Roland AP-2 vintage phaser repair help

[lion]

I've been given a vintage Roland Phase II (AP-2). The pedal is faulty and non working (the reason I got it for free).

As can be seen the AP-2 was designed for a bipolar supply using two 9V batteries. According to the former owner the phaser was functioning and sounded really nice up until he attempted to run it from a +/-9V PSU - and accidently got the polarity wrong. Goodbye AP-2!

Here's the schem:


I imagine several components might have died from the reversed power polarity. I'm thinking the IC's at least, also maybe transistors, and electrolytic caps - but not really sure how much damage to expect.

I see LM2902N quad opamps are still available - and I might as well replace the 40 years old electrolytics damaged or not - but what about the transistors?

If anyone could offer an educated guess on what's likely to have be damaged from the reversed power polarity - and what to check and/or measure before it's safe to apply power again - it would be really helpfull and much appreciated.

Erik

[Mark Hammer]

I can't see the transistors being compromised by a flipped power supply.  The 2902 quad-ops can be easily replaced by LM324s or even TL074s.  My first inclination was going to be suggesting you replace the op-amps used for the LFO with a lower-current variant, but I see that a) they used a pair of quad-ops, and b) they also used the trapezoidal LFO trick, via C23, to minimize audible ticking.  So it's a non-issue.

I think your hunch of installing new electros for the power supply is a good one.  At 40+ years of age they have earned a pension and retirement.

Once new ones are installed, Check to see that the biasing of the FETs is set correctly/optimally.

The bypass switching is cute.  I like the way they use one set of contacts to mute the feedback signal so that no cumulative noise enters the Q1/Q2 circuit, which remains on all the time.

[PRR]

These electrolytics would fry fast. Others have high value series resistors and would not fry.

Agree w/Mark that the transistors are protected by resistance. The JFETs hardly know what the power supply is.

I have seen reversed chips live, and die. Fix the caps and see if the chips come to a DC bias; else replace.

[lion]

Thanks Mark and Paul.

Replaced the 3 critical caps - but no change (the +/-9V DC supply still got pulled down quickly - didn't wait to take a final reading, but shut off the power within a second or two).

Took out both IC's and installed sockets. Fired it up (without inserting new IC's) - now he DC voltage kept stable - and I got sound through the Q1-Q2.

With TL072's installed the rest of the circuit came alive too. For the sake of it I replaced the remaining 3 electrolytic caps. So far so good, glad I didn't have to replace the transistors.

I'm gonna play with the phaser a bit more before I can decide if it's a keeper, but at this point I've noticed a couple of things which might need attention.

1) Less than unity gain/volume through the circuit. Well actually, looking at a test signal sine wave on my scope, the max wave top to bottom might be unity alright, but due to the modulation of the signal the percieved loudness is less. I would think that I could tinker with Q1 emitter resistor (R6) or reduce the 47K serie resistor (R7) into Q2. Not sure which would be the best way to get a bit more gain.

2) The dry/wet level is fixed by design, but - at least for my use - I think a blend control, or maybe even better, a way to mix in the dry with the wet, would be a worthwhile improvement.

3) There seems to be an issue around the resonance control. Turning the pot there's a scratchy sound  - like DC on the pot (but both C5, C6 and C22 have been replaced) and worse - through the total pot range there seems to be like "sections" where the sound - and modulation level - changes abruptly. Surely not the usual resonance control behavior. Suspection the pot itself I replaced it with a new one - and then again with a third one - still the same odd behavior. Any ideas what might be the problem here?

Erik

[Mark Hammer]

Congrats!

Note that Roland used the RC network between the drain and gate for each JFET that increases immunity to FET distortion.  Though I haven't had much contact with him in far too long, I recall Mike Irwin telling me that he found he preferred to gradual onset of distortion without the RC network to the near all-or-none distortion he observed with the network.  That is, the phase shift stages were clean, up to a point where they got much dirtier.  The qualifier here is that he works more in the area of synths, and may be thinking in terms of much hotter signals than guitars normally provide.  In other words, the issue of sudden-onset clipping may be less of an issue, or even a non-issue, with guitars.

[lion]

After a break I'm back on the project.

Thanks Mark for the heads up on the RC network for the JFETs. I might look into it when I get the other issues solved.

I've given up gettting the resonance control to work as I think it should for a front panel control. Only a small part of the total pot range yields useable sounds/phasing. So instead I've added a trimpot pot on the board and made the resonance control an internal preset.

Regarding the BIAS trim pot - again a sweet spot (or even spots maybe, depending on ones preferences) to dial in. I'm trying to find the best setting by ear - but I wonder if there's a calculatable optimal setting which can be verified by measurement? Any advise or insight on this would by helpful.

Erik

[Mark Hammer]

I find FET bias to be sort of like where you set the gear teeth in a wah.  A rack-and-pinion wah never lets you sweep the full rotation of the pot, for fear you might push it past its physical limit and damage the pot.  So the gears only cover about 70% or so of the full available rotation, leaving a bit at each end of sweep.  What some folks will do is open it up and move the pot a smidgen (a couple of gear teeth this way or that), such that the sweep range of the wah is maybe a little higher, or a little lower, to suit their tastes.

Same thing with phaser FET bias.  The FETs are rarely swept through their entire range of drain-source resistance.  There's always a bit more usable sweep, and the manufacturer decides to set the bias so as to reproduce a given "signature" sweep range for that product.  But you don't have to like that range or be restricted to it.  In my own phasers (generally P90 clones) I like to install an "offset" control to shift the range upwards or downwards, from swirly to gurgly.  Play with the bias trimmer and take note of how the sweep sounds to you in different settings.  You may find that you can make the bias trimmer a sort of panel-mount control to adjust sweep offset.  On the other hand, not being familiar with this pedal, I may be encouraging something that is not feasible, and the trimmer has a very limited sweet spot for musical sweep.  Still, play with it and see what seems possible

[lion]

Thanks Mark for sharing your insight and taking the time to explain so well - all's clear now. From what I've tried so far I think it's possible that the AP-2 has a rather limited sweet spot for useable sweeps. I'll dive in to more experiments - and with a more methodical and analytical approach.

Erik

[lion]

Just want to wrap this thread up.

Turns out the FET bias adjustment on this baby has a very limited range for useable sweeps. Dialed in the best setting and I'm quite happy with it. Also does a nice vibrato with the dry signal switched out.

On to the next project - thanks Paul and Mark.

Erik

[system100m]

Interesting thread. I was wondering if there's any layout available for this AP2?