DOD 640 Flanger

[miketbass]

Hello, I've picked up an old DOD 640 flanger. It was a one owner pedal bought from an estate sale and is in nice condition. I did the basic checks like I would with all BBD effects, measured clock frequency (50hz at low) and set the bias. Nothing seems amiss.

Anyways to the point, the "accent" control acts strangely throughout its rotation. Full CCW is a distinct flanger, and as I rotate the wet drops much quieter through the sweep up to full CW when we get the flanger with resonance cranked sound. It seems that either end of the rotation does what you would expect. Between the two extremes it sounds much more subtle and chorus like, but does continue to change throughout the rotation. The wet signal is present at all points but almost seems to "cancel out" and fade.

It is very difficult to find information on this pedal and my hunch is that this may be the way the circuit functions. I am not very familiar with the particular chip 4136 or its arrangement in the "accent" function but looking at the schematic it is designed differently than your run of the mill resonance knob which just controls how much of the signal is fed back to the BBD. Is this a strange panning function very unique to the 640 flanger? I have verified the pot via point to point and checking its functionality with a meter and it's working as it should. Any hot takes on what's going on here?

[ElectricDruid]

That schematic is a tangle which I can't untangle right now, but I'm willing to make a guess what it's doing.

You get a different sound out of a flanger depending on whether the feedback is in phase or out of phase with the original signal. I once messed about trying to do this on a single knob with a centre-zero, so turning one way gave you more negative feedback, and the other way gave you more positive feedback. I was never happy with it because the two different types of feedback gave rather different levels and hit oscillation at different points, so it was hard to tune it to get the zero actually in the centre. However, DOD might have done rather better than I managed, so I'm guessing that's what it's doing. The circuit is certainly feasible for such a thing, without analysing it closely.

[miketbass]

Thanks ED, you confirm something I've been thinking about. It appears that the clean signal is being panned throughout the pot rotation, which makes sense as the effect is most subtle with the accent knob centered. Listening back, the sounds at either extremes do sound like they could be negative/positive feedback. The CCW position is a much more pronounced effect than one devoid of any feedback. It would seem that the pot pans clean signal blend across a fixed value negative CCW and positive CW feedback path. At least that's my hypothesis. I wonder why this isn't used more? It definitely allows for a very wide range of sounds.

Also to amend my OP, the low frequency of the clock measured at the 4013 is really 50kHz, not 50hz.

[Govmnt_Lacky]

That schematic is a tangle which I can't untangle right now...

Hey..Hey..Hey..!!!

That schematic is a hand-drawn work of art there mister! 

[miketbass]

That schematic is a tangle which I can't untangle right now...

Hey..Hey..Hey..!!!

That schematic is a hand-drawn work of art there mister! 

GL, can you confirm that this is how the circuit behaves, considering you are the artist behind the schematic?

[thermionix]

Hey..Hey..Hey..!!!

That schematic is a hand-drawn work of art there mister! 

Might wanna redraw that bridge rectifier though.

[ElectricDruid]

That schematic is a tangle which I can't untangle right now...

Hey..Hey..Hey..!!!

That schematic is a hand-drawn work of art there mister! 

The *drawing* is beautiful, but it's the *topology* I can't get my head around right now...and that's DODs fault, not yours.

[Rob Strand]

Firstly,
- The point (A) needs to connect to a voltage divider to the supply and cap to ground.
The usual Vcc/2 thing  [EDIT: OK found it 22k's at sort of top left when schem in normal rotation]
- To me, the circuit largely follows the MXR flanger. The low-pass filter at the output is different.

Any hot takes on what's going on here?

The accent control reduces the amount of clean signal mixing back at the output
and increases the amount of resonance.  The level of the delayed signal is
left constant.  With Accent on max there is no clean signal mixing with the delay signal
and the resonant feedback on full so you end-up with "vibrato" and resonance.
With Accent on min. there is no resonance and the clean signal is mixed back with
the delayed signal in roughly normal proportions, perhaps with a little more clean than usual.

There is no "normal" resonance mode.   At mid positions it's kind of a pot-luck combination
of resonance combined with less dry signal (ie. more vibrato-ish)

For the 4136 IC, which was a common opamp in the day:
- The clean signal comes from output IC pin 10
- The clean signal passes through to the output mixer via the 10k+47k to IC pin 1
- The delayed signal comes from output IC pin 12.
   It mixes at IC pin 1 and is not affected by the Accent control.
- The resonance feedback input is the 47k on IC pin 8
- The output signal passes back to the resonance input via thr 18k + 47k resistors (to IC pin 8 )

The thing to realize is point (A) is like ground and the wiper of the Accent pot is basically this pan circuit except there are two inputs and two outputs.  The Accent pot block signals to ground at each extreme.
http://320volt.com/wp-content/uploads/2011/06/mono-ses-sinyalini-stereo-yapan-pan-balans-devresi.png

[miketbass]

Thank you for your input, I will take some time to study the technical points you have layed out for me. It is much appreciated. The effect is extremely subtle in the middle range of the pot and I suppose without hearing you may not be able to tell me, but I am worried that my unit is not functioning correctly. All of the pronounced effect is at the two extremes of the accent pot rotation and according to your input it sounds like this should not be the case. I will have a closer listen later tonight but I am curious if I have had a component failure that is causing this to not act correctly.

[Rob Strand]

Thank you for your input, I will take some time to study the technical points you have layed out for me. It is much appreciated. The effect is extremely subtle in the middle range of the pot and I suppose without hearing you may not be able to tell me, but I am worried that my unit is not functioning correctly. All of the pronounced effect is at the two extremes of the accent pot rotation and according to your input it sounds like this should not be the case. I will have a closer listen later tonight but I am curious if I have had a component failure that is causing this to not act correctly.

If the extremes work I'd guess it is actually working.

I just noticed something.   The resonance is output (or input) out of phase.   This is highly likely to produce the weirdness.

If you look at the MXR Flanger (or VOX Flanger)  the overall signal is inverted  but the "internal" signals are all in-phase.     The Boss Flanger keeps the overall signal in-phase but again the "internal" signals are all in-phase.

The key thing to note about the DOD 640 is the filter is non-inverting whereas the on the MXR and Boss they filter is inverting.    Boss have re-jigged the phasing to fix the MXR's over phase inversion issue.    It looks like the DOD 640 tries to fix the over-all phase inversion issue as well however the crack in their solution is the regeneration/resonance/feedback  has the wrong phasing.     The way that would be fixed is to feed the signal back to pin 9 of the 4136 IC so as to provide the correct phasing.    It's not that simple bu that's the idea. I'm sure there's a Flanger out there that implements this scheme, I just can't remember which one it was.

[TheKoala]

Hello there, just made an account to thank miketbass for tracing the schematics ! Got a DOD Flanger 640 recently and I repaired it/modded it to my taste so I thought I could help future people having this flanger to change a few parameters.

First thin is, as noted by some of you, the Accent CCW position is much louder than the CW position. The Accent pot center lug is connected to the VCC/2 voltage level so it is basically AC ground. Yup, the way it works is grounding more or less either the sumed clean/flanged signal or compute the difference of both signals, so that's just how you want to add the flanged signal to the clean signal. In the center position, you can something like half clean+flanged, half clean-flanged, so the sum of both would only give the clean signal. You can see on the schematics that the accent knob is tied to the opamp (+) input through a 47k resistor, and to the (-) input through a 75k resistor. This unbalance does affect the gain of the CW position. Paralleling a trimmer to the 75k resistor allows us to adjust the level of the CW position to match the level of the CCW position.

Second mod is related to the amount of feedback. When the Accent know is fully CW, the flanging effect is way more subtle than the CCW position. If you follow the accent pot wiring going up to the input stage, you see that the opamp has two 47k resistors : one in the feedback loop of the opamp for gain purposes and the other one coming from the flanged signal. Paralleling a trimmer with the latter resistor allows us to adjust the amount of feedback, helping to achieve some "jet flanger" sounds and making the Accent pot CW position way more versatile. Be careful not to short-circuit the resistor though, because this would lead to an increased volume and complete oscillation !

Also, the switch is just attenuating the input signal and increasing the output signal level to match keyboards and other instruments (Normal vs P.A., meaning pre-amplified ?). Just thought about adding this to my post.

[beedoola]

I just got one of these. Is there a guide to adjusting the trimmers for the bias and clock? Would I need a scope to do that or can it be adjusted by taking voltage measurements?

[Rob Strand]

I just got one of these. Is there a guide to adjusting the trimmers for the bias and clock? Would I need a scope to do that or can it be adjusted by taking voltage measurements?

To set the bias typically you would input a 1V peak sinewave from function generator (or PC or phone) at say 400Hz and observe the signal at pin 8 of the LM324 with an oscilloscope.  Adjust the bias pot for symmetrically clipping.  If there is too much clipping back-off the input level and you can't see any clipping at all increase the input level.

A much less precise method is to do it by ear.  Input a signal where you can hear clipping, back off the level so the clipping just disappears, now rotate the bias put up and down until you can hear clipping at both up and down positions.   Set the bias pot at the center of those two positions.

For the clock frequency you really need to know the delay specs for the unit.   Sometimes the published specs don't represent the built units.   You might also try to hunt down some old posts on this forum.

Typically the way the clock frequency is set is:
- Set the Width pot to minimum
- Set the Manual pot to minimum
  Using an oscilloscope, or frequency meter, measure the clock at pin 12 or 13 of the CD4013 chip.
  Adjust the clock to equal the maximum delay spec
  clock period = 1/clock frequency = delay time  / 512
  eg.  10ms => 20us (50kHz) , 13ms => 25us (40kHz), 15ms => 30us (33kHz), 20ms (25kHz)
- Set the Width pot to maximum, measure the clock (same pins).
  Check the clock frequency isn't too far the frequency for the minimum delay spec.

  If you use a DMM to measure frequency it is likely not to measure the upper clock frequency.

If you look at the to of the thread miketbass quotes a minimum clock frequency of 50kHz.
What's missing is if the clock frequency is measured at the input of the CD4013 (pin 11)
or at the output (pin 12 or pin 13).   The frequency on Pins 12 and 13 is half that on pin 11.
50kHz could be a either position pin, it's ambiguous where the frequency is measured.

It's best to dig-up the delay specs on this forum or elsewhere.   I'm pretty sure there's a thread
with details posted in the last three months.

You can set the maximum delay by ear but you need to have a good idea what the delay in ms sounds like.

If you have to guess go for 10ms max delay, or 15ms max delay.

[miketbass]

Hello, the clock frequency was taken from the output of the chip/clock pins of the SAD1024. It's been a few years but I would have used my Flike meter to measure frequency at this location like I do for all of my time based effects.