Twin-T oscillator

[alkopop79]

I'm trying to build a twin-T oscillator for synthesising drum sounds. The LM741 heats up pretty quickly and the circuit hardly oscillates even when the feedback is at 100 per cent. Even with an LM386 amp I hardly get anything (though I can see some oscillation on the scope). The circuit gets 10ms 5V pulses from an Arduino. Does the TL071/2 run from single supply? Can it replace the LM741?

[img]http://www.flickr.com/photos/alkopop79/8644385451/[/img]

---

twin_t.png
Description:
Filesize:	40.55 KB
Viewed:	482 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

[elmegil]

Have you checked to see if the chip is bad, by putting in another one?

TL072 does not operate single supply. LM324 does, but it's a quad chip. I believe LM321 would be the single amp version, but I've not used them.

[blue hell]

alkopop79

The speaker ... I'd not directly drive a speaker from an opamp .. if you want a speaker in the box make a separate power amp for it.

[JovianPyx]

As Blue Hell said - the speaker - it's a very low impedance load and could easily cause the problem. As Blue suggests, better to run the speaker from a separate amplifier.

[richardc64]

It still wouldn't work.

For single supply, almost any op amp would work in this circuit IF the (+) input is biased to 1/2V+. (Which some mistakenly call a virtual ground.) Triggers should then be capacitor-coupled in.

[JovianPyx]

It's possible that the signal generator symbol might be intended to account for the required bias. I can't really tell for the symbol in the circle.

However, the signal generator also indicates 500 mHz which is .5 Hz. I suppose that this filter is intended to operate at very low frequencies?

[Scott Stites]

[alkopop79]

Thanks for all the replies! I did further simulation and indeed, without virtual ground it won't work. THe circuit can be found here:


https://www.circuitlab.com/circuit/g52648/twin-t-with-trigger/

You can open it in the editor and play with it. The time domain simulation will even draw a nice kick drum. Eventually I found a sail splitter circuit here:

http://tangentsoft.net/elec/bitmaps/vgrounds/rdiv.png


However, when I build the circuit on the breadboard, it doesn't work. With 9V DC supply instead of getting 4.5V, 0V and -4.5V I get 9V, 4.5V and 0V. It seems as if the caps weren't even in the circuit. I checked the polarity and the value of the caps and even used different ones. Still don't get any virtual ground, the circuit acts as a voltage divider. A photo of the breadboard circuit:

http://www.flickr.com/photos/alkopop79/8646491414/

It's beyond me why doesn't it work. Any ideas?

[elmegil]

Are you using the "virtual ground" (yes I know I'm abusing the term) as your reference to see +/- 4.5V? If you're using "real ground" then what you describe is exactly what you would expect, yes?

[alkopop79]

It's working. I was a dumbass: the ground of the multimeter has to go to the junction, the middle ground of the circuit. Indeed, from a 9V supply I get 4.5V, 0V and -4.5V. I changed the 741 to a TL072 and it works! So far I've only checked it with an oscilloscope and it very much looks like an analog drum!

[alkopop79]

[elmegil]

I've done even more foolish things and even owned up to them here, so don't feel too bad

[alkopop79]

Tried it with the amp but unfortunately it distorts the crap out of the signal.

[alkopop79]

[alkopop79]

Is it possible that the amp distorts because it gets -4.5V to 4.5V signal? It's a single supply amp though.

[Scott Stites]

[Thomas_Henry]

Hi Gang,

My suggestion would be to jettison the op-amp altogether and use a transistor as the gain element for the Twin-Tee. That's exactly what I did for the Sirius Metronome published in Nuts & Volts in January. I too used the 386 for the output stage. It all worked out quite nicely with a very satisfying woodblock sound.

http://www.nutsvolts.com/forums.html?/magazine/article/january2013_Henry

Click the mouse icon to preview the article.

Thomas Henry

[alkopop79]

Thanks, neat circuit! I will give it a try! What's the difference between the LM380 and the LM386?

[-minus-]

Why is 'virtual ground' abusing the term? Can someone explain this?

[elmegil]

I believe (though Richard can correct me if I misunderstood his comment) that it's because the true "virtual ground" is when you hold one of your op amp inputs at ground and it forces the other one to also be at ground potential even though they're not shorted.

[Scott Stites]

I actually decided to subscribe to N&V because I know TH has actually published a string of articles there.

[alkopop79]

I've made a tiny workbench split power supply on a piece of strip board with a plug for the adapter. Super handy!

https://plus.google.com/photos/101922570428775677855/albums/5867116777743457841/5867116927902916626?banner=pwa

The supply has pins for GND, V+, split V+, midpoint and V-.

[alkopop79]

Friends,

the circuit works! It's pretty amazing, I get massive kicks on the output. However the circuit audibly 'buzzes' when I touch a pot. Also, the it tends to distort at higher frequencies and it's pretty noisy. Any ideas how to fix that?

https://soundcloud.com/alkopop79/twin-t-percussion-osc

[sonic]

I hope nobody objects to me resurrecting this thread. alkopop79 I tried to build your circuit but I'm having some problems and I wonder if anyone can give me some pointers. I used LM324N instead of LM741. I also had no 3.9M resistors and I used 3.3M. Apart from that I think I used everything the same. I get a basic kick sound but the Twin-T part seems to do nothing to filter the signal. No change when I tweak the pot, or even if I remove those parts of the circuit.

I guess either the substitutions I made are no good, or I have made some kind of rookie mistake in the build. The problem is that I don't understand how the current should be flowing and what I should be testing for in different parts of the circuit in order to debug.

The 555 trigger circuit is on the left side of the breadboard. On the right there is the 'virtual ground' circuit (yes yes I know. grumble grumble) at the top. Underneath the LM324N is the non-functioning Twin-T section. The main power supply is a 9V battery. Perhaps I should be building and studying a simpler RC filter in order to understand better what the circuit should be doing.

I've been kicking around here for a while but took an extended hiatus from electronics and am still to all intents and purposes still a total newb. Any assistance at all would be greatly appreciated. I am not necessarily fixated on this particular circuit. My plan is to make some kind of crude drum machine and I have a shedload of LM324N. Twin-T seems like a great way to make a kick sound. I will make a new thread for my project once it starts to take shape.

---

bpsXxaQ.jpg
Description:
Filesize:	649.17 KB
Viewed:	166 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

[elmegil]

I'm far away from anything to test with right now, but I'd expect the 3.3 to have more impact than the op amp. Do you have any 560k resistors? Put one in series and see if it improves anything, that should get you close to 3.8M....