# Bandwidth of triangular wave for desired linearity

I'm attempting to replicate the Wavetek Model 188 Function Generator with original circuitry. This function generator has a 'VCG (Voltage Controlled Generator)' input which allows the user to apply external frequency modulation. Here's a description of the input from the manual:

If the main frequency control dial is set to a minimum, then a +4 V signal at the VCG input will set the frequency to the top end of the range. Conversely, if the main frequency control dial is set to a maximum, then a -4 V signal at the VCG input will set the frequency to the bottom end of the range. So, the maximum excursion of the signal at the VCG input is 4 V (-2 V to +2 V, 0 V to +4 V, etc.)

From what I understand, the slew rate specification gives the maximum rate of frequency modulation as 2% of the range per microsecond, which translates to a 10 kHz triangle wave (if I'm not mistaken?)

Given that the maximum slew rate of the input translates to a 4 Vpp 10 kHz triangle wave, let's say that I have a transconductance amplifier that converts this voltage to a current and that the generator frequency is directly proportional to this current.

The question is: what bandwidth is required of the transconductance amplifier to achieve the required +/- 0.5% linearity if all other causes of non-linearity are neglected? If there is no simple formula to associate bandwidth with linearity, can anyone provide a decent rule of thumb for this situation?

• If I recall, that generator uses the RCA CA 3102 diffpair, which has 1GHz transistors. There is no fundamental limit to how fast the sweep signal can change. – analogsystemsrf Apr 27 at 13:24