The 'best' performing clock?

Status
Not open for further replies.
Jump to Latest
#101
Joseph K said:
poynton,

A very good reassumption!

Now, if you are not sticking it into XTI - then You are going to stick it into XTI! Too. But most importantly You are going to stick it into the CLK input of a D flipflop, which then itself disrupts the line between the dig filt / or decoder chip output and the DAC input. Which line, that's another question, depends on the DAC itself. Sometimes it's WCLK / WS somtimes it's BCLK.

In other words, now that you have a good clock, you have to duplicate it's output, and distribute it between XTI and the CLK of the D-flop.
Without creating ground loops.. Because the XTI in fact expects a sinusoid signal, what I would do is to start from a clock design which squares up a sinus oscillator, would buffer it, take one buffered output & square it up, then pass the other output through a small high freq pulse transformer, and forward this small sinusoid to the XTI. This way one can avoid of creating a new ground loop. In the process care should be taken to respect the setup / hold times for the reclocking flip flop, which usually means delaying adequately the signals. A usual trick is to invert / or not the branch which goes to the XTI. The small transformer provides this easily. One can also play with the length of cable travelling to the XTI.

Now, all these operations should be done in a way that does not disrupt the continuity of the fast signal current loops. [Supposed that there existed such a continuity in the first place.. which is almost never..]
Now, this sentence above is a pure contradiction in itself. So, it is here that I would add to the claim of RBFW [you still have to put it together competently..] - there is no such a thing that put it together competently..
About the difficulties of doing it properly I would suggest to read up the saga of Pete Goudreau about his dac mods. I have some huge text files..

Conclusion - go for a DAC design where all this has been taken into account right from the start. Hagerman? Guido Tent? Rakk dac? Pedja Rogic?

Ciao, George

Ps.: There exist somewhere also a design by Jocko..
Click to expand...


HELP , my DAC does not have a clock 🙁
 
#103
poynton said:
Whatever clock you chose can always be compromised by mediocre construction and installation.
Click to expand...
But the XO module requires no construction and the power supplies have been chosen.
Owners of the the XO2/3 clocks have achieved excellent results. The basis of the XO2/3 is the XO module. It works straight out of the box. Equally good results can be achieved with the Elso's and Jocko's clocks but unlike with the XO module, some user assembly is required and some users are more capable than others.

The 3.5 uses the TDA1305 a combined digital filter/dac so XTI is exactly where the clock is going. Jocko's point is about not usng the digital filter for clock distribution.
If you want a single inverter use a Picogate.

Joseph K said:

Conclusion - go for a DAC design where all this has been taken into account right from the start. Hagerman? Guido Tent? Rakk dac? Pedja Rogic?

Click to expand...

See post #1.
 
#105
Maybe I should have been more clear.......

Someone did it for me:

"Jocko's point is about not usng the digital filter for clock distribution.
If you want a single inverter use a Picogate."

Use a separate build-out resistor for each line. Simple as that.

You don't need a '04, with all 6 inverters firing away. Second best way to muck it up. Second most popular. Usually found in "high-end" designs only.

How did we get sidetracked into jabbering about D/A boxes?????????

Jocko
 
#107
That version was needed to take a signal sent over a 75R coax, to a CDP, and square it up. You don't need most of those parts if you put it next to the clock.

Like the HF termination.

Or the gain setting resistors.

The picogates came from Mouser.

Jocko
 
#108
rfbrw said:
Owners of the the XO2/3 clocks have achieved excellent results. The basis of the XO2/3 is the XO module. It works straight out of the box.
Click to expand...

Ah well, then there's the PSU.
The Tent XO2/XO3 has good regulation on-board, but you still need to make a good PSU for it. Dedicated, including the trafo.
I have used Guido's own PSU for the clock once, L-C unregulated.
Except that time, I always make an RCRC unregulated PSU (dedicated trafo, MUR1100 diodes), with very good results.

Someone can still throw way the baby with the water by not using a good dedicated PSU.
Just imagine, some will just connect the clock's PSU wires to whatever point on the player that has ~12V or above. 😀
 
#109
Originally posted by carlosfm
Ah well, then there's the PSU.
The Tent XO2/XO3 has good regulation on-board, but you still need to make a good PSU for it.
Click to expand...

But in this case there is a decent PSU

Originally posted by Hacker
The clock, powered from the above circuit + ALWSR will be connected into the holes left by the soon-to-be-removed oscillator. Now comes the question. Which clock will perform best under these conditions? If you have figures to back up opinions, that would be great.
Click to expand...

'above circuit ' being a C- multiplier.
 
#115
fmak said:


Guido

Are you going to try one? I may if it's available and I feel rich!.
Looks too big as in most cases.

I would like a surface mount miniature clock with PS, frequency configurable! Dream?

Fred
Click to expand...


Hi Fred,

No, I am not going to spend money on these.

Frequency configurability always affects performance.....

What would be your preference in terms of dimensions of a clock board ?

best
 
#116
Hi Fred,

No, I am not going to spend money on these.

Frequency configurability always affects performance.....

What would be your preference in terms of dimensions of a clock board ?

best [/B][/QUOTE]

It would be good to have two pieces. The clock module as small as possible to go in place of the stock crystal/XO which are now smd, and a PS not too big to go near.

The ALW version of the Jung Regulator actuallly has lower noise than any other I have come across including your own XO3 which itself can be improved (1/3x) with OSCon decoupling. This all in situ with XO coupled. I have found that this is the only way to both measure and inspect with a 400 MHz Tek scope.

If only this could be all SMD but I guess cost of manufacture and repairability are factors. But when you stick to 11.286 MHz it is OK to have largish modules. 45.1584 clocks are much more critical and also few have really good wave shapes.

If you want to see good clock signals thru'out, the dCS stuff is near impeccable! What you see in the manual is what you get. Even then, external clocking helps the sound a lot.


🙂
 
#118
Elso Kwak said:
Wave shapes are totally unimportant!
The best comparators and regulators are SMD.................
Agree on Oscons.
I like PCBs slighty smaller than the size of a credit card.......
Click to expand...


Why do you say wave shape not important - impedance matching ensures good waveform everything being ok. Trigger point depends on edge and averaging important.

Credit card too big 1/3 or 1/2 size max.

Best regulator may not be smd. There are BIG ground problems with some newer smd regulators ie LT. Really needs a LARGE plane, otherwise serious 50Hz pickup.
 
#120
fmak said:



Why do you say wave shape not important - impedance matching ensures good waveform everything being ok. Trigger point depends on edge and averaging important.

Credit card too big 1/3 or 1/2 size max.

Best regulator may not be smd. There are BIG ground problems with some newer smd regulators ie LT. Really needs a LARGE plane, otherwise serious 50Hz pickup.
Click to expand...


Hi Fred,

Elso never carries out measurements, so it doesn't make sense to discuss technical issues with him.

I agree on your remarks, but do not understand your note on the ground problems, could you eleborate ?

best
 
Status
Not open for further replies.
Top Bottom