SPICE is inadequate for this job because the new crystals have much larger nonlinear sensitivities. You need a general purpose model that starts from the material anisotropic stress/strain elasticity tensor and its dependence on a host of other factors such as drive level as basis and goes from there. The usual applications engineering linear time invariant model with fixed components is a joke.
And it's a cost issue, 70MHz is MUCH cheaper than 280MHz.
...Jim Thompson
Mesa etching.
John
Keerect !-)
The fab has now agreed to get a crystal model from the manufacturer.
As is usual with most of my projects, I come on-board when the s*it has already hit the fan... the original design team has floundered ;-)
So the fab has to keep "tight cheeks" while they request data they should have gotten long ago ;-)
...Jim Thompson
I guess that you'll tell me I never used 125 MHz fundemental crystals in TO-5 cans? Microdyne had them custom made for a long time, and I left there almost five years ago.
Sounds like great fun. I love to swagger in, in a modest sort of way, after somebody else has flubbed it. Especially if the somebody else is their in-house crew. You get to say out loud, in a kindergarten-teacher voice, the things they were afraid to mumble. The real trick is to wind up friends with the guys you're trashing.
No time left and big-budget consequences are icing on the cake.
John
Only got to fly JetBlue once, from Long Beach to Boston. Got stuck in traffic on the 405, and didn't get to the airport till 5 minutes after the flight was scheduled to leave... it was still there! Seems about
2/3 of the flight was in that traffic jam, so they held the flight...Charlie
[snip]
Yep, I love it!
"The real trick is to wind up friends with the guys you're trashing."
So true.
...Jim Thompson
Fred, do folk ever tell you that you have a rather strange sense of humor? I didn't get it the first time I read it, but now I do. ;-) Thanks.
Cheers, Tom
Hello Jim,
Then they should be the experts on this and be able to furnish all the model parameters. After all, they probably want the best oscillator in town and that first silicon is a success.
Regards, Joerg
"First silicon" WASN'T a success. That's why I've been brought in... fix the screw-ups ;-)
...Jim Thompson
"Welcome to the project -- you're late and over budget"?
Hello Jim,
That's the usual job of consultants. Same here, most of the time. Was the oscillator section a part of the wreck?
Regards, Joerg
Naaah! I'm the big-bucks hired gun ;-)
...Jim Thompson
Not necessarily. There have been 70 MHz fundamental tourmaline crystals for decades. $$$ though.
Hello Jim,
In Asia that is even more touchy. You see a messed up design, something that would never have worked and never did. Then you have to do the dance of extreme politeness. "Well, there are some wonderful ideas in this great design. Here is how we can add to those ideas to achieve an improvement". Meaning in reality most of it had to be scrapped and started from a blank sheet.
Regards, Joerg
I have one client who, for some reason or other, is fond of Chinese PhD's for chief engineer.
In my presence three of them have decided they didn't want to live in this particular east coast town ;-)
...Jim Thompson
In article , Jim Thompson wrote: [...]
I have some numbers. They may help. I assume you are working with an AT cut crystal.
When running: C0=7p R1=50 R0=6000 C1=1.2fF Q=33K
At startup, you should assume that R1 is higher. If your circuit will work with R1 = 150, chances are it will work with any reasonable crystal.
Okay- well that might have been a stretch, but JT does need to watch for the traps, like exciting various crystal vibration modes close to the
4th harmonic but not the 4th harmonic, or some tendency to chaotically switch into a different mode. The previous people cannot be complete incompetents and these kinds of issues are widely none thanks to authors like Parzen, so there must be something up with this particular crystal cut that makes a reliable design particularly challenging. If he doesn't get into the crystal physics on this one, his product will be no more than a hunch.
sense of
Okay- well that might have been a stretch, but JT does need to watch for the traps, like exciting various crystal vibration modes close to the
4th harmonic but not the 4th harmonic, or some tendency to chaoticallyswitch into a different mode. The previous people cannot be complete incompetents and these kinds of issues are widely none thanks to authors like Parzen, so there must be something up with this particular crystal cut that makes a reliable design particularly challenging. If he doesn't get into the crystal physics on this one, his product will be no more than a hunch.
Have something to add? Share your thoughts — no account required.
Ask the community — no account required