Op Amp Question

[breakshift]

I built the circuit attached. On the left is a buffered voltage divider to give a 2.5V DC supply. On the right is a non-inverting amplifier with a pot for adjustable gain. It works as intended with gain increasing with frequency - 3dB cut-off at about 10kHz.

When I first built it the amplifier was unstable - oscillating at about 250kHz I think. However it only oscillates if the input capacitor C5 is in-circuit but with Vin left open. It becomes stable if C5 is removed entirely and also if Vin is tied to ground, to a sig-gen or any source really. So why is the circuit unstable with C5 left open ended? Why does it become stable when C5 is removed? I'd have thought that C5, if left open circuit on Vin, would have no impact on the circuit since it's essentially an open circuit.

 

[Nigel Goodwin]

Is this in real life?, or just in a simulation?.

But in either case there's a lot wrong with the design, for a start you don't generally use 100pF for coupling capacitors, and it's a poor way to try and limit frequency response.

What is the source feeding it?, and what is the load? - for that matter what is it's intended purpose.

 

[breakshift]

Real life - built on stripboard. It's intended as a basic amplifier, to amplify the output of a microphone. I'm only interested in frequencies from around 20 to 100kHz (the mic is still sensitive in this region) so my choice of capacitors was an attempt to reduce the gain at lower frequencies. Why is the 100pF a bad choice? Would you set C5 higher and leave the freq response shaping to the feedback capacitor C6?

The source is the mic, which is about 400R output impedance. The load is an SA612 mixer, which is about 1.5k input impedance. I've not connected either to it yet though - I'm just testing it with no load.

 

[crutschow]

........................................
When I first built it the amplifier was unstable - oscillating at about 250kHz I think. However it only oscillates if the input capacitor C5 is in-circuit but with Vin left open. It becomes stable if C5 is removed entirely and also if Vin is tied to ground, to a sig-gen or any source really. So why is the circuit unstable with C5 left open ended? Why does it become stable when C5 is removed? I'd have thought that C5, if left open circuit on Vin, would have no impact on the circuit since it's essentially an open circuit.

It's likely stray capacitance from C5 to ground or to the output that causes the oscillation. Connecting the capacitor to a source or ground shorts out the stray and prevents the oscillation.

 

[breakshift]

Right. So I gather the circuit could use improvements, since it's on the edge of instability. How would I improve the circuit to make it more stable?

 

[ronv]

It's ok. It just needs an input. Make sure that you keep the output wires way from the input wires.
You need much larger caps to get down to 20Hz. 1Ufd on the input, 10 Ufd. for the gain and 4.7 Ufd for the output.

 

[breakshift]

I'm only interested in 20kHz to 100kHz so that's why the caps are smaller than usual - I've attempted to limit the frequency response to those higher frequencies.

Would it be better to put a filter on the output of the amplifier? I could get a sharper roll-off doing it this way, but it's not crucial so I thought I'd combine the filtering into the amplifier circuit.