This Week on the Breadboard: The 'lectric Mama Flanger

I have a little more testing / tweaking to do. I put the Flanger after a pedal containing a charge pump and the charge pump's switching freq was beating with the BBD's sample rate and making it "squeal like a piggy."
 
  • Like
Reactions: fig
Investigation & Findings:
The pedal that contained the charge pump mentioned above was a modified Chela as described here and here. I tried another pedal with a charge pump in it, the Crunch Captain Deluxe and observed the same squeal. Next, I replaced the 'lectric Mama with a commercial Flanger, a Monoprice yellow Flanger which appears to be a relabeled Mooer E-Lady. It exhibited the same squeal behavior.

The cause of the squeal is some of the charge pump's 40KHz switching noise sneaks out of the pedal via the output jack. When that 40KHz noise gets into a sampled circuit, like a BBD delay, chorus or flanger, the 40KHz signal is aliased down into the audio range when the sample clock is within ~10KHz of the charge pump noise. When the flanger clock is swept toward the longer delay end on the range, that's exactly what happens. Noise problems like this can be attacked on two fronts: the transmitter and the receiver.

We can make the transmitter end less noisy by filtering the signal before it gets out of the box. The modded Chela, which Cooder aptly named "Rippa," has a PRESENCE switch at the end which varies the high freq cutoff. Either the up or down position of the PRESENCE switch provides enough high freq attenuation to kill the squeal. The Crunch Captain Deluxe has no output filter on-board, so I will have to kludge one in.

We can make the receiver less susceptible by filtering the signal coming in to remove as much of the 40KHz switching noise as possible. I improved the 'lectric Mama's filtering upstream of the BBD by increasing R2 and C4, and adding another filtering stage with R7 and C5. I had to retune the filtering on the output side of the BBD (C9, C10 & R12) to restore the overall freq response. While I was at it, I tweaked the Vref voltage (increased R19) to maximize the BBD's headroom. While all of this helped, it was not sufficient to kill the squeal. It is also necessary to make sure that any pedal containing a charge pump does not transmit excess switching noise.

'lectric Mama v3.3 page 1.png

'lectric Mama v3.3 page 2.png
 
My 4 Knob Flange chip just came in so not to bad on shipping time. May order the StompLFO chip just to have it.
 
  • Like
Reactions: fig
Thanks, that answers a lot of questions (I think). If the charge pump is the source, wouldn't all circuits utilizing similar charge pumps, BBDs and clock chips exhibit similar results?
 
10 votes on the wish list guys. We can do better.
Remember, the voting is done on the top right area of the post, not the normal "LIKE" for a forum post.

wolf-of-wall-street-rookie-numbers.gif
 
Investigation & Findings:
The pedal that contained the charge pump mentioned above was a modified Chela as described here and here. I tried another pedal with a charge pump in it, the Crunch Captain Deluxe and observed the same squeal. Next, I replaced the 'lectric Mama with a commercial Flanger, a Monoprice yellow Flanger which appears to be a relabeled Mooer E-Lady. It exhibited the same squeal behavior.

The cause of the squeal is some of the charge pump's 40KHz switching noise sneaks out of the pedal via the output jack. When that 40KHz noise gets into a sampled circuit, like a BBD delay, chorus or flanger, the 40KHz signal is aliased down into the audio range when the sample clock is within ~10KHz of the charge pump noise. When the flanger clock is swept toward the longer delay end on the range, that's exactly what happens. Noise problems like this can be attacked on two fronts: the transmitter and the receiver.

We can make the transmitter end less noisy by filtering the signal before it gets out of the box. The modded Chela, which Cooder aptly named "Rippa," has a PRESENCE switch at the end which varies the high freq cutoff. Either the up or down position of the PRESENCE switch provides enough high freq attenuation to kill the squeal. The Crunch Captain Deluxe has no output filter on-board, so I will have to kludge one in.

We can make the receiver less susceptible by filtering the signal coming in to remove as much of the 40KHz switching noise as possible. I improved the 'lectric Mama's filtering upstream of the BBD by increasing R2 and C4, and adding another filtering stage with R7 and C5. I had to retune the filtering on the output side of the BBD (C9, C10 & R12) to restore the overall freq response. While I was at it, I tweaked the Vref voltage (increased R19) to maximize the BBD's headroom. While all of this helped, it was not sufficient to kill the squeal. It is also necessary to make sure that any pedal containing a charge pump does not transmit excess switching noise.

View attachment 19191

View attachment 19192
Strymon might have to look at this!!!
 
Thanks, that answers a lot of questions (I think). If the charge pump is the source, wouldn't all circuits utilizing similar charge pumps, BBDs and clock chips exhibit similar results?
I would have thought so, but I just tried a ModTone Space Flanger and a Moen Chorus after the Crunch Captain Deluxe and the modded Chela. No squeal. One possibility is that neither one sweeps the clock freq low enough. The Space Flanger may have more filtering, I'll have to open it up and see.
 
The 3-day course I attended was a subset of that material. Part-way into the first day, I turned to the engineer sitting next to me and said "You know, I was already doing some of this stuff and I just chalked it up to laziness." He said "Yeah, me too." It felt good to have our "lazy" techniques validated.

I'm going to download all of this stuff, thanks!
Late to this party, but what were the “lazy” techniques?
 
We would think of components in terms of their impedance or admittance, whatever made the equations simpler. We would express impedance and admittance as the Lapace transform. For instance, a capacitor's impedance is 1/sC. ("s" is the complex frequency, expressed in radians per second). We would express parallel components are their equivalent admittance and create a substitute variable to use in the equations. Example, the admittance of a cap and res in parallel is X1 = 1/R+sC. It's a lot easier and cleaner to write X1 instead of 1/R+sC. Lastly, we would arrange the equations so that each term had a clear physical meaning. This term represents the DC offset, that term represents the AC gain, that sort of thing. The whole idea was to turn a complicated calculus problem into a simple algebra problem. We didn't invent the technique, we just exploited it to make our jobs easier.
 
Last edited:
Ooooo wait a minute Chuck! We could feed mayonnaise to the tuna. This is big!

That's not what I was going to suggest....hang on......ahhh there it is...

Could the ESR of the caps in the circuit play a role in poor DC filtration of the charge pump's incessant swine wave?

Forgive any ignorance that may be involved with that thar question.
 
It's possible, but I use good caps for the charge pump. No cap is perfect, so there is always going to be ripple on the 9V coming in and on the -9V, +18V, or whatever voltage we make.
 
  • Like
Reactions: fig
I should change the title of this thread to "This Month on the Breadboard." I think I'm about done polishing this turd. It's come a long way from the minimalist Electric Mistress based design to where it is now. The EM is a good design and it sounds great, but it's susceptible to 40KHz noise from the charge pumps in other pedals. I've done what I can to filter out the charge pump noise, and it's pretty effective, but if you listen real hard you can still hear it when the clock sweeps down to the bottom-end of its range. Some pedals are noisier than others.

Presenting version 3.5.2. As I state in the notes, L1 is optional. It helps reject charge pump noise, but if you don't have a 750uH choke (Tayda sells an 820uH, that's plenty close), just use a jumper there. I changed BLEND to C-taper. B-taper works, but most of the useful range is up near the top. Just about any JFET will work for Q1, you want something with Vp below 2V and Idss above 2mA. You could even use a MOSFET. Voltages marked are for reference only. Gain is unity. The active filters before and after the BBD are 3rd-order 8KHz low pass. This circuit has the usual pre-emphasis (C4-R6) and de-emphasis (R13-C8) to reduce BBD noise.

'lectric Mama v3.5.2 page 1.png
'lectric Mama v3.5.2 page 2.png
 
Last edited:
Can’t believe I missed this thread for months somehow. Would I be correct in assuming that with minimal tweaks this could be adapted to accept the big-boy Tap-LFO chip for additional control over waveshapes via the distort input of the chip, as well as tap-tempo subdivisions, and an additional bank of 8 waveforms?
 
  • Like
Reactions: fig
Back
Top