Just a simple natural question
(I have not been following your problem...)
If you once erased the chips, did you restore the factory calibration?

To test the theory:
Have you measured the actual frequency, and compared to what the formula give if you read the constant out of the chip?
(is it the calibration, chip, formula or program routine that is wrong?)

Check to see how voltage variation affect the freq.
How voltage dependat is the calibration value, btw?

Maybe you have to measure the supply voltage (using internal ref. a resistor divider, and internal A/D) and add that using a lookup table for correction multiplier or something.

Regards
/Morgan

Hej William M. Smithers. Tack fšr ditt meddelande 14:38 1999-08-18 -0400 enligt nedan:
>I figured I'd take one more whack at getting the
>onboard oscillator to work on the 14000 before I
>give up and go external.
>
>The problem appears to be in the interpretation of
>the on-board calibration constant OSCM.
>
>I'm using this following the formula given in the
>app notes to determine the measured clock speed,
>and I've tried accessing OSCM both directly and
>via the call provided as part of the HiTech package.
>
>It obviously doesn't work, because no matter
>what I do, the serial bit-banging timing is
>screwed up.  Works fine with an external oscillator.
>
>Anybody have any ideas?
>
>-Will

Morgans Reglerteknik, HŠllekŒs, 277 35 KIVIK
  tel 0414-446620, fax 70331, mrt@iname.com