[sdiy] Sawtooth vs. Triangle core VCOs

[Ian Fritz]

At 12:59 AM 1/25/2013, Neil Johnson wrote:

>Because there is no reset time a suitably-designed triangle ....

There is still a switching circuit and a finite time required to reverse 
the ramp, no?

I get a reset/switching time of ~200ns for either Tri or Saw.  This 
corresponds to1.7 cents at 5 kHz. This is literally in the noise.

Also it must be pointed out that an accurate Tri waveform require that the 
up- and down-ramp rates be the same, and this can be a limiting factor in 
high-frequency performance.  For my Tri VCO I had to use a fully-symmetric 
discrete four-transistor current mirror with a pair of super-transistors to 
get adequate symmetry.

You didn't mention the charge-pump VCO.  These have no dead time because 
the integrator continues to integrate during the Saw reset.

---

I agree with all the comments about waveshapers.  If highly accurate 
secondary waveshapes are required, then starting with a Tri core might be 
preferable..

Ian