SV: Re: [sdiy] Buchla osc , Tri cores skew!

[Ian Fritz]

At 03:38 PM 8/31/2006, karl dalen wrote:
>One want to have the up ramp time equal to the
>down ramp time over the entire freq range of the VCO.
>
>Large Imbalance in the mirror skews the tri
>terrible therefore you trim it at low freq.

There are two major effects here.  1) If the current switch/drive are not 
perfectly symmetrical then the upramp and downramp rates will be 
different.  To first order, this should give the same skew at all 
frequencies.  2)  If there is leakage at the point where the cap is 
buffered then there will be a skew that gets worse at low 
frequency.  Eventually this will limit the low end operation of any design.

You can (at least in principle) compensate each of these effects 
separately.  The tri vco on my website has a totally symmetrical design, so 
the first effect is very small and didn't need to be compensated.  The 
second effect is compensated by feeding in a small current to buck the 
opamp input current.  This is adjusted to give reasonably good symmetry at 
around .05 Hz or so.

The details of all this depend on the specific design, of course.

   Ian
http://home.earthlink.net/~ijfritz/sy_cir4.htm