Discrete Low Jitter Clock GB

[pchw]

Hi Anton,

I will change from 6 boards to 4 boards + 2 DC kit.

Thank!!

[touchdown]

Hi Anton,

I only need the bare PCB since I already have in my hands almost all components, except the comparator and the crystal whose frequency is different from the ones you suppply, but I can get them easily and quickly.

Please count me in for n° 2 bare PCBS.

Thanks

[Jack C.]

and for me 2 PCBs, please...

[amc184]

Thanks to all those participating in the group buy. As planned I am ordering the PCBs today. I'm ordering 100. 52 have been accounted for so far and I'm planning on keeping 20 to 30 for myself, so there is still some for others who still wish to join in.

After the PCBs have arrived and I've tested them I will confirm prices and start taking payments. I expect this to happen in two to three weeks from now. Unless something dramatic happens, the kit and bare PCB costs will be exactly what I estimated in the first post.

@pchw: Change noted, that's fine.

@touchdown: Thanks for the clarification.

[tassosk]

I´m in for 1 DC kit and 2bare pcb.

[pidesd]

hi amc184,

i ve been looking for a clock upgrade for a while and found this thread.

can you give a little background of how you came to choose this solution if it s not too much ask? or maybe you could point me to some documentation?

also, what characteristics make the j309 a good candidate? would bf862 be a good substitution? it has reportedly very low noise.

finally, i have a tough time figuring out what voltage does VCC1 runs at.
i m very tempted to build this clock with my own psu.

regards,

-py

[BFNY]

Hi
If not too late, I'd like to buy 5 bare boards.

I've been building these sorts of clocks for a good while, and may be able to take a stab at measuring the jitter, at least to the limits I can (~3ps)
Bob

[EUVL]

I couldn't help wondering what the output impedance of the TL431 based power supplies might be even at 11MHz, and whether it can cope with the >50mA pulsed current for the oscillator.

I presume you have done some measurements for the above, and perhaps some jitter measurements on top ?

Also I wonder if you change the values of C101~104 when going to 33MHz, and what the loading on the Xtal would then be ?


Patrick

[amc184]

@pidesd: Check out the first post for a bit of background on the design.

I had a quick look at the data sheet for the BF862 and couldn't see anything that would make it unsuitable. It does have a very low noise level, but it's surface mount, which rules it out for this PCB.

Both VCC1 and VCC2 are +5VDC.

@BFNY: You're not too late, I've added you to my list. Thanks for offering to measure the jitter, what technique would you use?

@EUVL: Your concern over the performance of the TL431 at high frequencies is valid. No spec is provided past 4MHz, but in reality I found the performance to be satisfactory. There are decoupling capacitors beyond the shunt regulator.

As I have said before, I don't have the equipment necessary to produce a quotable figure for jitter. I have taken both objective and subjective measurements (comparisons of the jitter at a CD player's output and listening tests) that indicate that it has significantly lower jitter compared to a stock CD player master clock.

I chose C101 to C104 to best suit a frequency of about 18MHz, the centre of its range. I have experimented with differing values of C101 to C104 for each frequency, but found no performance increase. It's also very practical to have that value constant.

[BFNY]

Originally Posted by amc184

@pidesd:

@BFNY: You're not too late, I've added you to my list. Thanks for offering to measure the jitter, what technique would you use?

I would try a software method based on using a high speed digitizer.

First would be in high resolution spectrum analyzer mode, with a huge FFT (5Meg data record). Second using a couple hundred thousand clock cycles analyzed in special software. I may try other methods too. Both above are limited to the digitizer clock jitter, which in this case is about 3ps rms.
Best way for this sort of oscillator is to use a phase noise system, but I don't have access to one.