>      I think Steve has probably identified your problem.

I think Steve is on the right track, but not quite on the nail.
If the subject line is correct, the crystal is oscillating in the third
overtone of the marked frequency. If it was a third overtone xtal and
things were going wrong, it would be running at a third of the marked
frequency (ie fundamental).

Crystals are usually operated in fundamental mode up to around 20MHz where
they get too thin to be economical. I can't remember what frequency was
being used in this instance, but it seems unlikely that it would be that
high for a DTMF chip.
When you use a third overtone crystal, you have to put a filter in place to
suppress the fundamental, leaving the overtone the only one with the right
phase to make things happen.
IIRC, the original fault manifested itself when a finger was stuck on the
board.
If the normal trimming capacitors you put around are only 10-30pF (these
are pretty small values of capacitance), shoving 100lbs of skin encased
biological slop onto the board is likely to upset things. ie. A high pass
filter with finger-nails.

>      AFAIK they call a third overtone XTAL an 'XT cut' and a fundamental
>      mode XTAl an 'AT cut' ... anybody know what this is?

Again, almost but not quite. The cut (AT, XT, etc) describes how the
crystal physically moves in it's oscillation mode. Whether it gets
fatter/thinner or twists along one axis, etc. An AT cut is the most common
and has good temperature characteristics. The other common format for low
frequencies (32kHz) is the tuning fork.

>> Crystal is not a third overtone type is it. If so try fundimental type
this
>> may resolve the problem. Or could the value of bypass capacitors be
marginal
>> check device data sheets for correct window of values Or use disamilar
values
>> for bypass capacitors Or use ceramic resanator with inbuilt bypass
capacitors
>> Or add series resistor 500R or so with the crystal.

A couple of other things to check:
I think the layout was described as being good. ie. short tracks, etc.
Does this include a blob of copper ? Is this blob of copper grounded ?
If not, it makes a nice capacitor that you probably don't have on your
schematic.

Are there any fast switching things nearby ? For example, AC series logic
is a killer.

Is the chip designed to use the style of crystal in use ?
I used a Motorola RTC once,  that would power up configured for a 4MHz AT
crystal which would overdrive the 32kHz tuning fork I was using. That sent
the crystal into spasms and I couldn't get any sense out of it until the
clock was configured by software to use the lower power part.

Steve. (a different one)

======================================================
  Very funny Scotty.  Now beam down my clothes.
======================================================
Steve Baldwin                Electronic Product Design
TLA Microsystems Ltd         Microcontroller Specialists
PO Box 15-680                email: steveb@kcbbs.gen.nz
New Lynn, Auckland           ph  +64 9 820-2221
New Zealand                  fax +64 9 820-1929
======================================================