo to get the benefits of both.

Because this was such a step up from the standard Scarlatti and Vivaldi clocks, I decided to investigate what further refinements in clock accuracy were possible. Last week I received here a prototype clock using a BVA crystal. This is the ultimate development of the OCXO to improve phase noise by an order of magnitude.

Results? A huge increase in sound quality through a reduction of remaining digital edge. I am amazed how much cleaner the sound is. Sound has an analogue sweetness but with great resolution.

Downsides? Two main ones. First cost: the only commercial example made for the HiFi world is $35k although should be possible to build one for nearer $10-15k; second drawback is stabilisation is quite slow, typically 48 hours to get to peak performance.

It really is surprising how big an effect these timing refinements can make to sound quality.

Will an Esoteric clock work with your system? Mine was around 16K, it made a significant difference with my Esoteric separates. 35K for rubidium clock seems absurd. I use them in the radar system I manage and they are closer to 5K.

You misread my post. I have replaced the rubidium clock with a BVA clock. A different animal altogether. Short term accuracy is an order of magnitude ahead of the Esoteric clock!

Can you provide some more info on the BVA clock?

It's a prototype built around an Oscilloquartz BVA8607b double oven OCXO mounted in a damped enclosure with a stiff power supply. A bit of a bench top prototype but not so difficult to turn into a finished product.

Very interesting post. I've been creeping up the dCS stack...I know the clock and the clock cables make a BIG difference. I've used the dCS U Clock and currently the dCS Scarlatti Clock.

Clock cables i've tried are the stock dCS, Cardas Lightning 15 (excellent value and smashes the stock cables), Purist Proteous Provectus Praesto and now Transparent Ref XL digital.

All BNC to BNC

Speakers: Rockport Altair 2
Preamp: Robert Koda K10 Amp: Ayre MX-R
Cables: Transparent Opus MM2 Spkr, Opus MM2 & Trans Ref MM MM2 Intcnct.
Power Shunyata Triton/Typhon, Anaconda Zitron cords (x9)
Digital dCS Vivaldi DAC & dCS Scarlatti Clock & Transport with Transparent Reference XL Digital cables
Phono: Grand Prix Monaco TT, Schroder LT arm, Jan Allaerts MC2 Finish & Lyra Atlas, Nagra VPS/VFS
Racks: Grand Prix Audio Silverstone F1 (x2)

There are big improvements to be had by installing a master clock with higher short term accuracy than the DCS clock. Certainly clock cables also worth experimenting with. I've had good results with the Vertex range although they are not well known in the USA. Apart from those, my cables of choice are Siltech with MIT for power to digital components. But do improve the clock before the cables: there's more to be gained with doing that.

Are u talking about the master clock or the reference clock ?

Sorry about that, I will have to look into it.

I'm talking about adding an external reference in addition to the DCS clock. The additional clock outputs a reference 10mHz signal which the DCS clock locks to rather than its own Clock.

At those prices you have just forfeited your rights to making comments on the cost of Analog. :music:

Not something I have any interest in, having sold my Goldmund Studio and 5000 records many moons ago. These days even the hassle of dealing with thousands of silver disks is beyond me. Anything that can't be selected and managed from my iPad is far too much hassle!

custodian,

Thanks for the info. Any chance this will become a commercial product ?

I am using the Antelope clock & it does make a difference.

Regards,
Joe Ling

The Antelope rubidium clock has quite poor phase noise. There are much better rubidium clocks available with improved phase noise and the BVA clock is an order of magnitude better again.

There is no problem in producing as BVA clock for HiFi use apart from cost. Having looked at costings, I think it would be at least $20k, maybe a bit more. Do you think there would be a market at that level?

Anything commercially available at the moment better than the Antelope ?

I'll have a look at which rubidium clocks are used in the commercial units and get back to you. From memory, I think the Esoteric one uses a clock with better phase noise but I need to check. There is a commercial BVA clock for HiFi use made in Japan which retails at $35k but I think I could do an equivalent one for half that cost.

The longer the BVA clock operates, the better the phase noise performance. I must say the clock change to BVA has produced a staggering improvement in quality.

I have been using the Antelope Audio OCX clock master clock and a Stanford Research Systems Rubidium 10MHz oscillator (FS725). A material improvement to the internal clock of my Soulution 745 SACD/CD player. Interconnect and power cables also make a noticeable difference.

Davidny... Welcome to AA! :wave:

I am very curious here. I have the Vivaldi Transport ordered...hopefully it'll be here in a week or so. I have kept my Scarlatti clock for the time being...I got to demo the Vivaldi Trans...and it made quite a difference in my system vs my dCS Scarlatti Trans...I don't know if the Vivaldi Clock...the clock specs to the Scarlatti are the same..the Vivaldi clock adds dual frequency outputs...which can be useful (but I can get the same results by pressing a button on Scarlatti to change the frequency and don't have to buy another cable)

Using the reference clock input I've always been curious about...

Speakers: Rockport Altair 2
Preamp: Robert Koda K10 Amp: Ayre MX-R
Cables: Transparent Opus MM2 Spkr, Opus MM2 & Trans Ref MM MM2 Intcnct.
Power Shunyata Triton/Typhon, Anaconda Zitron cords (x9)
Digital dCS Vivaldi DAC & dCS Scarlatti Clock & Transport with Transparent Reference XL Digital cables
Phono: Grand Prix Monaco TT, Schroder LT arm, Jan Allaerts MC2 Finish & Lyra Atlas, Nagra VPS/VFS
Racks: Grand Prix Audio Silverstone F1 (x2)

The DCS clock is a standard OCXO of pretty good spec.

Using the master clock input on the DCS clock, you can use an external master clock which might improve sound.

Options for master clock are:

1. GPS clock - not a good choice because of poor short term stability

2. Rubidium clock- can be a real improvement however basic rubidium also has poor short term stability and you really need something where rubidium and OCXO are working together to get better performance.

3. BVA OCXO clock which is ultimate short term stability but at a price. Only commercial unit designed for audio use costs $35k!

BVA really makes a huge difference to sound quality in removing any hint of digital edge.

jfrech, according to my dealer the Vivaldi clock is better then the Scarlatti. I have not tried both but when they did the Vivaldi, they improved the internal circuits and power supply compared to the Scarlatti.

This is indeed an very exciting topic as pointed earlier in this thread. I am confused on this topic as a year ago I did experiment with the Antelope Rodium reference clock with my Vivaldi stack. At first I was impressed but after a few days of listening I felt the sound was colored a bit and could not understand why.

I called dCS directly since I had met the rep and he then had told me dCS apparently a while back looked into reference clock ... in the case the Antelope, he suggested its switching power supply was not ideal and even if you feed the dCS clock with a perfect 10Mhz signal, it still needs to be converted to 44.1 and 48khz internally using most likely a less then ideal chip set. otherwise not required when using the internal 44.1 clock directly. So while in theory using a reference should be better, they moved away from it.

I still would like to try it again and will look into some of the other choice you mention in your post. But, would it make any sense to any of you who experimented with this that the "conversion chip" that takes the 10Mhz to make it 44.1Khz inside the dCS clock may actually impact the desired result to the point using the internal being superior?

Are any of these reference clock you describe come with a linear power supply? I suspect why esoteric may have good result with their clock is because the Rodium reference is actually in the main Esoteric chassis, power by a linear power supply.

As you can see, many questions on this exciting subject!

Jacques

Jacques
Interesting post.

You are right, however DCS use conversion electronics for the inbuilt OCXO to generate the 44.1 and 48 signals.

The Antelope rubidium clock actually doesn't have particularly good phase noise which means it will not be major improvement when used as master clock for a DCS system. In simple terms, rubidium will have better long term stability but that has no real value in this application. There are techniques to link the rubidium source with an OXCO to produce an external master clock which improves on the basic DCS clock. I've done this on both Scarlatti and Vivaldi clocks with good results.

However at this level, a BVA OCXO gives much better results.

One other option I've not looked at would be to replace the OCXO in the Vivaldi with a performance optimised version of the same component. That might be worth looking at.

In the external master clock, psu and cabling are all of key importance in tweaking performance. Commercial problem with BVA remains the high cost of the crystal which means market very limited. Currently only available audiophile clock is $35k. I'm sure it could be produced at maybe $20k but even then, sales volume very limited.

Am I the only one that has no idea what this thread is talking about? It's in English but reads like Greek to me. Where can I go to get 'clocking' for dummies? This thread is also an other reminder that there are whole other levels of this science and hobby which I am clueless to.

crwilli, no worries I am confused too about this topic sometime! One good place to learn about digital clock and how they are use/important in digital playback for DAC like the dCS ones, is actually on dCS website.

If you go on dCS website and read a bit on the dCS Vivaldi "stack" (aka clock+upsampler+DAC+ transport) they explain what each unit does and why they have a separate component for the clock - to synchronize all the digital equipment together. Think of it this way: no clock or bad clock = everything bad about digital sound; while good clock = a digital playback that could rival with analog playback...

Hope this helps...

Sorry. Didn't mean to complicate it.

I only got into this stuff when I had to scratch the urge to try the external master clock socket on the DCS system. It is inevitably a tricky subject. All I can say is that better short term accuracy = better sound.

Rubidium clocks don't necessarily improve sound because most of them don't improve short term accuracy.

Any of the external master clock options cost money so if you want to jump in, make sure you can try for a while first.

A friend of mine has a recoding studio and I will ask him if he would know other brand of potential solution like you describe. Your comment about the Antelope would explain why I had mixed result with it. I kept scratching my head because logically I was so convinced there is a way to make the clock stock dCS clock sound better with a reference clock.

Would you be able or mind to expand what techniques/toold you use the improve on your Scarlatti and Vivaldi clock "link the rubidium source with an OXCO to produce an external master clock which improves on the basic DCS clock"?

How could one do that?

Jacques

+1 ! I also only started to wonder after getting the dCS clock - but it is indeed a subject with very few folks having the experience for these things it would seems. I know when I started to ask about a reference clock to my dealer he was looking funny to me: what do you mean I just sold you a clock!

:scratch2:

i will do that! Thank you.

Custodian - never apologize for bringing state of the art and beyond to AA. We all want to learn. Some just do it faster than others ;)

Having played with a few rubidium clocks and being confused with the results, I contacted a UK group which specialises in producing the highest accuracy timing devices for research. They were very kind and spent some time explaining the different options to me and they produced two prototype rubidium clocks with good short term stability. I thought they improved on the DCS clock. I did do a demo to the DCS sales director who was also enthusiastic about the results.

Sadly there is not much available "off the shelf" that I believe would really improve over the standard DCS clock. The Esoteric clock seems to have good performance although I haven't tried it. The Antelope is not, in my opinion, any improvement over the DCS internal clock. It sounds different but not better.

My view is that if you are going for an external master clock, go for a BVA. Expensive but really big improvement: it makes digital sound totally analog! As I posted previously, there is a Japanese company offering a commercial audiophile product but it is expensive. I got mine custom made using a standard Swiss made BVA crystal. A Scarlatti stack with a BVA master clock will comfortably outperform a Vivaldi stack using the DCS clock alone (in my opinion, for what that might be worth)

I have been trying to persuade the UK group to produce a commercial audio clock product but they are not convinced the market size justifies the effort which is a pity. The founder and owner of the UK group sadly died recently so there is some uncertainty over the future business direction. When things settle down, I'll revisit trying to encourage them to pitch an audio product.

I have had a similar urge with my Marantz SA7-S1. Not dCS I know, but...

Try it. That's a very good unit so worth doing. Does the clock input take a standard 10mHz signal?

Custodian, I was able to find the BVA master clock you reference to. at $37k it is very expensive indeed but seem extremely well built! The same company seem to have a smaller model but not sure it is a BVA master clock. Anyway I will try to educate myself a bit more a this...

It does look very well constructed. That machined case reminded me of my Ayre KX-r preamp case. The Japanese distributor was helpful and offered to fly over to demonstrate the unit! It uses the same crystal as I have in mine. The component cost for the crystal is around $12k, so it is always going to be expensive. I think a commercial product could be $20k

Wow! They offered to fly for a demo! Then there is surely some room in their margin :-) So you have not found any other company doing this right? I will definitively try to learn more on this...

I couldn't find any other company offering a ready made product for audio use. It would not be too difficult to source the crystal and build such a device. The BVA used is made by Oscilloquartz in Neuchatel, Switzerland. Model number is BVA 8607. It has a 10x better short term accuracy than the best OCXO clock and is sold as a replacement for atomic clocks.

Humm...ok here is a stupid question! If it is sold as a replacement for atomic clock, could one buy a commercially available atomic clock (not sure if Antelope has one but I think they do) and then buy this BVA part number to replace the generator of the 10Mhz signal? This may be an oversimplification but just a thought!

Jacques

Not a simple straight replacement but could be possible. Although Antelope doesn't have special case or great power supply so would be much better to build new unit with custom case and high quality power supply. The lock electronics are not the same for a BVA clock so that would need changing if you tried to do a straight swaps

Ok thanks. I think I just found my next DIY project!