[DEBUG HELP] violin piezo pickup preamp

[blana]

Hello everyone,

I just built a circuit that is supposed to be a piezo pickup preamp, the only problem is that is does not work.

The schematic is one of my own design and i did it after some research on the internet. Here it is:

https://www.dropbox.com/s/fs2qirefcgu2guo/piezo_preamp1.png

I did some simulations with LT spice and at least in theory it does what i need it to do: buffer, amplify and adjust the tone of a violin piezo pickup. I tried to keep it as simple as possible, using a small amount of parts.

I'll try to explain each block's function:

C1 is a separation capacitor.
R1 is supposed to be the big input impedance required for piezo pickups.
The first opamp stage should amplify the signal 10 times, bringing it to a bigger level.
R4 and C2 act like the first stage of a low pass filter - for tone control. It is tuned to be above the highest violin frequencies (4200hz) to eliminate hiss that can cause feedbacks.
The second opamp stage should cope for the losses of the first low pass filter amplifying the signal two times.
VR2 and C4 are the active low pass filter that should act as the "Tone" control of the preamp.
VR1 is the volume.
C5 is separation cap.

I'm not completely sure that I designed it correctly, so any help / advice would be welcome.

I built it, but unfortunately, it does not work. By this i mean that it produces no signal on output.

I measured the input voltage on the Tl072 IC and it is correct. Unfortunately i don't have any idea what i should do next.

Can you guys give me any advice?


Thank you in advance!

[PRR]

> I measured the input voltage on the Tl072 IC and it is correct.

"input voltage"?? Power, audio, bias? What is the number, and what does it mean?

Hint: Generally, signal pins should idle "half-way between the supply rails". The signal pins can not go ALL the way to the supply voltages and keep working. They can't even be close and leave room for signal, so yuo aim for the center. Here your supplies are (seem to be!) zero V DC and +9V DC. And your signal pins are also all at zero V DC.

[kevinh]

Hi blana,
PRR is correct. Most op-amps are designed to have positive and negative supply rails. For the TL072, pin 8 is the positive supply connection, pin 4 the negative.  You have taken pin 4 to ground, it should be connected to a negative supply of the same voltage as the + rail. If you do not wish to have dual supplies the op-amp has to be biased, or adjusted so the output pin is set to half the supply voltage. This can be done with a couple of resistors, of suitably high value, so as not to lower the input impedance of the amp too much. Suggest you follow the link below so you can gain a better understanding of the technical requirements. Hope this helps.
http://www.eng.yale.edu/ee-labs/morse/compo/sloa058.pdf

Best of luck

[Seljer]

Other than the aforementioned virtual ground issue, personally, I'd put the volume control and tone control in between the opamps for a constant low output impedance.

edit:
Also, aim for a higher knee on the first filter. Even the highest notes contain harmonics and such so make it go up to 10khz or even higher to preserve those, you can always get rid of high end later with the tone control. For a lower parts count, you could do away with R4+C2 and just place a capacitor in parallel with R3 for the same result.

[blana]

Hello everyone,

Thank you for your great help, i don't know what i would do without you guys, probably i would do a frustrated table flip once a day.

First of all some clarifications: when i said "input voltage" i meant the power supply of the IC, sorry for the vague meaning.

After reading your posts and the first part of the document from Texas instruments, i tried to do what seemed to be the obvious solution: a voltage divider to bring the + rail to half the supply voltage.
What i did was to add a 4.7M resistor between the supply and the + pin, like so:

https://www.dropbox.com/s/8pxmtxs0r6e09e8/divider.png

From Kirchhoff's voltage law, I deduced that the voltage would be split equally between the R5 and R1 resistors, giving me what i needed - half the supply tension between the + and the - rails.
After patching the circuit, i measured the voltages - big surprise for me, R5 took most of the voltage, leaving only 0.4v for R1. I don't understand why this happens, i presume my knowledge / logic failed me again.

Another thing that i noticed during my tests, is that if i hold my hand on the volume pot, i get a signal from the piezo pickup. I removed the pot and added a simple wire, for simpler testing. The same thing happens - i hold my fingers on the wire, i get i signal. I presume that my fingers bias the IC somehow, but i'm not sure how. Anyway, its weaker than the non-buffered pickup, so i presume the bias is not right even with my magic fingers.

So i', still stuck.

Thanks in advance!

[Seljer]

The other sides of the opamps need to referenced to your 1/2 supply voltage too. That or decoupled with sufficiently large capacitors. This would be regarding R2 and R6 in your schematic.
So you either add the caps in at those two places (about 10uF should preserve all the low end), or use a resitive divider and reference everything to that

Here I quickly redrew everything with the improvements I mentioned earlier

(And I juse realised C1 may need to be bigger, so put 47 or 100uF there)

[PRR]

> Kirchhoff's voltage law, I deduced that the voltage would be split equally between the R5 and R1 ... i measured the voltages - big surprise for me, R5 took most of the voltage

Observe everything!

What is the output pin of IC1a doing? By now you should suspect it needs to be near half-supply.

But I bet it is near V+.

Look what you told it to do. You apply V/2, say 4.5V, to the input. R2 R3 suggest a Voltage Gain of 11. 4.5V*11= 49.5V!! That can't happen on a 9V supply. Instead the output sticks at maybe +7 Volts. R3 R2 divide this down to perhaps 0.6V at IC1a's "-" input. The part your Kirchhoff didn't know about is internal diodes between "-" and "+" inputs. So the "+" input is pulled-down to a low voltage, over-riding the 5Meg bias resistors.

Another aspect. 5Meg bias resistors are large. Your meter, if a DMM, is a 10Meg load. So even without a slammed op-amp, 5Meg and 5Meg and a 10Meg meter gives 0.4 not 0.5. It is worse if you use a passive needle meter. The cheapest 1K/V meter on 15V scale is a 15K load! A 10K/V meter on 15V is 150K. And 150K meter load on 4.7Meg||4.7Meg is 0.3V or so.

Simon graciously re-designed your idea. The R2 R3 (now R4 R5) network returns to Vref not to ground, so the opamp does not try to amplify its Vref bias. Volume also returns to Vref. All audio signals ride on the same 4.5V DC reference. At the input and output coupling capacitors let the signal back down to zero DC so the DC does not annoy source and destination.

[blana]

Hello,

Wow, thank you very much guys, this is really, really helpful! Thanks for taking the time to explain simple things to a noob such as I.

Thanks a lot for the schematic Simon, you are very kind to do this!
Thanks for the detailed explanation Paul, now i understand better how opamps work and why vref should be on the minus side!

I'll get to work and rebuild the whole thing from scratch, I'll get back with the results as soon as i have them!

[blana]

Hello again,

Today I had the time to assemble the circuit. Unfortunately, it still does not work.

This time I get a signal, but it is a huge buzz. Only when i strike the piezo pickup there is a audible signal passing through. If i unplug the piezo from the circuit there is a variable buzz and crackle, similar to the sound the radio makes when its not on a station (but a lot louder and with a lot of crackles).

Another thing i think is worth mentioning, is that i used slightly different values for some components, because that's what was available for now. I dont think the behavior i described above happens because of these differences, but here is the list anyway (Simon's schematic): R1 and R2 are 50k instead of 47k; R4 is 8.2k and C3 is 2.2n; R11 is 500 ohms and C5 is 2.2u.

Again, I have no idea what to do next.

Thank you in advance!