LClock XO PSU

jackinnj · 2004-08-12 1:56 pm

Re: Re: Re: Re: Regulated Powersupply for Clock

Guido Tent said:
Again, I believe that a properly RC filtered voltage put in an emitter follower can be dead silent.

-

In Jan Didden's part of the Super-Reg series (3/95) he used an RC filter -- these old magazines are really worth saving.

fmak · 2004-08-12 2:05 pm

Re: Re: Re: Re: Re: Regulated Powersupply for Clock

In Jan Didden's part of the Super-Reg series (3/95) he used an RC filter -- these old magazines are really worth saving. [/B][/QUOTE]
----------------------------------------------------------------------------------
Indeed the ALW version of the superreg works very well at 5V with digital circuits. However, in implementaing supply, I always measure noise interaction with connected circuit and look for instability with a 400 MHz scope.

Elso, using your ears with your own clock is fine, but you cannot generalise on this basis.

:smash:

Guido Tent · 2004-08-12 3:25 pm

Re: Re: Re: Re: Re: Regulated Powersupply for Clock

jackinnj said:
In Jan Didden's part of the Super-Reg series (3/95) he used an RC filter -- these old magazines are really worth saving.

Sure, after all there is not that much news.....

The downside is that sub LF noise (<10Hz) is hardly filtered by RC filters.

Anyone seriously involved with jitter knows that sub LF jitter is most audible.....

cheers

carlosfm · 2004-08-12 3:44 pm

Re: Re: Re: Re: Re: Re: Regulated Powersupply for Clock

Guido Tent said:
The downside is that sub LF noise (<10Hz) is hardly filtered by RC filters.
Anyone seriously involved with jitter knows that sub LF jitter is most audible.....

cheers

Ah yes, spot on, as usual.

Jitter affects much more the bass than anything else.
Dacs, digital filters, digital receivers, etc. "reject" jitter (as per manufacturer's claims), but only at high frequencies. :clown:

Elso Kwak · 2004-08-12 5:05 pm

Re: Re: Re: Re: Regulated Powersupply for Clock

Guido Tent said:
dear Elso,

Since you are so convinced, could you indicate the measured noise ?

There is nothing against the Wenzel circuit, except that it needs adjustment.

(It is by the way not his invention - I found examples in older spectrum analyesers using similar techniques, and I even found valve amp supplies using this trick)

Again, I believe that a properly RC filtered voltage put in an emitter follower can be dead silent.

This is a copy of the post you point at:

Originally posted by fmak

-------------------------------------------------------------------------------
No need. I measured to 1 MHz bandwidth and it is as specified.

It's one thing measuring PS noise using a resistor; it's another measuring it with the PS connected to the XO and injecting the clock signal. I try to silence the PS using the latter.

The Audiocom Supetregulator collapses under XO load and spews out hf noise spectra like someone is going sick!!

It's useful to do a spectral analysis of both PS and clock signals.
Some XOs have odd harmonics just 10dB down from the fundamental in a more or less monotonic manner.

Hi

nothing to add...............

keep posting Mr mak !

Ciao
__________________
Guido Tent

So not a single reference to SOT 23 regs at all, or do I miss your point ?

best regards
-

Sorry, I meant this posts:
http://www.diyaudio.com/forums/showthread.php?
postid=191343#post191343

http://www.diyaudio.com/forums/showthread.php?postid=188512#post188512

These clocks are replaced to enhance the sonics so I judge them for that, not for measurements.
I have found no direct correlation between sound and noise level. I tried two SOT-23 regulators with the same noise spec. One sounded totally uninspiring and boring, the other was extremely good! And the result was much better than with a simple RC filtered follower.

Elso Kwak · 2004-08-12 5:18 pm

Re: Re: Re: Re: Re: Re: Regulated Powersupply for Clock

Guido Tent said:
Sure, after all there is not that much news.....

The downside is that sub LF noise (<10Hz) is hardly filtered by RC filters.

Anyone seriously involved with jitter knows that sub LF jitter is most audible.....

cheers

So Why then promoting the RC filtered voltage into a follower?

Guido Tent · 2004-08-12 5:26 pm

Re: Re: Re: Re: Re: Regulated Powersupply for Clock

Elso Kwak said:
Sorry, I meant this posts:
http://www.diyaudio.com/forums/showthread.php?
postid=191343#post191343

http://www.diyaudio.com/forums/showthread.php?postid=188512#post188512

These clocks are replaced to enhance the sonics so I judge them for that, not for measurements.
I have found no direct correlation between sound and noise level. I tried two SOT-23 regulators with the same noise spec. One sounded totally uninspiring and boring, the other was extremely good! And the result was much better than with a simple RC filtered follower.

first post

vBulletin Message
No thread specified. If you followed a valid link, please notify the webmaster

second post refers to LT1763 SMT which appears to give 10nV/SqrrtHz hence is a good one. Am I allowed to promote that ?

regards

Guido Tent · 2004-08-12 5:29 pm

Re: Re: Re: Re: Re: Re: Re: Regulated Powersupply for Clock

Elso Kwak said:
So Why then promoting the RC filtered voltage into a follower?

because it is way better than any of the 317 or 431 based rubbish and easy to implement for starting DIY

Life is all about compromises

Have you already measured LED noise ?

Guido

carlosfm · 2004-08-12 5:51 pm

This thread was oppened to discuss a good solution to pre-regulate (or not) to something around 12v.
I don't see anyone talking about this...

The regulation is already on the LClock XO, and it's not easy to change it. :clown:

Elso Kwak · 2004-08-12 5:55 pm

Re: Re: Re: Re: Regulated Powersupply for Clock

Guido Tent said:
There is nothing against the Wenzel circuit, except that it needs adjustment.
-

There is much wrong with the Wenzel circuit. As with all cancellation circuits it is extremely sensitive to misadjustment. Also the cancellation does not work at high frequencies like used in the clock.
It was all explained at this Sympatico site but unfortunately it is down.
http://www3.sympatico.ca/add.automation/spice/ripple.htm
:bawling:

Elso Kwak · 2004-08-12 5:58 pm

Preregulate

carlosfm said:
This thread was oppened to discuss a good solution to pre-regulate (or not) to something around 12v.
I don't see anyone talking about this...
The regulation is already on the LClock XO, and it's not easy to change it.

Spot off!
I found that pre-regulation makes soundquality worse.
So I don't use it anymore. This does not apply to PI-filters.
:bigeyes:

Elso Kwak · 2004-08-12 6:02 pm

Re: Re: Re: Re: Re: Re: Regulated Powersupply for Clock

Guido Tent said:
first post

vBulletin Message
No thread specified. If you followed a valid link, please notify the webmaster

second post refers to LT1763 SMT which appears to give 10nV/SqrrtHz hence is a good one. Am I allowed to promote that ?

regards

First thread must be:
http://www.diyaudio.com/forums/showthread.php?postid=191343#post191343
You may promote what you want.......
:bigeyes:

fmak · 2004-08-12 6:04 pm

Re: Re: Re: Re: Re: Regulated Powersupply for Clock

These clocks are replaced to enhance the sonics so I judge them for that, not for measurements.
I have found no direct correlation between sound and noise level. I tried two SOT-23 regulators with the same noise spec. One sounded totally uninspiring and boring, the other was extremely good! And the result was much better than with a simple RC filtered follower.

[/B][/QUOTE]

-------------------------------------------------------------------------------
Your posts refer to your system with your clock with no measurements that I know off. Therefore you cannot generalise.

I have stated several times that it is NOT just the clock, but the PS and the digital processors acting as a system. You need to optimise the whole lot to get the best sound, before you start talking about a specific component improving the sonics. In particular, it is sometimes necessary to couple the clock output properly by impedance matching for best wave shape. And please don't just say that this makes no difference. It does.

:bigeyes:

Guido Tent · 2004-08-12 6:06 pm

Re: Re: Re: Re: Re: Regulated Powersupply for Clock

Elso Kwak said:
There is much wrong with the Wenzel circuit. As with all cancellation circuits it is extremely sensitive to misadjustment. Also the cancellation does not work at high frequencies like used in the clock.
It was all explained at this Sympatico site but unfortunately it is down.
http://www3.sympatico.ca/add.automation/spice/ripple.htm

Elso,

I spiced that circuit so I know the ins and outs.

I wrote already that it needs adjustment. If you have analysed the circuit you would understand that it is not extremely sensitive.

It will never work at clock frequencies, but I explained already that I do not need it there. At clock frequencies the Q of the oscillator is what counts.

regards,

fmak · 2004-08-12 6:08 pm

Re: Preregulate

Spot off!
I found that pre-regulation makes soundquality worse.
So I don't use it anymore. This does not apply to PI-filters.
:bigeyes:

[/B][/QUOTE]
----------------------------------------------------------------------------------
Fine, I believe you but do not generalise.

I have found preregulation to be most beneficial technically and sonically, but I won't force it down anyones's throat!

:bigeyes:

Guido Tent · 2004-08-12 6:12 pm

Re: Re: Re: Re: Re: Re: Re: Regulated Powersupply for Clock

Elso Kwak said:
First thread must be:
http://www.diyaudio.com/forums/showthread.php?postid=191343#post191343
You may promote what you want.......

Elso,

This is getting very irritating. Again, in the thread you link above, not half a word is spent from my side on integrated regs in SOT package:

" Hello all,

Nice thread, to start with. Often, in supply discussions, I note that no difference is made between:

- The noise generated by the supply and / or other circuits, entering the circuit

- The interaction of the circuit (current) with the supply (and eventually other circuits)

Though both aspects have things in common, they again should be treated differently as well.

Furthermore, any common mode to differential mode conversion of "noise" currents, may affect the circuit, depending on layout. This could be considered as a 3rd aspect, whereas concept and implementation of any regulator can be spoiled, quite easilly.

Hum is a fine example of that........

all the best

__________________
Guido Tent "

In contrast, the given info in my reply is quite informative, as some among you have already confirmed.

Elso Kwak · 2004-08-12 6:21 pm

Re: Re: Re: Re: Re: Re: Regulated Powersupply for Clock

Guido Tent said:
second post refers to LT1763 SMT which appears to give 10nV/SqrrtHz hence is a good one. Am I allowed to promote that ?

This is getting very irritating. Again, in the thread you link above, not half a word is spent from my side on integrated regs in SOT package:

OK Guido, Sorry for irritating you. I was under the impression you do like the LT1763-5 (SOT-23 package)
:angel:

Elso Kwak · 2004-08-12 6:26 pm

Re: Re: Re: Re: Re: Re: Regulated Powersupply for Clock

Guido Tent said:
Elso,

I spiced that circuit so I know the ins and outs.

I wrote already that it needs adjustment. If you have analysed the circuit you would understand that it is not extremely sensitive.

It will never work at clock frequencies, but I explained already that I do not need it there. At clock frequencies the Q of the oscillator is what counts.

regards,

Interesting post. The guy at the Sympatica site who did Spice simulations and analysed the circuit came to whole different conclusions.
Referring to noise at or near clockfrequencies we differ in opinion.
Can't do much about the Q with available crystals.

Elso Kwak · 2004-08-12 6:44 pm

Re: Re: Re: Re: Re: Re: Re: Re: Regulated Powersupply for Clock

Guido Tent said:
because it is way better than any of the 317 or 431 based rubbish and easy to implement for starting DIY
Guido

The Wenzel circuit contains a LM317........
And the TL431 is in Cuno Snoeren's circuit recommended as DIY-supply for your clock.........
Wenzel Finesse circuit:
http://www.wenzel.com/documents/finesse.html
Cuno's circuit:
http://home.tiscali.nl/~t708955/schem/xoschema.jpg

Guido Tent · 2004-08-12 6:54 pm

Re: Re: Re: Re: Re: Re: Re: Regulated Powersupply for Clock

Elso Kwak said:
OK Guido, Sorry for irritating you. I was under the impression you do like the LT1763-5 (SOT-23 package)

I never listened to that one

The indicated 10nV/SqrrtHz at least is way better that what comes out of most other regs, which is what I confirmed

By the way, if I calculate the noise from the datasheet, I come to 63 nV - way off from what Fred achieved

regards

LClock XO PSU

Member

Member

Member

Banned

Banned

Banned

Member

Member

Banned

Banned

Banned

Banned

Member

Member

Member

Member

Banned

Banned

Banned

Member