Chorus and phaser in same box noise question

[Luke51411]

I'm working on a multi with a distortion, phaser, and chorus. I currently have the phase and chorus pcbs on the breadboard they are guitarpcb dr phybes (p90) and mini-me chorus. I'm running the phase into the chorus. When I switch the feedback on for the phaser I get noise from the chorus clock, it's a chirp sort of sound. Is this something that should sort itself out with careful wiring once boxed? It seems to be fine without feedback on the phase shifter. If the noise is inherent, I'll just leave the feedback switch off.

[Mark Hammer]

1) The chirp may well be your feedback loop resulting in momentary oscillation.  That is, for a moment what is getting fed back is juuuuusssttt above unity.  That can often be addressed by increasing the value of any series resistance in the feedback path.

2) Is your power for the LFO "decoupled" from the audio path?  Normally, if it wasn't suitably decoupled, you would hear a tick at the rate of the LFO.  It can help to use a low-power op-amp for the LFO.  This is why you will often see LM358 and TL022 dual op-amps used for LFOs.  They draw less current, producing much less of a spike on the power lines feeding both the audio and LFO subcircuits.

3)  I'm also kind of wondering if any spike on the power is also shared with the distortion, and thus amplified.

[Luke51411]

The chirp is in time with the chorus clock. Well, I'm nut sure if it's exactly in time but it's speed is affected by the chorus rate. I don't currently have the distortion hooked up to the breadboard as I'm trying out all possible combinations before committing to a box. I think both have low power opamps, one a tl061 and the other an lm358 for the lfo's. When switching on the feedback, the noise floor goes up significantly as well. I would expect it to increase some but probably not as much as it does compared to another p90 build I've done. Would using 1% resistors help lower the noise floor as the jfet stages would likely be closer to unity?

[Luke51411]

I have a p90 on vero as well maybe I'll try that with the chorus and see what happens.

[Mark Hammer]

What many designs do is include a cap in the feedback loop of one of the phase shift stages - generally either the last stage or the one the feedback loop returns to - to trim back the treble in the phase shift path and keep any hiss accumulated via the recirculation in check.

With a 10k feedback resistor, an 1800pf cap will give you a rolloff around 8.8khz, which should take some of the noise away without detracting from the phase shift sound.

[Luke51411]

In series with the feedback resistor?

[Mark Hammer]

No, in parallel.  Um, forgive my verifying, but you know which resistor I'm talking about?

[Luke51411]

The one I have on the switch I think... That's where the feedback goes through

[Mark Hammer]

Using this Tonepad drawing of a Phase 90, the suggested cap could go in either of the position where I show a thick red line.

[Luke51411]

Ok thanks mark. I was thinking it would be the 22k. I'll have to give it a try.

[Tony Forestiere]

Luke,
I was looking at some of the build reports at Tonepad. Did you make the 22K at the top "switchable" to add or remove it with a toggle you are calling the feedback switch?

[Luke51411]

Yes that's correct.

[Tony Forestiere]

Cool. Just wanted to get clear in my mind the feedback switch that causes the chirp from the chorus clock.
Try Mark's cap suggestion in one or the other (or both) positions.
The best to you.

[Mark Hammer]

I don't think the suggested cap will rid you of the chirp, but it should reduce the increased noise that occurs when regeneration is up too high.

Now, if one uses the 22k fixed resistor value shown (or 24k), I would not consider that "high feedback".  I tend to use a 12k fixed resistor with a 25k-50k pot in series, or else a 47k fixed resistor, if I'm using a 3-way toggle to yield high, stock, and low feedback.  Once the feedback resistance gets below around 15-18k, you start to hear a bit more hash in with the audio, and the cap will tame that.  But at stock feedback settings I don't usually hear any objectionable increase in noise.

My gutu says that the chirpis a result of the multiple effects sharing the same power line, with insufficient decoupling of power to the various subcircuits.  The LFO in your two modulation circuits may only produce a teeny spike, and largely inaudible tick, but if that spike and tick is amplified by the distortion, then that is likely where the chirp comes from.

A word about decoupling.  The decoupling of the supply lines is provided by tying each chip's power line to something like a 10uf-47uf cap to ground, and feeding the +9v to that power connection via a small value (10R to 47R) fixed resistor (instead of directly).  The RC combo allows the chip to get enough current, but the cap acts like a momentary storage battery or uninterruptible power supply, to deliver smooth power to the chip for a few instants...just long enough to get through that disturbance in the power line caused by the LFO's sudden draw atthe start of each pulse/wave.

[Tony Forestiere]

Now, if one uses the 22k fixed resistor value shown (or 24k), I would not consider that "high feedback".  I tend to use a 12k fixed resistor with a 25k-50k pot in series, or else a 47k fixed resistor, if I'm using a 3-way toggle to yield high, stock, and low feedback.  Once the feedback resistance gets below around 15-18k, you start to hear a bit more hash in with the audio, and the cap will tame that.  But at stock feedback settings I don't usually hear any objectionable increase in noise.

Great mod. Now I see the interaction.

A word about decoupling.  The decoupling of the supply lines is provided by tying each chip's power line to something like a 10uf-47uf cap to ground, and feeding the +9v to that power connection via a small value (10R to 47R) fixed resistor (instead of directly).  The RC combo allows the chip to get enough current, but the cap acts like a momentary storage battery or uninterruptible power supply, to deliver smooth power to the chip for a few instants...just long enough to get through that disturbance in the power line caused by the LFO's sudden draw atthe start of each pulse/wave.

I understand finally. Very clear and concise. Thanks, Mark.
Previewing my posts is not working for me. I apologize if I break something.

*edit* I just realized why the early Anderton dual supply circuits used quite a few electrolytics. Even with battery power.