better 4046 PLL

Couldn't you use XOR gates to multiply by 10 and then divide by 4 with a couple of flip-flops?

Regards, Ray

What's the accuracy spec on the 40 kHz?

If it fits within a +- 50 ppm tuning range, I'd tend to multiply up to, say, 8 MHz with a VCXO and divide down again. The 4046's PFD is better than good enough for that. VCXOs are pretty cheap, about $3 in onesies, e.g.

I've used LC VCOs in similar situations. Really dramatically better than a 4046's oscillator.

Cheers

Phil Hobbs

I think there's a 7046 from someone, but (a) I don't know if it's

Do they really need 100kHz that's better than what the "drecky" 4046 can supply? If not, why complain?

Don't know yet! My main contact on this is a software guy.

Yeah, a VCXO would be a great jitter cleanup part. I'll have to see if it would lock.

There are lots of RF type (varicap tuned) commercial VCOs around, anywhere from narrowband to octave range. I could put one on the board layout, as an option to the 4046 VCO.

Thanks

"Drecky" is a Larkin comment only because he thinks I designed it. I didn't. It's some poor copy, done by... I have no clue. The VCO, in particular, is a real kludge. ...Jim Thompson

You just made no sense at all. You agree with me that the 4046 is a pretty bad part. I never said you designed it, and never suspected that you did. It was an old RCA design originally.

We were talking about PLLs, and you start your idiotic squabbling as usual. You want to make trouble because you're a mean-spirted paranoid old git. You're not very good at it.

Hey, Mo won a 1-week all-expenses scholarship to a course at BU (over a thousand applicants, 20 selected from all over the world) and she has a very nice dorm room. You know, the very same dorms that you used to peep into the windows of, when you were there roughly a century ago. The only real compensation of having one's wife out of town is that I can eat serious meat three times a day.

Cirrus has a nice PLL clock chip, but it's OTP and requires a $170 board to program it.

Only goes down to 6MHz on the output, so it would be at least a 2-chip solution.

That looks a lot better than a 4046, and runs to 15 MHz typ. Thanks for the link.

Please convey my congratulations to Mo!

I never "peeped"... I never joined a frat. I just observed what the frats did with their front rooms... and you carried that observation to an all-time-high insane rant.

If you have to wait for your wife to be out of town to "eat serious meat" you must be pussy whipped >:-} (Not that I'm surprised :-)

As for you understanding PLL's... Bwahahahahahahaha! ...Jim Thompson

Under 3 ps RMS jitter, way under 1 ps/degreeC total-board timing drift, locking to an OC3 fiberoptic serial data stream. The Vectron OC3 fiber receivers went out of production halfway through the project, so I had to design my own. That was tricky, keeping the phase shift low over 20 dB of optical power level.

It's a bang-bang adaptive-loop d-flop phase detector in differential Eclips Plus SiGe logic. The major phase noise contributor was vibration, so I designed some tiny wireform springs to isolate the nice massive DIP14 ecl vcxo from the PC board. Sort of little bobblehead dolls like you see in the backs of cars, mechanical resonance around 10 Hz.

There are a few hundred of these boards firing a couple thousand client devices in NIF, picoseconds of precision across acres of facility.

Many of our boards have PLLs, analog and digital. I've posted about some before, so your statement above is stupid, as usual.

ut

us

y 2

d,

The data sheet emphasises "No dead zone of PC2" and "Charge pump output on PC2, whose current is set by an external resistor Rbias" which means that y ou can put together and integrator-based (second order) phase-locked loop w ithout having to provide an op amp to do the integration.

It's been recommended here from time to time over the years.

The VCO frequency may be tolerably linear function of input voltage - over the input voltage range for which it works, but it stops working at low inp ut voltages, and only offers a roughly 3:1 frequency range.This shouldn't b e a problem in your application.

Digi-Key Australia stocks it (19,901 in in stock), for $A3.18 in small quan tities, which is comforting - IIRR Farnell gave up on it years ago.

So why didn't you know of the modern HC/HCT versions?

And you're still pussy-whipped >:-}

And a bird-brain. ...Jim Thompson

PC2, whose current is set by an external resistor Rbias" which means that you can put together and integrator-based (second order) phase-locked loop without having to provide an op amp to do the integration.

input voltage range for which it works, but it stops working at low input voltages, and only offers a roughly 3:1 frequency range.This shouldn't be a problem in your application.

quantities, which is comforting - IIRR Farnell gave up on it years ago.

A resistor to ground from the PC2 output of a 4046 accomplishes the same thing, with a part that costs $0.17 in qty 25 (Arrow).

The metal-gate 4046 is quite pretty in a 1970ish way--the oscillator can cover way over 100:1 in frequency. Later models, starting with the HC4046, are OK as phase detectors if you remember the resistor, but the oscillators really stink. We went round that mulberry bush in fairly gruesome detail about 8 months ago in this very boutique.

Cheers

Phil Hobbs

Because it's a discussion group, where people help one another out. You clearly don't get the concept. All you do here is boast and call other people stupid. And spin technical discussions into bitch-fests.

No complaints. I can't imagine a better woman to do it to me.

I'm reminded of a frat party I went to once (as a guest, not a member.) A guy had a tee shirt with two big letters PW on the front. The beautiful girl next to him had a shirt with one big letter, P.

Seven, nine, eh, what's the difference? It's an odd number, and bigger than "many".

(Thanks for taking my imperfect recollection and turning it into something accurate).

It looks like DigiKey has oodles.

It's 2.236 times better than a 4046.

That's actually funny... so rare for you ;-) ...Jim Thompson