Dual OP-amp tremolo

[joehill]

Have anyone built this circuit?

http://www.elektroda.pl/projekt/tremolo.gif

I found a schematic in a swedish DIY electronics book from the 1980s ("501 elektroniska byggprojekt"), and found almost exactly the same schematic and with a PCB layout but in polish
http://www.elektroda.pl/rtvforum/topic122059.html

So I built it, but used 220nF in parallel with 100nF instead of 330nF, and GE oa1182 instead of GE aa118. It makes A LOT of noise, flowing in tempo with the oscillator. Most of it disappears when I take the tone knob on my Tele down to zero.
I have not really understood how the diode modulator works with the amplified signal (IC 01A) and the oscillator (IC 01B), but could it be the wrong diodes causing this?

I guess it would work good on a guitar with the input buffer and the op-amp, but it seems like the signal out is weak and it distorts when turning the signal op-amp (IC 01A) gain up.

[PRR]

What kind of noise? Hum, hiss, thumping, grinding....?

What power? Battery, good supply, dubious supply?

The guitar tone control "should not" have any real effect. If it does, I'd suspect a miswiring between guitar and T1.

> it distorts when turning the signal op-amp (IC 01A) gain up.

I think that's the way it is.

> I have not really understood how the diode modulator works

Signal must flow from C7 through D1 and D2 to R19 (1K) to get to the output.

How can audio signal get throut D1 and D2?

In your mind, grab the top of R18. Pull it down to ground and up to a positive voltage.

At ground, there is no steady DC current through the diodes. They tend to be "off", high impedance. What little signal gets through is absorbed by the low-resistance R19.

A very large signal may turn-on one diode but turn-off the other diode. Very little signal will get through.

Now bring the top of R18 to a positive voltage, perhaps 3.2V. You will get 2mA steady current flow in R18, splitting to two diodes D1 D2 and resistors R16 R19, 1mA each leg. The resistance of a diode with 1mA flowing is about 30 ohms. Signal passes through 30+30= 60 ohms, loaded with 1,000 ohms, so 1000/1060 or about 95% of signal gets through.

And this will be pretty much the same with *any* diode. There's reasons why Silicon may be a bit better than Ge. Try 1N914 or even 1N4007.

Actually the 100K and 220K network won't flow anywhere near 1mA to the diodes. I'm guessing the diode impedance actually varies from >1Meg to about 700 ohms each, 1400 ohms total. With the 1K load, signal level varies from 1% to 42%.... tremolo! Because the average of 1% and 42% is about 20%, or a 5:1 loss, we want gain of 5 to get back to unity gain. Indeed IO1A seems to offer gain of 1 to 6.

But. There's real issues. And some sign that it isn't totally thought-out.

What does T2 do?? The signal is already at a 1K impedance, super-low by guitar-cord standards. We could take an output right from C8.

Likewise a minor mod of IO1A could eliminate T1 _and_ get a lower hiss level.

Diode bridges distort at 20mV. This is below normal guitar level. Why are we boosting *before* the diodes? If the output is not much over 20mV, but guitar level is often 50mV-200mV, don't we need some gain after the bridge?

It is interesting that it seems to be a fully worked-out project with PCB. I can't follow the Polish discussion to see what people are saying.

[joehill]

Thank you for taking your time answering. I will read it through again tomorrow and will put out some recordings. But I guess I would describe the noise as a cricket (with the tremolo going).

The swedish text book I found the same shematic in is very serious, and in the short text they say it would be good for ex. an organ.

Have to go and see a band of friends right now playing in town.  (http://www.youtube.com/watch?v=KiUbxxez9mg)

best regards / J

[joehill]

So I got my hands om some AA118 today and I switched them. They have a bit higher VF about 500mV, the oa1182 had 280mV.
Also I changed buffer transistors to bc547A and BE voltage went up to 650mV (from previous 580mV) on Q1/Q2. (Used 2n5088 previously)

but the noise is still there.

Sound clip me strumming a chord with an Strat in noice cancelling position (neck/mid pu)
https://www.dropbox.com/s/2a965m2wwtsv6h2/trem1.mp3

Any ideas anyone?

[pinkjimiphoton]

is it on a breadboard? holy cow, that's noisy as....

[joehill]

No, only input/output jack goes through a breadboard right now. Ive soldered it onto the pcb with pcb mount pots

[joehill]

Are Q1/Q2 starving? Should I change the resistors? Which ones?
As I said, its 0.64-0.65V BE drop. (BC547A)

BTW using pretty much exactly 9V battery


best regards / J

[joehill]

Here's my voltages measured from ground
opamp is tl072.

v+ 8,79

q1 B 3,89
   E 3,28
   C 8,79
(BE lower than when measuring between legs)

Q2 B 3,88
   E 3,28
   C 8,79

op
pin1 4,39V
   2 4,39V
   3 4,38V
   4 0V
   5 4,37-4,41 (oscillator)
   6 4,39V
   7 3,66-5,26V (oscillator)
   8 8,79V

[joehill]

Now bring the top of R18 to a positive voltage, perhaps 3.2V. You will get 2mA steady current flow in R18, splitting to two diodes D1 D2 and resistors R16 R19, 1mA each leg.

The output voltage (pin 7) gets eaten by about 4V by the R13 100k, so the top of R18 is about 0.2V

Mybe the diodes arent really active, explaining the high noise/low signal ratio?

[PRR]

> top of R18 to a positive voltage, perhaps 3.2V.

That was just to get the thinking started. As you say, as I wrote later, the voltage will be very much lower due to 100K-220K network.

> top of R18 is about 0.2V

About 50uA so 0.05V in the 1K resistors, say 0.15V in Germanium diodes, 0.2V may be just what we expect.

[thehallofshields]

Necro bump. PRR's description of the circuit was very good. I found another drawing of an AM modulator that might better illustrate the concept.