LM317 as power supply for kwak-clock

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Not easy to make it right!

In your private email to me you have attached a sketch over the the "Kwak"-clock. It's for replacement of the original oscillator in a CD-player.

I don't want to dissapoint you but there are a couple of things you must know:

It's VERY difficult to make a low litter oscillator which is substansially better than the original. If you make an stand-alone oscillator, you get probely more problems with RF-emission (= worse sound) than you get improvements. The Kwak-design isn't any special at all!

The oscillator subject is very wide and hard to get close (if you don't are satisfied with 1 crystal, 2 caps, 1 inverter). It's very important to have stable power supply, constant temperature, well controlled drive level (power) of the crystal, not to tune the crystal. Mnay things to condsider. The Kwak-design is tuned with a cap in series and that's a death sin!

If you really want an improvement you can seek information at many crystal and oscillator manufactures.
 
KWAK-CLOCK

Hi Peranders,
The KWAK-CLOCK is a dramatic improvement over the CMOS oscillator circuit used by 99 % of CDplayer manufacturers.
I received dozens of enthusiastic reactions from DIY enthusiasts from all over the world.
I also had the opportunity to compare my clock with the LCaudio clock. The KWAK-CLOCK outperformed the LCaudio!
I think you have absolutely no clue what makes a good <B>sounding</B> clock!
Powersupply makes a difference as I tried many different schemes: LT1086-5; BC550C/TL431 plus PI-filter; Jung regulator with LT1021-5 reference and OPA27 opamp etc. It makes a difference in sound which powersupply you use but it is not a big effect. The cap in series with the crystal can be omitted but the sound is better with the cap.
For your information I got much inspiration from crystal oscillator manufacturers and also from a former employee of a crystal manufacturer.
Peranders, why don't you build the KWAK-CLOCK and judge the <B>SOUND</B>?! It will not cost you a arm and a leg, which can not be said of low jitter cantype oscillators.

Jayel, I believe you can use the LM317 with bypassed adjustment pin successfully.

;):(
 
Hi Elso. Is it possible to send to me an schematic of that BC550C/TL431 plus PI-filter power supply?

Also in my power supply, I connected a 220uF electrolytic from Vout to ground just like in the 1st page of the datasheet of LM317. But I also soldered a metallized cap from Vout to ground as most power supplies I've seen have this. Is this a good idea?

Thanks

jayel
 
LM317

The LM317 has pretty high "overhead" requirements. (Discomoded as I am with a busted ankle, I don't have my Nat Semic manuals at the computersite.) but I think that you need at least 1.5 volts of "overhead". Further, the LM317 is gonna need to see a lot of several m.a. to operate correctly -- so you have to burn watts (or milliwatts) to have it optimal. You might want to pull the Vin from a higher source. You can, of course, use a low dropout regulator. These have advantages and disadvantages compared to the LM3XX series -- both Nat Semi and Texas Instruments discuss the pros and cons on their websites.
 
Nice guy

Elso, I hope you got not too offended. I'm a nice guy normally.

I must admit that I haven't done any "crystal tweaking" but I think that you haven't been able to test two CD-players of exactly the same model, with and without oscillator trimming. It's very hard to remember different sounds if the difference is very small. The ear tend to adjust.

The difference between my DENON DCD-1520 with B-B DAC and 5532 output and my CS4328 separate DAC is VERY, VERY small for most sounds but for some particular sounds you can actually hear a small, small difference. But what is the cause of that?

Don't you agree that it's wise to tune the crystal as little as possible. The Q-factor (jitter, stabiltiy) gets rapidly worse with tuning.

I agree that a simple inverter oscillator is rather "slapp" as we say in swedish (not very exciting!)

You can't deny that it can be a problem with a stand-alone oscillator above the main pcb can be a problem. 16 MHz with fast flanks radiates and the signal can go to places where it shouldn't be. Ground and power supply for this extra pcb can be critical.

The best oscillators are those for series crystals (not parallel) and I have a schematic somewhere. I have made an oscillator with series resonance and at that time I notice it was VERY unsuitable to use a parallel crystal in a series oscillator and then tuning the frequency to the parallel value. Don't do it! The Q-factor is reduced to nothing!
 
I have made an oscillator with series resonance and at that time I notice it was VERY unsuitable to use a parallel crystal in a series oscillator and then tuning the frequency to the parallel value.

That is why when you design the oscillator, you specify either a series or parallel mode of operation to the crystal manufacturer.

You can still pull them enough to tune it without screwing things up.

Jocko
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.