From Oakley Sound via analogue Sent Tue, Jul 16th 2019, 12:40
> Since the triangle oscillator ramps up and down with controlled current sources, it doesn't have the problem. It doesn't have the same problem but it does have similar but lesser problem which also causes it to go flat at high frequencies. With the sawtooth oscillator you have both the time taken to discharge the timing cap (which needs to be fast to give us a nice sawtooth shape) and the time taken to detect the requirement to discharge the cap. Since both are of a fixed time it is usually compensated for either by forcing the VCO to run faster by injecting more current into it, or by reducing the amplitude of the sawtooth at higher frequencies. The reset or flyback time for a sawtooth VCO can indeed be significant and especially so with a large value timing cap. Some VCOs require the timing cap to be large if they need to be LFOs as well. eg. the third VCO of the Minimoog. Early series minimoog have poor tracking on the third VCO because of this. In the triangle wave VCO you don't have to rapidly discharge the timing capacitor but you do have to monitor the output waveform so you know when to flip states. This means that the speed of the detection circuit, a comparator, and the slew rate of current source itself will also give rise to a flattening of the VCO response. It also means that the waveform will increase in size as the frequency increases. Thankfully, both of these can again be compensated for and will be smaller than the sawtooth VCO. Triangle VCOs do, however, suffer from mismatch between their rise and fall times. This is not generally a problem at audio frequencies but if the VCO is required to go into very slow LFO mode as well then the mismatches can be significant. Tony www.oakleysound.com