Build Your Own Clone Message Board

About a year ago some of us went around a bit about some mods to the Analog Vibrato kit--tape warble, chorus, flanger.

http://www.byocelectronics.com/board/viewtopic.php?f=8&t=59510

Finally found a weekend to return to this. I finished stuffing the board and included extra flying wires to allow breadboarding components into the Chorus Mod and Flanger Mod insertion point suggestions for the Analog Vibrato. The vibrato just by itself is very nice. I find for the chorus mod I might want the vibrato'd part to be a little below the dry part and using a pot in the place of the 47k resistor I found that as expected a slightly lower value there does the trick. This causes a bit of gain through the effect but I can live with that. If I choose a fixed resistance I will probably go with 33k or 39k, though I found values as low as 27k to be nice. (I think I'd rather do this than replace the stock resistors that set the gain for the vibrato'd signal. At some point maybe the right solution is a chorus pedal that was intended to be a chorus pedal. Good thing I have a couple of those from BYOC.)

My bigger interest though is the Flanger mod, which I've breadboarded but for which so far I can't tell that it's doing anything. I've tried to explore as much of the depth/rate/"additional pot setting" parameter space as possible but I don't really know where to prod. "Gross user error" is not ruled out yet.

If anyone out there has done this mod and has some words of wisdom I would gladly read them. I don't understand yet exactly what the extra loop does--maybe just feeds back some version of the signal into the path, though for flanging I'd expect that somehow the size of the delay is supposed to be smaller than for chorus or vibrato. Maybe I'll have to put the 'scope on the job but I'm not sure what to look for.

Typically, a chorus will have a ramp, saw-tooth, or some sort of triangular wave LFO. The VB-2 has a perfect sinusoidal LFO. So it will always have a different "flavor" than a standard chorus.

As for the various blend resistor values, you may want to look at our bass chorus schematic. If you exclude the oscillator circuit, and the 47k dry blend resistor, the VB-2 and CE-2 are pretty much identical. Adding the 47k resistor for dry blend is pretty much turns the VB-2 into a CE-2. The VB-2 might have a tiny tiny bit more gain because of the 7.5k resistor instead of the 10k resistor. Our bass chorus is just a CE-2 with a passive cross fader blend mod. So if you look at that schematic, you and compare it to the our regular chorus (or the CE-2), you can see how to replace the wet and dry 47k mix resistors with a blend control.

A flanger is just a chorus with a feedback loop. Or another way to look at it is that a flanger is just a modulated delay with a very very short delay time. If you replaced the 4096 stage BBD in an analog delay with a smaller 1024 or 512 BBD, and manually moved the delay time back and forth, you would literally have a flanger. Anyhow....yes, you are just adding a feedback loop from the end of the wet signal path back into the original signal before it's split into wet and dry, i.e., the input of IC1B. It may be better to feedback into the output of IC1b rather than its input. If you look at the BF-2, you will again see the same sort of similarities that are shared between the VB-2 and CE-2. The primary difference is that they BF-2 has an additional buffer at the beginning of the wet signal path. It also has an IC-based low pass filter at the very end of the wet signal path instead of a transistor-based low pass filter. I don't view this as something intrinsically inherent to flangers. I think it was more a case of, "we needed that extra buffer at the beginning and now we have half an op-amp left over, so let's save a few cents and also save space on the PCB and just use that instead of a transistor". Anyhow...I'm digressing again...The point is that maybe the feedback signal would like a buffer between it and the wet/dry split. So you could also try feeding the signal back in at the base of Q2 or maybe even directly into the BBD at pin 3. I mean...if you've got it on bread board and you're experimenting, you might as well try all points between the base of Q1 and pin 3 of the BBD.

_________________
*patience is a virtue*

Please do not PM me. email is prefered. keith@buildyourownclone.com

Thanks Keith for the insights. I too think of a flanger as the short-delay side of the family that includes chorus, and so far with this Analog Vibrato the most flanger-like sounds are found when the delay trim pot is set near the minimum. That setting makes it harder to get the trippiest vibrato sounds but that's ok. I see in a BF-2 schematic I found that there is a (resonance, I guess) feedback line like the hand-drawn one for this flanger mod so I think I see now what the general idea is.

To be clear I built the Analog Vibrato as stock onto the PCB (and into the enclosure with all the rest of the parts, it's a completed and working kit build) and the only breadboarding is for the components in the hand-drawn additional loops, for which I soldered in some long wires to access points at the PCB (onto some cap leads before clipping them). (Just the chorus and the flanger; I'm skipping the tape warble.) The chorus mod of course just needs test clips to jumper in the 47k (or 39k or 33k or ...) resistor but I have three wires to a tiny breadboard to try out components like the 100nF coupling caps and the pot(s) for the flanger mod. I guess I need either to add some wires to other access points in the "all points between" section or else learn to be deft with a fine-tipped 'scope probe applied to points on the PCB. Since I find no audible evidence besides popping of caps that my flying wires & breadboarded components are doing anything I may still have some other underlying error I need to track down.

I don't have enough hands to do all these things at once so if anyone is reading this and wants to recommend an inexpensive but serviceable looper I could buy, kit or otherwise ... thanks in advance ...

After studying some other flanger schematics (Boss BF-2, Pearl FG-01, BYOC Stereo Analog Flanger) I get the impression that my problem could be, at least in part, a phase issue. Moving my injection point to one of those later junctions puts it on the other side of an inverting amp and though there are still some RC phase shift opportunities in the chain it would nominally make the phase relation more like what I think I see in the other schematics. Or am I all wet again?

I don't really know. Guitar effects pedals are notoriously frequently out of phase, yet still work. But I've had experiences while experimenting with BBD effects where the wet effect was very weak when signal was out of phase. So sometimes it matters. Sometimes it doesn't. I can't tell you why, though.

Maybe try adding a buffer to your bread board, just for kicks. Maybe look at the feedback circuit for the FP-777 flying pan (byoc soaring skillet) for inspiration http://byocelectronics.com/skilletschematic.pdf

_________________
*patience is a virtue*

Please do not PM me. email is prefered. keith@buildyourownclone.com

Trying to ease myself into the whole "nuts for guts" orientation--
OK gentlemen here's a high-altitude view of my Analog Vibrato build
that is starting to look more like something from John Carpenter's "The Thing."



I tack-soldered a few more flying wires in there in hopes of finding just the right
spot to inject the delayed signal back into the stream, for flanger-resonance purposes.
I used my high-tech continuity tester (love those film cans) to check my work,
and eventually those leads will plug into that little breadboard you see on the side.
Probably I'll put some of those signals on the 'scope as well,
to find out just what is going on with all the RC phase shifts along the way.

Trying to sort this out is very educational.
Between this and the Mimosa I am really learning a significant
amount of electronics.
--Which is good, because I don't think I'm shredding any more masterfully.

Looking forward to your updates on this project.

_________________
*patience is a virtue*

Please do not PM me. email is prefered. keith@buildyourownclone.com

Stayed up almost to midnight last night trying to learn something, in advance of family visiting for the week (--> no more experiments for a while), and it's clear that I have a lot more to investigate ... but since you ask ...

I looked through the other schematics I could find for flangers that seem to use the same BBD chip set and basic circuit topology (Morley Sapphire Flanger Box, Ibanez FL301, Pearl FG-01, Boss BF-2, BYOC Stereo Analog Flanger). What they seem to have in common is, not surprisingly, delivering a delayed copy of the signal from a low-impedance point through a coupling cap to an "amount control" and a series limiting resistor, through another coupling cap, back to a higher-impedance point before the MN3x07. I've given up for the moment injecting at the low-impedance locations (4558 pin 7 op amp output; Q2E = MN3207 pin 3 input) and am trying instead:
• the C2/R5/(C3orR4) junction (per the original suggestion)
• one end of R9 (R9/R8/C5)
• the other end of R9 (R9/C6/Q2B)

For the feedback I have so far used 0.1 uF to a pot terminal 3 (terminal 1 to ground) with terminal 2 ("the AC coupled signal to be returned") through another resistor and then another 0.1 uF cap, then to one of the junctions listed above. This is not exactly per the original suggestion but is more like what the other schematics show. I will likely try the original suggestion as well, and maybe also some other variations (sometimes I see a resistor or cap dropped to ground).

I've tried both a 10k and a 100k pot in that role so far. The main difference seems to be a difference in the needed minimum value for the feedback-damping series resistor (marked "?" in the original suggestion): about 2.2k for the 10k pot, about 5.6k for the 100k pot. Of course it's not surprising that these would be different. I note that that series resistance is definitely needed when injecting the fed-back into Q2B; otherwise it sounds like somebody is stretching an annoyed cat. Inserting at the earlier end of R9 isn't as much of a problem nor is inserting at C2/R5.

Insertion at either end of R9 seems a bit noisier than at C2/R5 and in both cases yields more audible LFO artifact--presumably because the C2/R5 point has more filtering following.

For the setups I have tried so far, more or less summarized above, I think I have a preference for the effect as produced at the C2/R5 injection point--it's cleaner and perhaps a bit more intense. The character is also slightly different from that produced at either end of R9. I think I will, for now, attribute that to the phase issues I mentioned earlier. All the other schematics show injection at a point earlier in the circuit that (according to the number of inverting amplifiers in between) has nominally the same phase as the tapped point, while in the original suggestion here the phase is reversed. However, given the other phase shifts that might occur in the whole path, and that *some* kind of flangey-sort-of frequency cancellations will occur no matter what additional phase difference is introduced, I won't at this point be snobbish on that point. "If it sounds good, it is good."

I am still convinced that a key part of this is having the delay times low, so a lot of my fiddling is with the delay trim. I don't have enough hands (or lab assistants) to cover this entire parameter space adequately. Also I am still trying to get a grip on exactly how the chorus mod (the 47k from the 4558 pin 7 to pin 2) interacts--sometimes it doesn't seem to matter; sometimes it makes a huge difference. (Could be user error though.) I am reminded of all the switches/choices in the Anderton (EPFM) envelope-follower-phase shifter from way back when.

I would not be averse to hearing some specific suggestions to try but can't promise I will actually try them!

Just a teaser, gentlemen, but I have now taken over the dining room table to set up for exploring this Analog Vibrato a bit more:




Probably couldn't get away with this if my wife were in town.

(iPhone as function generator!) Haven't figured out a lot yet. The phase shift through the entire pedal of a pure sinusoid, with effect minimized (depth zero, delay trim anything), changes by almost two full cycles between about 20 Hz and 800 Hz (that's just the cumulative effect of all the filters along the way). That must have some effect on the the way flange filtering will sound(*), but I don't think it should be so different from what happens in the other similar flanger pedals. I will resist dumping too many numbers here in the thread but at some point may attach a document with some measurements if I think they may be of interest.

I am already wishing I could get the rate lower (slower) and will probably eyeball the schematic to see if I can make an easy & reversible mod. (I think now I intend to finish this box off as a vibrato with the chorus mod, per known results, as I really do like the way it sounds in its intended configuration; but I will still try to figure out, if I can, how to get it to sound like a flanger.)


(*) edit to add: The concurrent discussion about the Phase Royal v1.0 in a different thread eventually sent me back into the archives to learn more about that particular circuit, including Tark's mod suggestions and critiques, but it wasn't until duhvoodooman's comment
https://www.byocelectronics.com/board/viewtopic.php?f=29&t=60399#p504666
and pointer to
https://www.byocelectronics.com/board/viewtopic.php?p=352826#p352239
that I realized there had been a similar observation made concerning the regeneration path in that circuit. Tark only showed SPICE simulations but I think made an appropriate point about how phase shifts along a signal path need to be considered when trying to figure out what effect feedback will have on a scheme in which it is essentially phase relationships in summed signals that lead to the filtering that makes the effect.

I think I've done about as much as I can before needing to resort to more surgery, so here's an update.

(1) I'm taking a feed from the emitter of transistor Q4 (Q4E), per the original suggestion. I have tried injecting this, coupled through a 0.1 uF cap (and a resistance in series) to the top of a pot ("regeneration control"), with its foot at GND, and the wiper coupled through a resistor and 0.1 uF cap to several places further upstream. I have done this more systematically than in what I described earlier, and the only really promising injection point is the base of transistor Q2 (Q2B). This feedback loop is in accord with what I see in other flanger schematics, both from the point of view of crude relative phase between the two circuit points, and also from the point of view of injecting a low-impedance signal into a high-impedance juncture.

(2) Without some damping on the fed back signal, the circuit oscillates, according to the amount of delay dialed in. To suppress this, you can keep yourself from turning the pot wiper all the way to its upper end (this is essentially what the original suggestion implements, with a fixed or trimmable R between Q4E and the top of the pot and in series with one cap), or else you can put some resistance in series with the pot wiper on the way toward the second cap and Q2B. The latter method is what is typically found in other flanger schematics. I did not find any obvious difference between using 50k and using 100k for the REGEN pot. For both damping topologies, it was necessary to have at least 3-5 kohm in place, but it's sufficient to use just one or the other. After I make some other changes (that will require surgery) I may end up revising that R_damp upward.

(3) I can confirm that a "C" taper is preferable. You need to have the pot turned most of the way up to hear a strong effect. I think this argues for having the damping resistance connected to the wiper rather than the top, but it's not a big deal. With the current state of my breadboarding, the effect is so slight for the CCW half of the REGEN control pot that it may make sense to put a large fixed resistance between GND and the "bottom" of the pot. At present I've got a C50k pot sitting atop a 47k resistor to GND.

Everybody likes pictures, so here we go. The long breadboard gives me a selection of injection points (most of which I am about to discard), and also has the chorus mod on it. The lower breadboard has the flanger mod on it.





I find that the slowest "RATE" I can dial (full CCW) is about 2 Hz, which to me is too fast for a flanger. As a result, I am not really certain that the modulation effect I am hearing is what I was looking for. Additionally, the character of the modulation depends verrrry strongly on the setting of the DELAY trimpot (which is not at all surprising; and I think the shorter end, CCW, is supposed to be the desirable end) and on keeping the "DEPTH" knob nearly full CCW. So what I intend to do next is implement the "Slow" part of the "Fast/Slow/TapeWarble" mod that SteveO apparently implemented successfully. While I am in there I might replace the 500k ("474") DELAY trimpot with a 1M trimpot, hoping to get some longer delays, which I think I might prefer in the chorus mod. FWIW I believe I measured the range of delay for the stock trimpot and R48 (47k) combo as from about 1 ms to about 8.5 ms (edit; originally I had written "0.3 ms to 8.5 ms" but I can't replicate that lower end). That actually sounds about right for flangers, but maybe is short for some chorus versions.

Obviously I am not accomplishing this stuff very fast, so if anyone want to make suggestions, feel free--I could perhaps do some other experiments too.

You could try replacing the V3207 with a V3208 if you want longer delay times. Might be easier than swapping pots. If that gives you the desired effect, then maybe switch the pot. However, when you wind out the delay time, you also increase the noise floor. So if longer delay time is indeed what you need, then I would even more strongly suggest using a bigger BBD.

_________________
*patience is a virtue*

Please do not PM me. email is prefered. keith@buildyourownclone.com