Simple Non-inverting amplifier makes a weird sawtooth-like buzz...

[guyshermannz]

Hi all,

I'm new here, and I'm pretty new to playing around with stompboxes. Anyway, I have a problem with a very simple circuit.

So I started off following this tutorial from Brian Wampler

And that was all fine, it behaved as advertised. I then decided to go a step simpler and just play with the basics, a buffer, a simple amplifier, an active high pass filter and an active low pass filter.

Doing a buffer works fine. Then it comes a bit unstuck.

The problem I'm having is that if I go with a really basic amplifier (fixed gain) circuit, where the inverting input on the opamp is tapping off a voltage divider between the output and ground, I get a really weird sawtooth-like buzzing, and nothing resembling the input signal whatsoever.

Here's a jack-scope capture of the signal:



If I put a cap just before ground on the path from the non-inverting input, I get largely what I would expect, but this is also a highpass filter, and I'd kinda like to avoid the bass roll-off.

Is this one of those times where theory and practice differ, and you just have to have the cap in there?

Here is the circuit that doesn't work:



And here is the circuit that does work:

[11-90-an]

Welcome to the forum...

Are you only using a simulator? Or are you also breadboarding also...?
Simulating can sometimes have errors...

If you want to reduce the bass rolloff, you can increase the capacitor value... maybe to 2.2uF or more.

[guyshermannz]

Hi, I'm breadboarding as well, that scope capture is from running pink noise out of my audio interface through the circuit and back in.

Oddly the simulator has no issue without the capacitor.

[Steben]

The cap decouples DC. You can remove the cap by connecting the point to 1/2 bias between the the two 10k resistors (4,5V).

[ElectricDruid]

The cap decouples DC. You can remove the cap by connecting the point to 1/2 bias between the the two 10k resistors (4,5V).

+1 agree. The problem is to do with the DC bias of the circuit. The op-amp expects both its inputs to be biased at the same level, so having one connected to the midpoint level and the other tied to ground doesn't work. Putting the cap in blocks that DC connection to ground and allows the circuit to work.
The solution is either to connect the bottom of the feedback network to the 4.5V bias, like Steben suggested, or to use a larger cap like Nathan suggested. 10uF should put the cutoff down at 15Hz, which you'll never hear. Alternatively, you could use 20K/10K resistors for the feedback network, in which case you'd only need a 1uF cap to get the same 15Hz cutoff.

[PRR]

> Is this one of those times where theory and practice differ,

No. You have to get your DC right before you think about audio. 4.5V bias, DC gain of 3, the output wants to go to 13.5V as ordered. But it can't even reach 9V. It is bollixed before it starts.

A lot of simplified lessons ass-ume a split +/- supply, and zero V DC on most signal nodes. But in battery pedals we often run a single supply and 4.5V DC bias. Keep that in mind.

[tonyharker]

You need a capacitor of about 25uF in series with R5 to isolate the negative input. Put it between R5 and negative.

[guyshermannz]

Thanks for the answers everyone, this is very helpful.

So, to make sure I'm understanding correctly: because I'm biasing the non-inverting input around 4.5V, but I'm not biasing the inverting input to the same voltage, I effectively get a much bigger difference between the two inputs (assuming 0V coming from my guitar) than I was expecting, so the opamp is trying to go wide open (and hitting its limit wrt the supply). In the circuit that works, the capacitor is decoupling the input from ground, but I could solve the problem without the capacitor by biasing the inverting input. Is that correct?

[antonis]

because I'm biasing the non-inverting input around 4.5V, but I'm not biasing the inverting input to the same voltage, I effectively get a much bigger difference between the two inputs (assuming 0V coming from my guitar) than I was expecting, so the opamp is trying to go wide open (and hitting its limit wrt the supply). In the circuit that works, the capacitor is decoupling the input from ground, but I could solve the problem without the capacitor by biasing the inverting input. Is that correct?

Νο.. 

As Paul showed you above, singnal needs to "sit" on a reference voltage (+4.5V here) so it can swing symmetrically up and down (from +4.5v to +9V and to 0V).
For the particular op-amp configuration, +4.5V are set on non-inverting input and op-amp tries to make both inputs voltage equal..
In case of R6 been a short, output should stand on +4.5V 'cause there shouldn't be any voltage drop across R6 hence output voltage should be the same with inverting input one..
With the presence of R6 & R5, there is a voltage divider on inverting input of Vout = R5 / (R5+R6) so for inverting input voltage to stay at +4.5V, Vout should be 13.5V..
You can see than we have a DC gain of 3 here..!! 
By placing a capacitor between R5 and GND, you block DC flowing to GND so practically set R5 out of the NFB loop..
(R5 lower leg isn't connected anywhere, from DC point of view..)
The above mentioned voltage divider doesn't exist anymore, so we result into a unity DC gain..
(Vout DC is +4.5V..)
That leaves a margin of +/- 4.5V (ideally) for signal undistorted voltage swing.. 

[Steben]

Electrolytic caps are not very strong components on the long term. And those are what you need if you want high capacitance at a fair price. Thats why it is interesting to avoid using them.