Help understanding a Schaller TR-68 Clone Phase Shift Oscillator Circuit

[bushidov]

In regards to an older post I made, I am going back to this circuit for some more analysis. See old post here:
https://www.diystompboxes.com/smfforum/index.php?topic=123088.msg1162846#msg1162846

So I've started looking up transistor based Phase Shift Oscillator circuits to get a better understanding of how the Schaller TR-68's oscillation is working. I first found this link: https://www.circuitstoday.com/transistor-phase-shift-oscillator and noticed the RC Phase Shift Oscillator resembled the TR-68's to some degree, if I shifted the components around on the schematic. As the Schaller Tremelo had a switch for "fast" and "slow", I broke the schematics up into two circuits so that I could look at them separately.

Fast:


and
Slow:


I noticed that by decreasing the capacitor C8's capacitance and reducing the resistance that comes after C8 that goes to ground (R13 and R14), the oscillations are generally faster. Conversely, increasing that capacitance (Parallel C8 and C7 = 26.7uF) and increasing resistance (R14 without a parallel resistor to reduce its value), the oscillations are generally slower.

But I am still a little uncertain about a few things, if using this guide below is accurate:


1. Why do I have two transistors?
2. Using the circuitstoday diagram, what part on the TR-68 schematic is equivalent to R1?
3. Using the circuitstoday diagram, what part on the TR-68 schematic is equivalent to Rc?
4. Using the circuitstoday diagram, what part on the TR-68 schematic is equivalent to Cout?
5. Using the circuitstoday diagram, what part on the TR-68 schematic is equivalent to Re?
6. Using the circuitstoday diagram, what part on the TR-68 schematic is equivalent to Ce?
7. Why do C7 and C8, have their polarity facing one way, while C9 and C10 are facing the other?

I am guessing these are some pretty noob questions, but this is the way I learn, so I have to ask.

Thanks for any and all help in having me understand this!

[11-90-an]

Not so sure so these are just guesses...

1. ??
2. Q3 and R17
3. R18
4. C5
5. No Re (for gain)
6. No Ce (for gain)
7. Because if you see closely there are some connecttions to +9v... C9 an C8 can actually be turned around... you can see that both points are connected to ground in some area...

Forgive me for asking, bushidov/erik, but how did your chopped oc2 build go?

[antonis]

Not been able to follow items correspondence for the moment, but for 1. the answer is impedance matching..
(Q3 is an Emitter follower effectively isolating RC phace shift oscillator from Q4 CE Amp)

[bushidov]

Forgive me for asking, bushidov/erik, but how did your chopped oc2 build go?

It's going on the back burner. My day job is getting ridiculously busy and has been taking me away from a lot of my hobby time.

[bushidov]

1. the answer is impedance matching..

Alright, but then how did the circuitstoday.com circuit get away with using only 1 transistor? What would I have to do to reduce the TS-68 down to one transistor? I'm not saying I would, just, "if I had, to, I'd have to do..."?

[antonis]

Alright, but then how did the circuitstoday.com circuit get away with using only 1 transistor?

I presume it's only an example ..

[PRR]

The 1-transistor job barely works.

2 transistors can be more likely to work with marginal parts and large changes of rate set parts.

Transistors have been CHEAP since Jimi left us. Don't be stingy.

[bushidov]

Don't be stingy.

No worries there. I was just trying to understand it is all.

[bushidov]

So, I still have some questions. Perhaps, I am drawing this incorrectly, which is why I am not getting it, but if C5 is the Cout of the circuit, what are C6, R11, and R12?

And how is VR2 + R16 altering the frequency of the oscillation?

I am guessing this is more than a 3 stage RC oscillator?

[11-90-an]

I realized my first most here has quite a lot of errors...
Even now there still might be errors, so take my conclusions with a grain of salt...

R1 from the circuits today is essentially R16, R18 and VR2 in series, and also adjusts the feedback going back to Q3

R11 and R12 form a voltage divider to limit some output signal that can enter back into the circuit for more feedback (i think)

[Rob Strand]

Those parts look weird because it looks like it's drawn wrong.

This makes a bit more sense,



Where did get your schematic?

At this point I don't trust the schematic I posted 100% either so there could be a bit of truth in both.


The oscillator looks like a phase-shift oscillator, except it's twisted a bit with the way the frequency pot is wired.
I think that type of wiring cause a notch.  For the oscillator that probably skews the phase.

[bushidov]

Where did get your schematic?

I don't remember exactly, but it's one I built up and made a JLCPCB run of. It works great. Now all I am trying to do is understand, "why/how it works"

I know it is a phase shift/RC oscillator from what folks here have said in the past, so I was looking up examples of phase shift oscillators, and they all tend to draw their schematics the same way, so I was trying to draw this one the same way everyone else was. This way, when sites try to explain the math to me about how what values do what, I can apply it to this pedal.

The problem is that I am not getting a 1 to 1 match with the Schaller's Phase Shift Oscillator and "everyone else's".

Any help explaining how it works would be much appreciated.

[Rob Strand]

The problem is that I am not getting a 1 to 1 match with the Schaller's Phase Shift Oscillator and "everyone else's".

Because it doesn't follow the standard phase-shift oscillator 100% I wouldn't be overly concerned with that!  It only takes a few changes and it will oscillate a different frequency.

In the schematic I posted, I'd (loosely) call these the RC components the "phase-shift" network:
C5, C6, C7, R10, R11, R12 (and R8 has some impact)

That's clearly an oversimplification since it completely removes the frequency adjustment components R15 and P2.
But that's just it,  it doesn't quite follow the common phase-shift circuit.   So that's too simple.

Those bridging components,  R15 and P2,  often have a *large* effect.   In fact you could kind of "improve" the (mental) approximation of the circuit as three T-neworks:
- C7 + R12 + half of C6  ; half of C6 means  4.7uF = 9.4uF in series with another 9.4uF
- the other half of C6 + R11 + half of C5
- the other half of C5 + R10 + C4
Then on top of that you have R15 and P2 bridging the lot.

All this is really just circuit gymnastics and it's not that helpful either.

At best all you can say is it looks like the network is big enough that there's enough phase shift to make an oscillator!

The main problem is it's a large network with too many parameters for our little minds to process.   There are ways to calculate the response of this by hand by it's a lot of work and it won't give you any more insight than simulating the circuit.

Any help explaining how it works would be much appreciated.

What you can do in a circuit simulator is,
- break the connection between the second transistor's collector and the junction of R15, C8, C7.
- connect a sine source in series with 2k2 (to duplicate the collector impedance) to the
  junction of R15, C8, C7
- measure the AC voltage and phase at the collector of the second transistor.

What this does is shows the gain and phase around the loop.  From the phase you can work out the oscillation frequency.

After that you can vary the components in the network say 5% in value and observe their effect on the gain and phase.   That will give you a good idea what components affect the oscillation frequency and the gain at that frequency.      The parts that affect these conditions the most are the ones that contribute to the behaviour of the oscillator.    There will be some parts which vary the frequency but don't affect the gain so much, those are the best parts to vary when you want to change the frequency.

In a very long winded way what I'm saying is it's too hard to come-up with nice formulas and you are better off getting an idea of what parts are most important by experiment.

The form for the frequency might end-up with something like,

f = k / {2 * pi  * [ C4 * (P2+R15) * C5 * R10 * C6 * R11 * C7 * (R12+R8) ] ^ (1/4) }

So roughly the frequency varies as 1/ (part value)^(1/4).   The part's effect on the gain is going to be the determining factor.