Phase 90 de-ticking advice?

[mdh]

I built a Phase 90 from the Tonepad layout (http://tonepad.com/getFileInfo.asp?id=42), and I love it, but it ticks. Well, it's more like it swooshes on slow speeds and thumps on fast speeds. This occurs whether I'm playing through a clean channel or with some gain, but it can be worse with gain. My understanding is that there are two typical causes: (1) high current draw from the square wave generating part of the LFO and (2) capacitive coupling between nearby long wires, particularly if the wires from the rate pot are long and come close to input or output wires. I think I've read all of the pertinent posts on this topic, and I'm still stuck. Generally it doesn't seem to tick when the lid is off and the board is allowed to pop out of the box a little bit, so I tend to suspect that my problem is capacitive coupling of wires to the board.

Here's what I've done so far:

1. Swapped the op-amp in the LFO section (TL072) out and replaced with a TL062. The rationale was that TL022's seem to show up in lots of LFOs, and the TL062 was the closest I had to the TL022 in terms of current consumption. I'm uncertain whether swapping the op-amp should help here, though, because I don't know what determines the realized current draw of the op-amp.

2. Added a 100 nF cap across the supply pins of the LFO op-amp. It didn't seem to have any useful effect, but I haven't taken it out yet. I'm mentioning it in the interest of full disclosure, and in case it was a stupid idea. So please tell me if it was.

3. Made the wires to the rate pot extremely short. They're maybe 1.5-2cm long, and don't come close to any of the other wires. The closest other wire is +9V, which connects to a nearby point on the board. But pushing the board into the box, if anything, moves that wire further away from the rate pot wires. The ground to the board is connected to the ground on the output jack, and this wire runs sort of parallel to the +9V wire. They're not bundled together or anything, though.

4. Replaced the battery with a regulated 9V supply (9V DC wall wart which shows ~12.5V with a 50 mA load, LM317 dialed to 9.06V, tests same under load from pedal; big 3300 uF filter cap in the supply). The rationale here was that ticking seemed to get worse as the battery went flat, so I thought I could eliminate that with a stable supply.

5. Replaced wires from stomp switch to board input and board output with RG174 coax. The input cable shield is grounded at the ground on the stomp switch; the output cable shield is grounded on the board near the solder pad for output. All the other wires in the box are unshielded, including those from the stomp switch to input and output jacks, and to the LED. Generally, though, these wires aren't any longer than they need to be, and they don't in general run parallel to each other. The one place in the box where the wiring is sort of undisciplined (long wires going all over the place) is south of the stomp switch. I'm paranoid about battery snap wires breaking at their solder joints, so I tend to leave them a bit long. Add to that the fact that I added the DC jack the quick 'n dirty way, by wiring a battery snap to it backwards, and connecting that to the existing battery snap. So that area is a little ugly.

I have not yet tried shielding wires to rate pot, but they're so short it would almost seem impractical. I'm not sure where I would connect the ground for their shields (the back of the pot itself, like in the control cavity of many guitars?).

I'm hoping to find a way to fix this without major surgery to the board itself. One suggestion I've heard that I'd like to avoid is to put a small resistor in series with the V+ to the LFO op-amp and a large electro from there to ground. The layout is sufficiently cramped that I don't really see how I could do that in a way that would work in the long run and continue to fit in the box (it's crammed in a 1590B).

If anyone is able to help but needs more details on the wiring (e.g., the grounding scheme, or how the wires are physically routed in the box), I'm willing to post photos, drawings, whatever. I really like the sound of this pedal, but it's not really useable if it's ticking all the time. Any tips for diagnosing the specific cause of the problem would be most welcome.

[jonathan perez]

mine only "ticks" when im not playing anything...weird...

[Ardric]

I found ticking with the Vref trimmer at some places, also using a TL062 for IC1.  It seemed a fine line between ticking on one side, poor or no phasing on the other.  I used a 10-turn 100K trimmer so it was easy to find a sweet spot, but not without a little ticking still remaining.  I could only get three fets closely matched out of my collection of 20... three with a Vgs within about 15mV, the forth about 50mV away... so I shelved the board on the belief that when I found a really well matched set I'd simply dial the ticking out.

[mdh]

I found ticking with the Vref trimmer at some places, also using a TL062 for IC1.  It seemed a fine line between ticking on one side, poor or no phasing on the other.  I used a 10-turn 100K trimmer so it was easy to find a sweet spot, but not without a little ticking still remaining.

That's a good point, maybe I should play with the bias trimmer some more. Back when I had it out on the bench wired up to a breadboard, the bias trimmer definitely seemed to affect whether it ticked or not, but as you say, it was really hard to dial in a good phase without ticking. However, in my experience matched FETs ticked over a broader range of the trimmer than unmatched FETs. Maybe that's because the matched FETs phased over a broader range as well.

Still, it seems to me that since it doesn't tick with the lid off, there must be something going on with the offboard wiring.

TBOM, yeah, mine definitely ticks even when I'm playing. Sometimes it seems like it's not as noticeable when playing, but it's always there.

[jonathan perez]

TBOM!

noone gets that 

right now its in a wah shell...

im going to take it out and put it in a box, see what it does.

im also going to check the wiring, make sure its all dandy...

[Fp-www.Tonepad.com]

If the pot wires are close to the input/output wires/jacks, there's a chance of some noise leaking into the audio path from the LFO. Make sure pot wires are as short as possible and that they're away from other wires, using shielded wire for the input and/or output should prevent any noise.

Some swooshing is to be expected when not playing, but shouldn't be loud enough to call noise.

[mdh]

If the pot wires are close to the input/output wires/jacks, there's a chance of some noise leaking into the audio path from the LFO. Make sure pot wires are as short as possible and that they're away from other wires, using shielded wire for the input and/or output should prevent any noise.

That's the thing, though, the pot is turned around so that its pins (alas, it has pcb pins, not proper solder lugs) are toward the top edge of the enclosure, with short wires going straight to the board. It's nowhere near the jacks, and both the input and output wires (from the board to the stomp switch) are shielded. The wires going to/from the jacks are not shielded, but as they're nowhere near the rate pot or its wires, I figured it wouldn't help.

The truly strange thing is that when I have the box open for awhile and do just about anything to the inside of the box (move wires around, tell it that it's a good little phaser, whatever), it usually sounds fine for a little while, but after playing it for awhile, the damned thing starts ticking again. Future sessions always seem to be plagued with ticking if I don't open it up in between. Maybe it's just getting lonely or something. 

Some swooshing is to be expected when not playing, but shouldn't be loud enough to call noise.

This is definitely more than a little background swooshing. My DOD flanger swooshes when you're not playing, but it's not noticeable enough to be a problem. The Phase 90 is way louder than that, even when playing.

[Fp-www.Tonepad.com]

Sorry. You seem very upset. A lot of people have built this project from the pdf with success. Take it easy. I guess you're the 1% unsuccessful builder.

From what you describe, maybe you have a bad cap somewhere, since you mention that 'it usually sounds fine for a little while'.

Good luck.

[nordine]

i had a ticking problem with a phase 45,
turned out to be that 10uf cap that goes from the oscillator opamp to ground.. it suffered from sloppy soldering... resoldered it, and everything went fine

[mdh]

Sorry. You seem very upset. A lot of people have built this project from the pdf with success. Take it easy. I guess you're the 1% unsuccessful builder.

Not upset, exactly, just kind of frustrated. I've tried so many things to address this problem, and nothing has worked. I guess I'll just keep on plugging away at it, but perhaps another break is called for before I worry about it too much. Make no mistake, I'm very appreciative of all your fantastic layouts, Francisco, this is nothing against you! And don't worry, I won't come unhinged over this

Thanks for the cap tip. I swapped out both of the electros just to be safe, but no luck so far. I'm assuming that the film caps are less likely to be at fault. I will double check the solder joints as well, as nordine suggests.

[mdh]

I just thought I'd post the resolution to this problem, as solutions are often lacking in this sort of thread. The short version is that I removed the 22k feedback resistor, and the ticking went away. The phasing is still satisfying to my ears, so I'm happy.

The longer version is that when I audio-probed the circuit, I heard ticking on pin 7 of IC2 (the output of the second phase stage) and pin 7 of IC3 (the output of the fourth phase stage). I didn't hear any ticking on IC2, pin 1 (output of first phase stage), nor IC3, pin 1 (third stage), so I figured that there was something funky going on with feedback between stages 4 and 2.

If anyone with theoretical knowledge cares to expound on why this solution works, why it shouldn't work, or when it should/shouldn't work, I would be interested to hear about it.

If anyone wants to hear about the various crazy theories I entertained in the process of debugging this problem, be careful what you wish for, I will oblige.