DIYHiFi.org

This one?
Do not use bigger than 220uF for C16 on drives other than Philips.
I will try Jocko's suppy on it.

How 'bout something simple like a 1-transistor Pierce?



This is the one place I would use a JFET instead of a BJT for a oscillator. C2 is variable, but I could not come up with a quick way of drawing one. L1 is just a RF choke of some kind.

Jocko

Apparently I had a senior citizin moment, as I uploaded the wrong scheme.
Here is the single supply one:

Hi, would the crystal not be over-driven?

That basic Pierce idea is older than either of us. I don't think it would have survived if it cracked any crystals.

Jocko

Interesting comparison, guys. I remember Jocko's Pierce circuit, or something a lot like it, from my brief forays into ham radio in my yooth.

Elso, how do your circuit's waveforms look without the comparator?

Mine is a variation of a Colpitts, where you take the signal through the crystal. Normally, the crystal goes to ground, and you pick the signal form the emitter. Requires a low impedance in series with the crystal. Upsets loaded Q a bit, but filters out noise sidebands. And gives me yet another reason to use a common-base stage.

Can't teach an old ham new tricks.

Jocko

I meant too much gain, bad for phase noise.

Somewhat like a sine-wave.
Does it matter?

The best publication on this topic I have seen so far is
Oliver Collins: "The Design of Low Jitter Hard Limiters"
IEEE transactions on communications, Vol. 44, No 5, may 1996

Unluckily, it's in the IEEE vault of wisdom.
I stumbled across it on the internet, nevertheless.

He shows rigorosly, that you need a multistage approach
with different bandwidth for each stage with 4 or 5 stages
typically being the the point of deminuishing return.

Sadly, it's probably not OK to offer help to stumble.

regards, Gerhard

_________________
Everything has been said already - but not yet by everyone. (Karl Valentin)

http://energiainfinita.110mb.com/00494304.pdf

Regards

What I meant was: Have you tried using it with a real-world load without the comparator? How does your FET crystal driver stack up against Jocko's? You could argue that they'd both benefit from a high-speed comparator on the tail, but let's see how they do without (it might be reasonable to answer "exactly the same, depends on the crystal and the power supply" but let's find out first...)

I'm on vacation in a place without components, but next week I'll throw them both on the protoboard and see, and then give them both a try with Jocko's regulator when it arrives...

For once, this is something that uses on-hand junk box components (since I have one of Guido Tent's crystals somewhere), so worth a taste test.

I just sent the 2 domestic orders. (No, Gordon.....yours isn't ready yet.)

Anyway, I tried to use the self-serve kiosk, as tonight is the first time I have been able to leave. I tried to mail them as large envelop, but the postage was ony $0.83, and it would not let me mail anything less than $1. No, could not get delivery confirmation, or any other reasonable add-on. Nor would it let me stack orders, to get meet their minimum. What a lot of crap.

But.........if I mail it as a parcel, and not as an envelop, it costs $1.17. Which meets their $1, and I could proceed from there.

I hate it when that happens.

SO, looks like no matter what I send, to whom, and where, it costs a buck-something to send.

Jocko

I will be gone all weekend on family matters. Maybe Monday, too. Probably next weekend as well.

So, you probably won't see me here, and my mail will undoubtedly stack up in the meantime. It will get answered upon my return, but it may take a while.

So don't e-mail me, and I won't have as much crap to wade through on Tuesday. Have mercy on a miserable old sod.

Jocko

I have tried the FET Colpitts without any add on, with 74(V)HCU04 logic and with comparators. The latter were slightly better.

Guido does not sell crystals but complete oscillators in a square metallic box!