Have Questions about Crystal Parameters

[X-Rated]

Feedback welcome on spreadsheet I made.

http://www56.zippyshare.com/v/4268579/file.html

It is a work in progress but mostly done.

[N8YX]

I've got a non-parameter-related one:

A design which I'm looking at improving uses a 2.560MHz crystal in an HC6/U can as part of a timebase reference oscillator. Needless to say, this sucker drifts...the rig which incorporates the circuit has always been known for it. Say, 1.6KHz from a cold start.

So...I have a few options:

1) Temperature compensation of the existing crystal
2) Design a separate TCXO operating at 10.240MHz then divide it by 4 to reach the initial frequency
3) Find an HC-18/U version of the 2.560MHz crystal and temp-compensate that
4) Incorporate a circuit which applies +5v to the existing oscillator at all times, thereby keeping it in a stable thermal regime
5) Chuck the crystal oscillator altogether and replace it with an AD9851 or similar DDS, controlled by a PIC
6) ???

Give me your thoughts, Jerry.

[X-Rated]

I guess we need to know a couple of things.

1. Why does it drift? I am guessing the crystal gets hot and the frequency drifts.
2. Is that an AT-cut crystal? If you have an AT-cut crystal, the frequency shouldn't deviate by more than 20ppm or so. (230Hz max offset or so at 10.24MHz)
3. How close do you need the frequency to remain?

Temperature compensating your own TCXO is a bitch. Schemes I have seen utilize thermistors, varactors, and resistors to trim the frequency back on. Not saying it can't be done, but alignment is difficult.

AT-cut crystals have a temperature curve that bottoms out at some frequency at ~40C to 75C. If you expect the radio to be at some temperature, you can get a crystal with the right cut for your temperature needs. The frequency would drift down at warmup and settle down to be pretty steady after it has warmed. If you have a 40 degree environment, it will not drift more than 50 Hz at 10.24MHz and remain there pretty steady.

I have seen some 10.24 MHz OCXO's on ebay. They will drift badly for 5 minutes or so and then be steady to well under 1Hz.

So, am I in the ballpark?