News:

SMF for DIYStompboxes.com!

Main Menu

Ultra Flanger

Started by Arfman, September 18, 2007, 08:52:52 AM

Previous topic - Next topic

Arfman

Hi, this is my first post here. I signed up a couple of weeks ago and have been lurking ever since. I consider myself to be at an intermediate level with respects to pedal and amp building...still have a *lot* to learn. I mostly etch my own PCBs and have done some elementary distortion boxes from scratch, though the circuits were not particularly innovative. All told I've built over 20 stomp boxes.

I have just finished the circuit for John Hollis' Ultra Flanger using the latest project file from GEO. I'm getting a faint ticking sound and in searching the forum here that seems to be a fairly common problem. Sensitivity to extraneous noise sources, like fluorescent lights, certainly seem to exacerbate the problem, and largely I can dial in a good setting where it's hardly audible. Adjusting the trim pots helps (and I did install the 5K trimmer and cut the pad traces).

Mark Hammer (did I get the name right?) has an article on his site talking about a quite similar problem with the Zombie Chorus due to the Vref on the delay chip (I think). The fix was quite simple by installing a voltage divider instead of using the LED's for that chip. I'm just now wiring up a Zombie Chorus, coincidently and will probably be able to start testing tonight.

Would the same fix for the Zombie Chorus work for the ultra flanger?

Thanks and I'm glad to be here...from what I can see this appears to be a really good board w/good people and a great culture that encourages camaraderie and frowns on egocentric behavior....

mdh

QuoteMark Hammer (did I get the name right?) has an article on his site talking about a quite similar problem with the Zombie Chorus due to the Vref on the delay chip (I think). The fix was quite simple by installing a voltage divider instead of using the LED's for that chip. I'm just now wiring up a Zombie Chorus, coincidently and will probably be able to start testing tonight.

Would the same fix for the Zombie Chorus work for the ultra flanger?

Am I missing something, or isn't that how the Ultra Flanger is set up already?  The LFO and the audio portion of the circuit have independent Vrefs: a zener or LED reference for the LFO and a voltage divider for the audio path.  Have you boxed it up yet, or is it a naked circuit board with long leads running all over the place?  Could be that it quiets down when it's in a metal box with short leads.

Mark Hammer

The Zombie fix I posted is really only a partial fix.

There are a few separate matters to be addressed:

1) Using a single source to both bias the BBD chip and use as a reference voltage for biasing op-amps may "work", may save parts and cost, but is a compromise and a bad idea.  The two biases are different voltages.

2) Using a quad op-amp chip like a TL074 to provide the 2 op-amps needed for the audio path and the two needed for the LFO is a bad idea.  You should use a separate dual op-amp for the LFO and audio portions.  This will let you decouple them separately, instead of having whatever the LFO does impact directly on the other 2 op-amps in the same hunk of epoxy.

3) The LFO and audio path need to be decoupled from each other.  You should run something like a 100R resistor from the shared +9v source to the V+ pin of the various CMOS and op-amp chips, and run a 10-100uf cap to ground from that pin.  This will give each chip a small reservoir of current to hold them over in case one of the other chips makes a sudden demand for current.  Those sudden demands are what you hear as ticking.

Arfman

Quote from: mdh on September 18, 2007, 01:56:42 PM
QuoteMark Hammer (did I get the name right?) has an article on his site talking about a quite similar problem with the Zombie Chorus due to the Vref on the delay chip (I think). The fix was quite simple by installing a voltage divider instead of using the LED's for that chip. I'm just now wiring up a Zombie Chorus, coincidently and will probably be able to start testing tonight.

Would the same fix for the Zombie Chorus work for the ultra flanger?

Am I missing something, or isn't that how the Ultra Flanger is set up already?  The LFO and the audio portion of the circuit have independent Vrefs: a zener or LED reference for the LFO and a voltage divider for the audio path.  Have you boxed it up yet, or is it a naked circuit board with long leads running all over the place?  Could be that it quiets down when it's in a metal box with short leads.

Well, certainly nothing like bursting into a board for the first time with a question demonstrative of ignorance...maybe I need to notch down my self-evaluation from the intermediate level.

You are quite correct; frankly I was scanning/searching this board and other sources for a fix. I found Hammers article on the Zombie Chorus and just assumed it was the same problem and potentially the same fix *without looking* at the schematic... :icon_redface:

No, it's not boxed and I did recognize that the problem was exacerbated with proximity to certain electronic fixtures...including my cell phone...Even at it's worse it didn't seem to me to be as bad as some other folks have described. It probably will be significantly better once I get it boxed. I tend to use a flanger in moderation...but when it comes to any noise or unwanted behaviour in a pedal or amp I feel compelled to keep at it until it's as good as possible...

So, thanks for setting it straight and good to meet you...

mdh

Are you kidding?  I've asked much dumber questions here, and hell, I might have even asked that one before.  As Mark alluded, there may be other issues.  I'm interested in improvements to the Ultra Flanger myself, as I've been meaning to use it as a building block for a dual-delay TZF unit.  What I'm definitely wondering about is the degree to which using the same bias point for the BBD and the audio path op-amps is a problem, as Mark seems to be suggesting.  One approach I've tried is to stare at the schematics for the Ultra Flanger, Zombie Chorus and Small Clone (which is dead quiet, by the way), and try to understand what makes the Small Clone so quiet.  Mark has discussed the issue before, but I'm still trying to come to terms with it.

That said, I think my experience with the Ultra Flanger (and R.G.'s layout in particular) is similar to yours.  Some noise, but not as much as the various other threads here made me expect, even when unboxed (still haven't boxed mine).

Mark Hammer

The UF does appear to have addressed the issue of different bias voltages for different parts of the circuit, but it still needs to tackle the decoupling.

Arfman

Quote from: Mark Hammer on September 18, 2007, 02:16:56 PM
The Zombie fix I posted is really only a partial fix.

There are a few separate matters to be addressed:

1) Using a single source to both bias the BBD chip and use as a reference voltage for biasing op-amps may "work", may save parts and cost, but is a compromise and a bad idea.  The two biases are different voltages.

2) Using a quad op-amp chip like a TL074 to provide the 2 op-amps needed for the audio path and the two needed for the LFO is a bad idea.  You should use a separate dual op-amp for the LFO and audio portions.  This will let you decouple them separately, instead of having whatever the LFO does impact directly on the other 2 op-amps in the same hunk of epoxy.

3) The LFO and audio path need to be decoupled from each other.  You should run something like a 100R resistor from the shared +9v source to the V+ pin of the various CMOS and op-amp chips, and run a 10-100uf cap to ground from that pin.  This will give each chip a small reservoir of current to hold them over in case one of the other chips makes a sudden demand for current.  Those sudden demands are what you hear as ticking.

Mark, thanks for jumping in. Are we talking about the Zombie, the Flanger, or both (my guess is both but already stepped in it once today by assuming)

I'm using TL072's for the opamps on either circuit so no issue there. Those are what I have laying around in my junk drawer. I call this "zen harmonic" construction...my wife will kill me if I bring any new parts home so the junk drawer ensures peace and harmony.

Any more suggestions greatly appreciated, in the meantime I'll roll up my sleeves and see what I can come up with and ask questions as I go...Thanks!!!

Arfman

Quote from: mdh on September 18, 2007, 02:39:10 PM
Are you kidding?  I've asked much dumber questions here, and hell, I might have even asked that one before.  As Mark alluded, there may be other issues.  I'm interested in improvements to the Ultra Flanger myself, as I've been meaning to use it as a building block for a dual-delay TZF unit.  What I'm definitely wondering about is the degree to which using the same bias point for the BBD and the audio path op-amps is a problem, as Mark seems to be suggesting.  One approach I've tried is to stare at the schematics for the Ultra Flanger, Zombie Chorus and Small Clone (which is dead quiet, by the way), and try to understand what makes the Small Clone so quiet.  Mark has discussed the issue before, but I'm still trying to come to terms with it.

That said, I think my experience with the Ultra Flanger (and R.G.'s layout in particular) is similar to yours.  Some noise, but not as much as the various other threads here made me expect, even when unboxed (still haven't boxed mine).

Decoupling as Mark suggested seems straight forward enough. I'll take a crack at it and post how it went. Then maybe we could make suggestions on changing the layout to accomodate...

oldrocker

Mine doesn't click or tick at all.  I perfboarded it so maybe I got lucky.  My Zombie does tick a little but it's hardly noticable really.  I love both of those pedals.

Mark Hammer

Decoupling is a smart idea for any modulation pedal that uses an LFO. 

It is also a smart idea to use a low-current dual op-amp for the LFO.  Why?  Because the "standard" 2-opamp LFO first produces a square wave, and then stretches that out to form a triangle.  What you hear as ticking is the sudden current draw as the square wave rises, placing noise on the power line.  The decoupling reduces the extent to which that spike "bleeds" into other parts of the overall circuit.  Using a low-current op-amp reduces the magnitude of that spike in the first place.  Think of the decoupling as being like ear muffs, and the low-current op-amp as being like smaller speakers.  Use of either will keep the volume of undesirable sounds down, but use of both will reduce it even further.

You will note in glancing through schematics of commercial phaser, flanger, chorus, and tremolo pedals that many of them will use either an LM358 or a TL022 dual op-amp for the LFO, and use different op-amps elsewhere in the audio path.  They use those particular op-amps because they draw less current.  There are other functional equivalents, I'm sure.  I just wanted to explain why you often see those ones.

The third strategy for deticking involves a slight change to the LFO design itself.  Nicholas Boscorelli explains this nicely in one of his Stompboxology ( http://moosapotamus.net/IDEAS/stompboxology/stompboxology.html ) newsletters, and you can sometimes see this in commercial designs.  The gist is to change that initial square wave into what is more or less a trapezoidal wave.  What we hear as ticking depends on the suddenness of the rise-time.  If the rise-time can be slowed down even just a wee bit, the current draw of the LFO is essentially distributed over a long enough period of time (in electronic, not human, terms) that it does not produce disturbances elsewhere in the circuit.

So, reduce the amount of current required to produce a square wave, make the rise-time of the square wave less sudden, and isolate the chip requiring the sudden current draw from the rest of the circuit.  That's how you get a tick-free modulation pedal.

Morocotopo

 Arfman, something else to consider is using shielded cable for the inputs and outputs from the board to the switch and from the switch to the jacks, with the shield attached to ground in only one end. might help, it helped me to get rid of the ticking totally in my CE 2 chorus build.

Morocotopo
Morocotopo

Arfman

Quote from: Mark Hammer on September 19, 2007, 10:25:03 AM
Decoupling is a smart idea for any modulation pedal that uses an LFO. 

It is also a smart idea to use a low-current dual op-amp for the LFO.  Why?  Because the "standard" 2-opamp LFO first produces a square wave, and then stretches that out to form a triangle.  What you hear as ticking is the sudden current draw as the square wave rises, placing noise on the power line.  The decoupling reduces the extent to which that spike "bleeds" into other parts of the overall circuit.  Using a low-current op-amp reduces the magnitude of that spike in the first place.  Think of the decoupling as being like ear muffs, and the low-current op-amp as being like smaller speakers.  Use of either will keep the volume of undesirable sounds down, but use of both will reduce it even further.


Thanks for the lesson, Mark! Sounds like I need to buy more parts anyway...Is it essential that I decouple both the audio and LFO IC's from the 9+, or could I leave the audio connected straight to it and then isolate the LFO with the 100R/100uf? Only reason I'm asking is that this would be much easier for the board that is already etched and populated...otherwise I would have just changed the layout...

But now I'll know better next time...I may just go ahead and pick up a couple of low current op amps, throw one in there and box it first just to see if that took care of it.

Arfman

Quote from: Morocotopo on September 19, 2007, 11:06:32 AM
Arfman, something else to consider is using shielded cable for the inputs and outputs from the board to the switch and from the switch to the jacks, with the shield attached to ground in only one end. might help, it helped me to get rid of the ticking totally in my CE 2 chorus build.

Morocotopo
If I continue to pick up interference after it's boxed I'll keep this in mind. I do this on my amp builds and have some nice, thin PTFE shielded cable I use.

Mark Hammer

"Anti-tick" protection is a bit like S/N ratio; more is preferable, but every little bit helps.  Happily, this is a problem that can be attacked on many fronts.  Personally, I could not find a TL022 in my city of 1,000,000 unless I bought a pedal and cannibalized it.  But I can find 100R resistors and 100uf caps. :icon_biggrin:   Do what you can do, using the various suggestions offered, and let your ears decide if it is enough or whether one of the unexplored avenues needs to be pursued as well.

Arfman

Quote from: Mark Hammer on September 19, 2007, 02:15:49 PM
"Anti-tick" protection is a bit like S/N ratio; more is preferable, but every little bit helps.  Happily, this is a problem that can be attacked on many fronts.  Personally, I could not find a TL022 in my city of 1,000,000 unless I bought a pedal and cannibalized it.  But I can find 100R resistors and 100uf caps. :icon_biggrin:   Do what you can do, using the various suggestions offered, and let your ears decide if it is enough or whether one of the unexplored avenues needs to be pursued as well.

I was able to easily couple the LFO opamp and it helped quite a bit and it wasn't the worst hack job I ever had to do. I just put the 100R/100uf components on pin 8 after cutting the 9+ trace. Now it only appears at all if I get real close to a source of interference so I suspect once it's boxed it will be quiet as possible. Just to be safe, and considering I have other LFO's I can use them for, I've got some TL022's ordered.

And in the future I'll know to modify the layout before etching the board...

Mark, thanks for the assist.