Injection Locking an Oscillator (was Novel Zeitgeber devive for antique clock)

In PSpice I skewed the starting of the "pump" to be purposefully off sync, then you can watch it walk in. The LM339 version is much easier to see and understand, because the alternate pulses change amplitude so rapidly. ...Jim Thompson

Works even better (faster "lock") if you toss the diode, reverse input to LM339 and connect resistor from its output to the top of tank. (Makes you wonder what it'd do if you switched a cap in and out.)

My boss, when I first went to Motorola in 1962, was Jan Narud (co-father, with Walt Seelbach, of ECL). His PhD thesis was on injection-locked loops... about 1/2" worth ;-)

I'll try and find it. ...Jim Thompson

So you don't know how to post an .ASC schematic file?

You always make it hard to verify your circuits, presumably because they are usually defective.

Best of all, pull-down only, narrow pulse...

...Jim Thompson

As usual, you post an image of a tiny time slice of a simulation.

Post an LT Spice .ASC schematic, something most people here can play with.

My apologies. I guess you are incapable of running a simple .CIR file, something any freshman EE student can do >:-} ...Jim Thompson

Wow Jim , Your splendid Bass Boomer design wins the award for most Op Amps in a circuit.

Regarding I.L.L. vs P.L.L., both can be made with similar capture range, Q and phase noise if input signal goes into a limiter.

I once tried to get good phase sync using a few bits of data sync for burst communication of a 4Mb/s modem using the vertical interlace lines in TV fo r data during the early 80's The data edges were detected with an exclusiv e OR gate with one input slightly delays by the desired 1shot duration and the crystal oscillator data rate. Back then cheap crystals were 50/50ppm. S o I changed the design to an analog PLL using a Sawtooth VCXO waveform and with a low data transition density of 1/1000 a Sampled Sawtooth formed a go od feedback signal for the VCXO

Now a 1 ppm TCXO is cheap and even with only 1 data transition pulse in 100

I miss reading all those old HP journals when they talk about Mesa transistors and 0.1 millimicroseconds

"Any resonator" can be used for an Injection Locked Loop (ILL). ( the oscillator only helps get it started but f-err affects capture range)

The tradeoff is the ratio of signal injected to the stored signal resonating. If the Q is high enough, the resonator will ramp up slowly to saturation and appear as am oscillator or clock and sustain itself for many clock cycles without drift.

The jitter is related to the frequency error and harmonic ratio of injected signal.

I won a subscription to the HP Journal at a science fair, when I was in high school. It was when a lot of conventional and sampling scope and calculator stuff was happening. I thought that Page Mill Road must be a gateway to heaven.

There was also a couple of Tek newsletters, and the G R Experimenter and BSTJ and stuff.

It was still a pretty cool place when I interviewed there in 1987, for a job working on magneto-optic storage that I (fortunately) didn't get due to their hiring freeze. I preferred Watson, despite its being on what was then the wrong coast ;) but since I had a family already, I'd certainly have taken a permanent job at HP over a postdoc at IBM.

A couple of years before, I got a summer job there to do some programming for a multi-head ultrasound system, but couldn't do it on account of being on an F1 visa. (The guy I would have been working for was Waguih Ishak

When I was a kid, I had a subscription to the old Motorola Update--about four times a year, I'd get a box full of databooks, full of a few things I knew a bit about, like transistors and op amps and radio parts, and lots of exotica that I had no idea of: MOS memory drivers, MCMOS, MDTL, MTTL, MECL, power devices, microprocessors, and so on.

At 14 or so, it made me feel (almost) like a real designer. It was sure an encouragement to become one. ("Who'd a' thought, forty year ago, that we'd be here drinking Chateau de Chasselais?") ;) (*)

Cheers

Phil Hobbs

(*) Monty Python, "Four Yorkshiremen skit" (The original was "thirty", not "forty", but oh well.)

I prefer Pinot Gris ;-) Or my new discovery, "Sofia"...

...Jim Thompson

I interviewed at HP, when I first moved to California. The guy looked at my resume and gave me a sour look and said "The first thing you need to do is decide if you're an engineer or a programmer."

It was obvious then that, for HP, I was neither.

Here's an even better performer...

Well, not working at HP wound up being a good thing for both of us, I suppose. (Of course it's possible it was a good thing for HP as well, but I digress.) ;)

Cheers

Phil Hobbs

Sounds like win-win-win.

HP won?