> I'm all for making pics do as much as possible, but you'll have a tough
> time making a cost-effective replacement for a <$2 part like this.


I can dedicate a senior engineer to solving this problem for a year if it lowers
the parts cost of my product by $1. The trick of course is weighing the
 difference
between buying a beefier PIC to solve the problem versus using the dedicated
chip.

I was talking to a telecom guy about this problem and he claims that 90% of
the signaling gear he encounters, generates the DTMF with a syntesizer chip.
 This
means that the two tones are essentially equal in amplitude and their phase
relationship is constant. You can't design telco gear assuming this, but you
also can't design RS-232 interfaces with just a diode :-)

--Chuck