ASM-1 LFO symmetry problem

[John Speth]

Some time ago a builder of an ASM-1 described a problem that he had with the LFO
of the ASM-1 at low frequency settings.  I had the same problem and decided to
investigate.  I found that this was due to slight mismatches in the power supply
rails about ground.  You can show that the charging rate of the timing cap  is
proportional to the input voltage of the integrator and if the output of
comparator op-amp in the negative swing is different than the positive swing,
you'll get different charging rates.

So I thought I could use this to my advantage by being able to set the skew of
the output waveforms.  This was easy by simply adding a constant current
adjustable with a pot to the summing node of the integrator.  I verified that
you can change the skew of the triangle output from + saw to - saw this way and
likewise the pulse width of the square output.  The only problem (and it's a
biggie) is that the new skew control also affects the rate of the LFO.  Ideally
you'd like the rate and skew control independent.  Does anybody have any ideas
on how to achieve this?

Also does anybody know which manufacturer of 7915/7815 voltage regulators
produces the most accurate (in terms of output voltage) devices?  I checked my
data sheets for these and found they are allowed to vary between 14.6 and 15.4
V.  I guess they are true "regulators" but not necessarily "voltage references".

John Speth