Superclock in SL-1210

[MartinT]

Thanks - that's good, if sobering, information to have.

[sq225917]

See if Guido Tent has any suggestions he seems to be able to source a variety of do an tcxo's for his own tentlabs stuff.

[Paul Hynes]

I will bear that in mind when sourcing my own crystals.

[Marco]

Few companies make their own crystals, these days. So, when you approach someone about making a custom part, they usually are not interested. Their suppliers only want to make parts that are cheap and common. Trying to get someone to make a small handful, at an oddball frequency, is a pain.

And that's exactly what pisses me off about audio these days: it's all about quantity over quality, driven by the bottom line!

Is it therefore any wonder why many of us build our systems around quality vintage equipment, designed in an era when overall standards were much higher? Whatever happened to striving for excellence, genuine innovation, and creating something that is truly unique and special??

To my huge disdain, we've largely become an undiscerning nation of mindless sheep, who are easily dictated to and satisfied with blandness and mediocrity. And they call this progress!!!

Marco.

[Reid Malenfant]

Ok guys, I have just had a thought

If getting an ovenized cystal is going to be problematic it might be worth considering simply keeping the standard crystal at precisely the right temperature to produce exactly the right frequency.

It wouldn't be too difficult to use a peltier cooler to maintain an exact temperature. They work both ways, with DC current fed one way one side heats up & the other cools, feed DC in the opposite polarity & the side that heated up previously cools & vice versa.

Shouldn't be difficult to make a control circuit to do the job

[MartinT]

My thinking says that the exact frequency is not so important (certainly not the highest priority as it is for radio operators), it's the stability and jitter that are critical. Will any kind of temperature controlled enviroment improve that, or is it down to how optimally the flip-flop circuit uses the crystal? This I think is where the Superclock comes in.

[ar-t]

Ok, the first thing you guys need to decide is what exactly it is that you want to accomplish, and why.

I know..........."I want it to sound better!"

Right. And what parameter(s) make it "better". Stability? Sort-term, or long-term? Phase noise?

Once you know the answer to those, and why, then you can think about making it "better".

If "better" is long-term stability, then what you have may be good enough. Short-term..........??? Phase noise.......I can promise you it isn't. (And I don't even know what you have, what it does, and how it is made.)

OK......someone said jitter, etc. Yes, if you divide down, the jitter gets better. But, the higher the reference, the harder it is to get a good crystal. And therefore, low jitter.

If whatever this thing is already has a divider chain, then it probably will not get any better. Eventually you hit the real-world noise floor, and that is as low as you can go. Divide any clock enough times, and that is what you get.

[sq225917]

There is a divider chain on both the 1210 and the Hercules board I'll be using. To be honest I'm more interested in psu quality and splitting each stage than I am about the quality of the do reference. Crystal effects will be swamped by motor cogging in the 1210 and by cogging and belt incOnsistencies in my application.

It can't hurt though, can it? I think most of us just like fiddling with stuff. I actually give away far more stuff than I build for myself.

[MartinT]

What I want is not long term stability (which will have precious little effect on sound quality), I want jitter reduction to give the cleanest clock possible. Some of that will come with the phase 3 regs I'm about to install, to give the AN6680 control IC a cleaner supply, and some of it could come from a Superclock circuit.

In the case of the Technics, jitter reduction could help reduce rotational noise amplified and fed into the motor by the AN6675 motor drive IC.

[Pete]

I think knowing what the time constant of the speed control loop is, in relation the the frequency used in the digital circuit would help to understand if jitter is a significant factor.
If for instance the crystal oscillates at 4 Mhz, but the control loop operates at 100hz, the jitter noise may not even be discernible to the control loop.

I'm a mech engineer, not an EE, but that is what comes to mind.