Mutron Flanger dissection.

[StephenGiles]

You need dirt for a great sounding flanger - thus the Electric Mistress and ADA sound so good!

[Derringer]

Great job!

Gut shots when you get the time?? And you know what would be very useful.....I don't believe I've ever seen any documentation of the foot pedal shutter and surrounding mechanics. A drawing with dimensions would be great!

I like things that go woosh!  

If you look on the "other" site in the vintage effects section, there's a whole dissection of the optical pedal that goes with the Mutron Bi-Phase
It's got to be very similar to what's going on in this flanger's pedal.

[armdnrdy]

If you look on the "other" site in the vintage effects section, there's a whole dissection of the optical pedal that goes with the Mutron Bi-Phase
It's got to be very similar to what's going on in this flanger's pedal.

Yeah...I have those files on my PC. I took a look at them last night and compared the info to the flanger pedal. It is similar but....one thing that's missing is the measurements of the shutter. I think that it would not have the same effect with a modified shutter and with different tolerances between the LDRs.

[digi2t]

Alright guys, more measurements.

SAD1024 frequencies from pins 3-14 (8-10 are identical);
MIN: 63kHz
MAX: 537kHz

LFO frequency
MIN: 0Hz (or pretty damn close to it)
MAX: 11.1Hz

I tested the LDR resistances, and I found them to be in the same range for both LDR's. Mind you, my testing was rather unscientific. In a dark room, I shone a bright white LED flashlight into each LDR, and recorded the light and dark resistances. Both LDR's gave me the same ranges, with the same rates of end to end resistance speed. What I got was;

MIN (light): 18K, but could be lower because I could not shine the light directly into the tube. It was at an angle.
MAX (dark): over 40M, my DMM doesn't go higher than this.

Like I said, but LDR's gave me the same readings.

Here's a quick sketch of the tubes, and the arrangement of the LED and LDR's;



I'm thinking that the LDR's that we used for the EH Talking pedal conversion (Waitrony KE-10720) might be suitable for the job. I fooled around with one, and although it gives a lower light resistance, about 5K in similar conditions, the speed up to 40M dark is pretty similar to the originals. Besides, this could be fine tuned with resistors, LED selection, LRD angle, and there is a range trimmer as well. Nothing insurmountable here.

[Dave W]

There's a lot of room to play with; higher and lower; with the clock f's....
Thanx for all this work Dino! 

[digi2t]

There's a lot of room to play with; higher and lower; with the clock f's....
Thanx for all this work Dino! 

Looking at it now, and looking at the ADA flanger docs, I realize that. As I mentioned before, when Mark and I fired it up, it was poo, and what you hear in the video is my tuning by ear. I know now that two of the trimmers adjust the high and low end, and I have it adjusted to what my ears were telling me was "decent flanging". Of course, with Fender3D's direction, I can now scope it, and tune it even better. Up until now, I was only using the scope to check waveforms and voltages, this is my first frequency probe. Thanks for helping me bust my cherry on that one.

[12Bass]

Yes and no.  In truth, many, if not most, of the notches created at the "top" of the sweep (i.e., at shortest delay) are inaudible, largely because they are beyond the range of the guitar and most guitar-amp speakers.  So whether they are compromised because the BBD output drops a db or two or even three, once it heads for the <.4msec range, can often be timbrally moot.  I suppose one would likely notice it if playing white noise or something with similarly wide bandwidth through a wide bandwidth system (e.g., studio monitors), but not likely with the typical bass or guitar rig.

The expected question here is: so why the heck aim for delays that short?  And the answer is that dramatic slow sweeps have to feel like watchng slow-mo video of a car airbag inflating.  One moment there's absolutely no trace, and all of a sudden the car cabin is filling right up.  The transition from no audible notches to notches absolutely everywhere is a big part of the "drama" of flanging.  That nothing-to-everywhere traisition is a result of the overall sweep ratio (max delay divided by min), so both "start" and "stop" frequencies/delays play a role.  But the more critical one seems to be the minimum delay.  All other things being equal,  a flanger that seeeps from .5msec to 10msec will seem more "dramatic" than one that sweeps from 1msec to 20msec.

Sure... the 0.5 to 10 millisecond range covers the "sweet" zone of the sweep, IMO.  Flangers that cannot produce delays shorter than 1 millisecond don't seem to get up high enough and really lack that dramatic swoosh when they start sweeping back down.  Around 0.3 ms seems short enough in my experience.

My point was more about how the decrease in BBD gain with increased clock frequency affects the comb filtering.  The most profound (deepest) flanging effect occurs with an exact 1:1 mix of delayed and straight signals.  When the BBD gain drops off at higher clock frequencies, this alters the mix ratio from the ideal and makes the upper part of the sweep less exciting.  I've found this gain decrease to be a bit of an issue even when running the SAD1024A in parallel-multiplex.  Right when the flanger starts sweeping up into the upper midrange, the BBD gain starts to drop noticeably and the effect loses impact.  This is an area of the sweep that I happen to like a lot, so I've tuned my A/DA so that the delay/straight mix is 1:1 in this range.  Running the SAD1024A in series would only exacerbate the problem at higher clock rates, leading to a weaker sounding flange in the upper range.

Just want to add this for thought....
IMHO; a big part of the "drama" in the upper part of the sweep of the A/DA; aside from all the points that have been mentioned; is due to it's bare bones minimal post BBD low pass filtering. You can go up, up, up; but if you filter out too much sonic information those short delays are just not going to sound the way one would want them to.
Just my 2 cents from a hack with a soldering iron and a 'scope....

Agree on the minimal low pass filtering.  I've gone even further with my build, allowing practically full 20 kHz bandwidth in the delay path.  I tend to use my A/DA through a custom full-range 3-way fEARful bass enclosure.  The intensity of the A/DA's high frequency comb filtering is not lost through this monitoring system.  Looking forward to an MN3007 version of the Mu-Tron in the future!

[Fender3D]

Alright guys, more measurements.

Thanks Dino

[Mark Hammer]

I'd have to read through all those articles to be accurate about it, but in the various scanned issues of DEVICE on my site (hammer.ampage.org, pages 10 and 11), I think it is the late Steven St. Croix who proposes that a decent flanger needs to have at least a 40:1 sweep; something which Craig Anderton reiterates in the Hyperflange article.

Again, two ways to get that: extending the maximum delay time, or decreasing the minimum delay time.  The former gets you into the challenges of a) requiring a slow enough clock frequency that the clock becomes more audible, and b) moving into the delay range where pitch deviations become more obvious.  The latter imposes the burden of buffering the clock signal, but that is much less of a problem with the SAD1024 than it is with the Panasonic chips. That caveat aside,  every little bit one can dip below 1msec delay has a much greater impact on the sweep ratio, than adding a msec or two at the other end.

[armdnrdy]

Very useful info!

Thanks Dino!

[Gus]

Looking at the schematic I noted some things

The Z2B and Q1 and Z2A based circuit fragments look to be some kind of noise gate.  Z2B is a full wave rectifier to a R10, C8 attack, C8, R11 decay Q1 fet switch.  Z2A is a lowpass filter.

The 4046 is being controlled via pin 9 look at the VCO part
http://www.ti.com/lit/an/scha003b/scha003b.pdf

The foot pedal is a buffered voltage divider  Note R72 and R62 are the same value and R61 and R71 are the same value

The buffered voltage divider output of Z6A is not set at 1/2 +7.5VDC and -7.5VDC but more to +7.5VDC

Now the tricky thing is what is ground in the schematic?

Is ground the same set by the output of Z11B or is it a mix of grounding?

The 15VDC supply is split by the Z11B and passive parts circuit fragment and the ground looks to be the voltage divider output that is buffered by Z11B.  Not sure what C50 is about.

[digi2t]

Alright, I have a couple of questions....

1) I noodled around with it last night, and I found that the best sounding frequency range is between 50kHz to 800kHz. I can't get it to go higher than about 800 and change, and in any case, I can't hear any flanging higher than that (though my dog probably can). Below 50kHz, the sound starts to warp too much, giving me an almost "boing" sound, which really isn't useful. Looking at some tuning procedures/video of other flangers, I can't get into the MHz zone with this unit. I'm close, and my ears are telling me it's pointless, but does it matter?

2) Is this unit capable of thru-zero, or, do you need to be running in parallel to achieve thru-zero? Looking at the schematic, I see that the SAD-1024 is running in series... I think. I'm trying to be as educated with my guess as a hack can possibly be.

Bonus question) Why don't I have +/-15vdc in this circuit, like the schematic implies? The whole thing is running at +/-7.5vdc. Bad caps?

[12Bass]

1) I noodled around with it last night, and I found that the best sounding frequency range is between 50kHz to 800kHz. I can't get it to go higher than about 800 and change, and in any case, I can't hear any flanging higher than that (though my dog probably can). Below 50kHz, the sound starts to warp too much, giving me an almost "boing" sound, which really isn't useful. Looking at some tuning procedures/video of other flangers, I can't get into the MHz zone with this unit. I'm close, and my ears are telling me it's pointless, but does it matter?

Does it matter?  Probably not.  If it sounds good, it is good, IMO.  Clock much higher and the BBD signal will drop off rapidly anyway, making the comb filtering less intense.  There's not a lot of spectral content to flange up there anyway.  For me, the best sounding range is between around 0.5 ms to 5 ms.

2) Is this unit capable of thru-zero, or, do you need to be running in parallel to achieve thru-zero? Looking at the schematic, I see that the SAD-1024 is running in series... I think. I'm trying to be as educated with my guess as a hack can possibly be.

Not capable of through-zero flanging (TZF), at least not without adding another short delay line in parallel and sweeping against that.  To go through zero, the delayed signal would have to sweep ahead of the straight signal, then behind it again.  

[digi2t]

I agree whole heartedly on the "sounds good = is good" front. I mean even at 800kHz, it seems a bit high/weak to my ears, but I can always trim that back with the pot, so that's fine. Besides, if I use it with keyboards, pumping lower frequencies (than guitar) through it might be fine at 800kHz. I prefer to have the maximum usable range available, and then cut back from there.

That's what I thought concerning the TZF. Like I said, this is brand new territory for me, but I kind of figured that it would require parallel signals, not series. Since the whole concept was derived from parallel reel to reels, it only made sense in my noggin.

Thanks!

[Fender3D]

Bonus question) Why don't I have +/-15vdc in this circuit, like the schematic implies? The whole thing is running at +/-7.5vdc. Bad caps?

'cause designer made a mistake 
you have just one 15V regulator, then you have 15V from positive to negative
Z11b will provide half supply (GND).

BTW
you did invert labels @ power supply... 

@Gus

Z6a simply set LFO reference, not circuit GND.

The Z2B and Q1 and Z2A based circuit fragments look to be some kind of noise gate.  Z2B is a full wave rectifier to a R10, C8 attack, C8, R11 decay Q1 fet switch.  Z2A is a lowpass filter.

ADA docet...

@12Bass

I wouldn't matter so much for BBD signal dimming...
Every analog flanger ever built suffers (well, TCE aside  ) from this...
We're used to, aren't we?

[digi2t]

'cause designer made a mistake  icon_mrgreen
you have just one 15V regulator, then you have 15V from positive to negative
Z11b will provide half supply (GND).

OK, makes sense.

BTW
you did invert labels @ power supply...  icon_wink

Do you mean, why I flipped the arrows for the +/- 7.5vdc? It's because the schematic convention states arrow up = +, and arrow down = -. Yet, taking my readings from the board, the power section did not reflect the original schematic. Pin 2 gives -, and the original schematic showed arrow up, and vice versa. I found this confusing, so I amended the schematic.

Now...

Can anyone point me in the right direction as to how to graft an MN3007 in place of the SAD1024? Will it work with the 4011 providing the clock signals? As Jimi would say, I'm outta my pay grade here.

EDIT: Actually, is it possible to use the Moosapotamus 3007 daughter board here? Take the clock signals from the 4011, and plug the in/out between C17 and C19?

[Fender3D]

Do you mean, why I flipped the arrows for the +/- 7.5vdc? It's because the schematic convention states arrow up = +, and arrow down = -. Yet, taking my readings from the board, the power section did not reflect the original schematic. Pin 2 gives -, and the original schematic showed arrow up, and vice versa. I found this confusing, so I amended the schematic.

My bad, I just stared regulator pins...
Yep, another designer mistake...

Now...

Can anyone point me in the right direction as to how to graft an MN3007 in place of the SAD1024? Will it work with the 4011 providing the clock signals? As Jimi would say, I'm outta my pay grade here.

Check this http://www.diystompboxes.com/smfforum/index.php?topic=79154.msg652910#msg652910

Pay attention:
MN30xx needs an output pull-up resistor instead of SAD's pull-down.
You'll need Vgg (SAD hasn't it) @ pin4

Dissect my schematic and you should be done...

[digi2t]

OK, thanks, I hadn't seen that one yet.

Do I still retain the original 4011 for the clock signals?

Is the pull up resistor R8 on your schematic? Can I use a trimmer like R80 on the Mutron, but to positive?

[Govmnt_Lacky]

So is this circuit working from 0 to +15V with a +7.5V reference?

Or... Is it working from +7.5V to -7.5V with 0V as reference?

[armdnrdy]

This power supply isn't making much sense to me.

Something's rotten in Denmark!

I don't know how one would go about getting a negative voltage from the output of a positive voltage regulator, (MC7815CT)
or 7.5 volts out of a 15 volt regulator, or positive 7.5 volts from the center tap. What am I missing?

By looking at this power supply schematic the output voltage should be +15 volts and ground. Then there is the schematic section to the left of the PS that looks like a virtual ground reference.