FL301 Flanger on Tonepad, is it possible to... (RG, Mark?)

[RedHouse]

I'm looking into building a Flanger to stuff into an old wah case and would like the manual sweep option to use as a special effect.

1.) the FL301 Flanger over on Tonepad, is it possible to add a manual sweep ctrl to it?

2.) alternatively, I've seen the Ultra Flanger designed by John Hollis which does have a manual sweep ctrl, but noticed many who built it reported noise issues (ticking?), but I've also noticed some revised PDF's of the Ultra Flanger which mentions decoupling the digital noise from the flanger, anyone know if that was the cure to the noise complaints?

[gorohon]

Where have you seen these revised Ultra Flanger PDF's?  I'd like to look at them.  I built a UF.  More tick tock than a grandfather clock.  Sorry to deviate from the topic.  I've got nothing for you on the 301 flanger.

[Mark Hammer]

Followup on the Ultraflanger resides at GEOFEX.

The introduction of manual flanging adjustments to the FL-301 should not, in principal, increase audible ticking.  I qualify my response with "in principle" because I have no way of knowing about your layout and where you run your wires.  As you'll learn below, one of the wires to be routed carries the LFO signal, and running it too close to the wrong thing *could* result in some audio bleedthrough of LFO-related ticking.

The general principal of manual control is that the clock generator (MN3101 or CD4046) is responding to a DC voltage.  That DC voltage could be coming from the LFO (DC from an LFO?  Sure.  Think of it as an always positive voltage that just happens to change a bit over time.), but it could *also* be coming from a steady DC source like the power supply.

The most commonly found arrangement across both DIY and commercial products is that the LFO depth/width and manual delay controls interact.  Within such an arrangement the depth control, which normally serves as simply an attenuator (volume control) for the output of the LFO, gets transformed into a panning blend control that pans between the LFO and the steady DC voltage.  The more you pan towards the steady DC, the less LFO you get, and the more LFO you dial in, the less impact the steady DC can have.

Why does it do this?  Simple.  Because the DC voltage going to the clock generator should never go higher than a certain level.  The LFO is preset to swing by a certain amount.  By not having this tradeoff, there is the risk of summing the DC voltage to the maximum swing of the LFO and that will just confuse things.  I'm not the expert here, but my sense is that if you add a DC offset to an LFO and don't attenuate the LFO, you end up with a clock that sweeps up and down for one ened of the sweep cycle, but reaches a plateau and lingers unchanging for a while on the other end of the sweep cycle.  Not good.

Having the depth control be a "tradeoff" control eliminates any possibility of a DC voltage ever being mixed in with a max-depth LFO.  Conversely, it eliminates any possibility of the user doing a broad manual retune and offsetting the LFO so much that it screws up the sweep.

As near as I can tell, the 10k clock-adjust trimpot on the board (VR101) functions the same way as the manual control on a great many flangers.  It taps the supply voltage and divides it down to provide a DC voltage to fine-tune what the clock generator sees.  The node of R146, 149 and 148 is the mixing node of the LFO+DC sources.  R146 and R149 are essentially the "mixing resistors.

You will note that while VR101 is there to be able to inject some DC offset, there is no foolproof tradeoff depth control like you'd see on the BF-2, for instance.

This leaves two alternatives.  One is that you simply mount a 10k pot on the chassis to replace VR101 and use good judgment and your ears to adjust appropriate amounts of DC and LFO.  A second is that you set the width control to max and figure out some sort of panning tradeoff control to go where R149/R146 are.  Insomuch as there is less R&D work involved, my recommendation is simply to relocate VR101 to the chassis.  

If you want to decrease the likelihood of catastrophe, I would suggest trying the following:

1) Set LFO depth to minimum.
2) Retune VR101 and listen to the impact on delay and notch location.
3) When you've reached about as high as you'd want to tune it, measure the resistance from the trimpot wiper to the relevant tie point (and I honestly forget whether this will be V+ or ground).
4) Now readjust manual delay to about as low as you'd ever want to tune it, and measure the resistance between the wiper the OTHER relevant tie point.  (If ground for #3, then V+ for this one, and vice versa).

Why are we doing this?  To make sure the optimal pot value is used that will insure meaningful and nonproblematic adjustments across the full rotation (or at least most of it).  

Let's assume that we want to aim for the same overall resistance between V+ and ground that R147 and VR101 currently have (11k).  If the resistance measured  at maximum useful VR101 rotation between the wiper and V+ is, say 4.3k, and the resistance measured between the wiper and ground is 2.7k with VR101 rotated as much in the direction of ground as you find useful, then we have 11k - (4.3+2.7) = 4k of usable resistance adjustment range to work with.  The nearest standard pot values are going to be 5k (4k7) and 2k.  So, to make VR101 into a chassis-mount pot for manual control, with as much of the pot's rotation as possible providing useful adjustments, pick a 5k pot.  I don't think linear vs log makes too much difference here, but let your ears guide you.  

Now, depending on which side the critical adjustments are, select either a 2.7 OR a 4.3k fixed resistor for that side, and then make up the difference to 11k with another fixed resistor (or combo).

So, assuming that tuning VR101 towards ground slows down the clock (i.e., tunes the notches lower in the spectrum), and that the lowest usable wiper-to-ground resistance is 2.7k, you'd stick with a 2.7k resistor, and hang a 11- (2.7+4.7)= 3.6k on the V+ side.

Bear in mind, this is AN ILLUSTRATION and not necessarily the values you HAVE to use.  Also note that you will still have to use good judgment in adjusting things.  Should work, though.

[ethrbunny]

Errr.. ok..  :shock:  ...umm... so that will get rid of the ticking?

[RedHouse]

Thanks Mark.

About that updated UltraFlanger PDF, is that a fix for the ticking problem?

[gorohon]

When I first heard of the "updated" version on this thread I thought that it was updated from mine.  It's not.  I have ticking,  but I think that if I try to layout the wires better and use sheilded wire it will reduce the tick.  When I use my boutique fuzz in front of the UF it is quiet, but that is not fixing the problem with the flanger.   Despite the unwanted noise, it flanges nice.  I get a little distortion though using the (-) feedback selection.  I like the (-) feedback sound.  I find that the UF (-) feedback isn't very pronounced with it until you reach a certain point, then you go just a little too far and it self occilates.  The (+) feedback selection is quite swooshy and jet like.  I guess the UF was supposed to be a minimum parts A/DA flanger-alike, but I have no idea if it even sounds like an A/DA.  I guess all of those extra parts make a huge difference in performance.

[Fp-www.Tonepad.com]

The 301 Flanger project on tonepad doesn't tick at all.

It's a very nice flanger IMO.

I'd say... build them both if you can!

Fp

[Mark Hammer]

My FL301 board is etched and populated, with a bunch of wires and pots hanging off it.  I have yet to install it into a chassis so I can't comment on the sound.

What I wrote, though, had nothing to do with ticking.  It was entirely about the issue of being able to do manual *initial delay* adjustments as are done on other commercial units that have such a control in the form of knobbed pot.

As far as I've been able to understand it, ticking in LFO-modulated pedals comes from a couple of sources.  One source is layout, where the tick that does exist is somehow "sensed" by the audio path or some portion of it running too near the control path.  The other source revolves around the decoupling (or lack thereof) of the semiconductors in the control and audio portions.

Keep this fact uppermost in your mind at all times: THE LFO IN ANY MODULATED PEDAL IS GENERALLY A SQUARE WAVE GENERATOR.  The square wave produced by one half the LFO circuit is smoothed out by the other op-amp in the circuit so that it *behaves* like a triangle.  The fact that it starts out life as a square wave is important.  Why?  Because a square wave implies a very sudden and jarring swing in voltage from one extreme to the other, and consequently the current required to do that.  When such instantaneous current draws come from a source SHARED by multiple semiconductors, what happens to one is experienced by the others.  It's like everyone seated at the dinner table, and your 6'8" football player cousin sits down and declares "Man, am I starving!!" and eats just about everything there.  Everyone else at the table goes to bed with an empty feeling in their stomach.

The solution to this is to "decouple" what is supplied to the different chips, such that they have their own "secure" current source to tide them through these momentary current draws from other parts of the circuit.  Whenever you see a subcircuit in a schem for the "Vref", and it has two equal-value resistors and an electrolytic cap from Vref to ground, that cap is there for decoupling purposes.  The two resistors may be enough to divide down the supply voltage to extract V+/2, but the cap is there to store up charge and assure that Vref remains rock solid when there is a sudden current draw  on that puny little battery from somewhere else in the circuit.

Inference: ONE of the solutions to reducing ticking is to provide separate Vrefs for different parts of the circuit, each with their own decoupling cap.

[RedHouse]

Yes thanks Mark, I understood your first post, mine was a 2-part question about the manual ctrl of the FL301 -and- the fix for the UltraFlanger, the reason I was wondering about the PDF is because of RG's verbiage on the second page of the PDF (bottom right corner "Notes on the Layout") where he says:

...I've added a couple things to the design, R27/C11 decouples digital noise from U3 and U4...

and I had understood that was the complaint folks had with it.
(I wish RG would chime-in here since he wrote it)

Anyway I want to be fairly certain of either/both these questions of mine before I go etching up a board and buying parts etc. It would be nice to have the luxury of extra time to breadboard everything and deciding but I've been so busy these days yaddayaddayadda.

Some day (when I have some time) I'd like to investigate using the XR2206 for the LFO on both a Flanger and a Univibe, it has Sine, Triangle and Square wave outputs available, adjustabe symmetry and easily adjustable speed.
(the square would be handy to drive a buffered LED for speed indication)

[RedHouse]

(bump)