New VCO,linearity!

[Tony Clark]

> Talking about VCO's, what is the highest linearity
> percentage error one could accept over a 20Hz to 10KHz
> range? Is 0,4% bad? Talking only about the VCO
> not the expoconverter error introduced.

   Well uncompensated VCOs tend to run in the 10% error range over 20-30 
degree Farenheit.  Obviously this is pretty bad.  At 1% (over a 20-30 
degree differential with a typical compensated VCO design) it's not so 
bad as typically your ambient isn't going to change by that much during 
the course of use.  At maybe 5-10 degree differential we're talking 0.5% 
or less in practical terms.
   In the testing I have done I have noticed that the drift is not linear 
over the whole frequency range.  Someone else may have experienced 
otherwise, but there appears to be a definate "zero" point where there is 
no drift.  Above and below this center frequency the drift changes 
directions.  It really doesn't mean much except for the fact that the 
extreme frequencies have the worst drift!  And depending on the design, 
the "zero" point may not be in the center of the frequency spectrum!

> And how about the highest PPM Temp value for
> the frequency of the VCO?

   This is totally related to the particular design.  If you are talking 
about your stereotypical sawtooth integrator VCO, it's something 
inbetween 3300ppm and 3800ppm.  I've been told that 3500ppm is ideal, but 
it is a little more difficult to obtain Tempco resistors with this 
value.
   I'm sure others can shed more light in this arena as it is quite a 
prevalent topic here at DIY.  :)

   Tony

------------------------------------,----------------------------------
I can't drive (my Moog) 55!         |     The E-Music DIY Archive
------------------------------------|  
Tony Clark -- clark at andrews.edu     | aupe.phys.andrews.edu/diy_archive
http://aupe.phys.andrews.edu/~clark |     Contributions welcomed!
------------------------------------'----------------------------------