My partner and I (he does software, I'm hardware) have been developing a guitar tuner based on ESP32. Yesterday he sent me this plot from excel of data he captured at 3 diffferent frequencies. Y axis is digital 12-bit (4096). The voltage range is software set, it's somewhere between 1.0 and 3.1v, not sure yet. Wherever it is, the lower part appears to get clipped at about 1/2 the range. This behavior could be the phase inversion from exceeding the common mode input voltage on a TL072 that @Audioguru again hinted at.



Here's another plot, barely clipping at the top of 4096.

 

An LM358 dual opamp works with a supply as low as 3V and has inputs that work at very low positive voltages. but its input bias current is 6 times too high for your 2.2M and 1M input bias resistors that should be no higher than 360k and 160k.

 

Ok good to know. So 5v source and maybe 100k/220k voltage divider or so.

 

An LM558 dual opamp works with a supply as low as 3V...

Is that a typo? LM558? Or LM358. I can't find a 558.

 

Yes, I had a typo error, use an LM358 dual opamp.

 

I'd like to band-pass filter it before it gets to the esp32.
Bass guitar goes down to 40, guitar goes up to mid-300's, mandolin goes up to 660 Hz.
Below is my attempt:
LPF on the right, C1, R4 give a crossover of 589 Hz
HPF on the left, C2, R3 give a crossover of 34 Hz.
Gain may change, but in this configuration it's 47/3+1 = 16.7
Generic op amp in Eagle, I need to do this in spice with a real LM358.

 

Pretty similar results in Spice:

 

I was looking at this and realized that C3 and R2 also create a HPF. So I got rid of C2 and R3 (see post #27) thinking I didn't need it (and changed C3 to give me the right crossover. But the output went WAY down. Lost like 100 db. What gives?

 

Hi rp,
On your 2nd Sim check the DC level on the +NI input of the OPA and the DC output voltage of 358, without any AC input
E

It helps if you always post your LTS asc file

 

On your 2nd Sim check the DC level on the +NI input of the OPA and the DC output voltage of 358, without any AC input
It helps if you always post your LTS asc file

OK you might have to help me on this. I ran a DC operating point analysis. Doesn't matter whether AC = 1 or 0, results at v1, v2, v3 are the same:
v1 = 1.48vdc
v2 = 4vdc
v3 = 0.24 vdc


Here's my asc file.

 

Hi rp,
That's not the asc file for the circuit you posted in post #30???

E

 

I think I need C4, I don't remember why. It does something important to do with AC or DC voltage. And I don't understand the value of C2. I understand all the other values... but not C2 and C4.

 

Doesn't matter whether AC = 1 or 0, results at v1, v2, v3 are the same:

The AC voltage value is used only for the AC analysis and has no effect on the DC or transient voltage values.

I don't understand the value of C2. I understand all the other values... but not C2 and C4.

C2 and C4 are there because the amp is powered from a single supply and thus is biased above ground to handle ±AC signals.

C2 blocks the DC voltage at the amp input so the input doesn't affect the bias voltage.
Its value needs to be large enough to not attenuate the lowest frequency signal.

C4 blocks the DC output voltage so the output bias point is not affected, and also needs to be large enough to not attenuate the lowest signal frequency.

 

Should C4 be electrolytic or film? 1u is kind in the middle...

 

Should C4 be electrolytic or film? 1u is kind in the middle...

Either is fine.

 

R5 and C1 are the LPF. But what components are acting like a HPF?

 

what components are acting like a HPF?

C2 with the input impedance (R1||R2), and C4 with R4.

Look at the AC response at the output of C2 and across C4.

 

Hello,

The attached pdf will show you a lot of single supply opamp circuits.

Bertus

 

Wow that is an excellent resource. I will study it.

 

Here's my finished circuit. Probably send it off to print soon.