NOOB CV input / attenuation / summing / buffering question

[Tombola]

Sorry in advance for this question.

I'm putting together this FM radio module. The CV input picks up stations between 1.5v and 3.1v.

I'm using the 3.3v source that powers the FM module on the 'start' pot - the op amp is powered with a normal +-12 split supply.

On the module I want a 'start' tuning pot and a CV input pot (attenuator) just like any standard oscillator.

I'm trying to follow the summing amplifier from Art of Electronics.




The 'start' pot works as I'd expect, until I connect the CV pot (the 38k resistor is to keep the voltage above 1v, to increase the useful sweep of the pot)

My problem(s):

- When the CV input is zero (i.e. the output from a pressure points row with all pots at zero, or just unplugged), changing the CV pot still changes the output voltage - it rises around the middle of the sweep, falling at each end

- There seem to be other interactions between the pots that I don't expect - i.e. the total voltage doesn't seem to get much beyond 3.3v. I'd assumed that 3.3v + 5v from the Pressure points would give 8.3v - not ideal for this implementation, but it's what I expected - I'm just testing this part of the circuit without the radio module connected



What am I doing wrong?

Thank you!

[EATyourGUITAR]

you could do it by adding resistors to ground on the ground lug of your CV pot and another resistor to ground from pin 2 of the opamp. anything 10k or 100k maybe. the correct way and the best way here is to just buffer everything. remove and replace R2 with an envelope follower like a TL071 or half of a TL072. same thing for R4. if you do not use buffers, you have a voltage divider that is adjustable in two places and responds to CV. everything affects everything with the schematic you posted. buffers remove the interaction since you have a buffer for each adjustable voltage divider.

[iopop]

Increase R2 and R4 to 100k. That's how its done no most other CV input circuits Ive seen.

I can not give you any good theory behind why. Sorry. The impedance needs to be higher for the input signals to not bleed over?

[EATyourGUITAR]

This exactly how a passive mixer works. 100k resistors. It changes the impedance. Opamps have low impedance outputs so you could say that the opamps are the best way to add a mixer without changing the impedance of the original design.

[Tombola]

So I got it working nicely like this:

'Start' pot -> follower -> 10k -> inverting summing amp
CV input -> follower -> 10k -> inverting summing amp
inverting summing amp -> inverting amp

That's all four quarters of a TL074 - is there an easier way?

I tried iopop's technique, but it reduced the overall levels a lot, so I'd need to boost them back up, and there was still a bit of bleed.

[daverj]

Increasing all resistors to 100K should eliminate needing to buffer the individual inputs. If you only increase R2 and R4 to 100k, but leave R3 as 10K then the circuit has a dived by 10 gain ratio. Making all three 100K maintains the 1:1 gain ratio. If you want to reduce the 5v CV input down to a 3.3v level, then make R4 a 150K and the other two 100K.

[Tombola]

Thank you all for your advice and patience. This is now working very nicely - and looking again is pretty much the input system on the CGS VCO schematic.

It's things like this that remind me just how poorly a politics/journalism degree has equipped me for designing analogue electronics.



I used 200k for the external input, as I don't really need more than 2.5v from the CV, as it's all offset by 1v.