AquaBoy DX build problem caused by diodes?

[BillyBoy]

Not sure I'm on the right track with changing the resistance in the feedback loop.

I've been changing the size of T1.  I thought 150K would do it, but that only moved the runaway feedback from about 8 o'clock to noon on the Feedback knob (with the trimmer full up).  So I just tried a 500K T1.  That got it over to around 1 or maybe 2 o'clock, still with the trimmer maxed.  I doubt even a 1M trimmer would stop the runaway.  Going from 50K to 1M+ seems highly suspicious.  If I was changing the resistor instead, it would be over 500K - 750K at least (from its original value of 5.1K).

The repeats are also "falling apart" more than they should, even for a BBD delay (I have a 3005 Deluxe Memory Man and a couple of other BBD delays to compare to).  They turn into all static/noise after just a couple of repeats.  If I increase the Feedback, the repeats turn into a distinct "wooden tocking" sound that doesn't sound anything like the note played.  That happens after just a couple of repeats.

The shorter the delay time, the more pronounced the problem is.  Maybe adding that much resistance is causing the loss of fidelity on the repeats.  Maybe that C31 cap at 220n is now too big?

Or maybe Q3 and Q4 are gaining it up too much, causing the need for so much resistance in the feedback loop - ?

[BillyBoy]

Some of my last post is incorrect or misleading.

The pedal sounds better when I play a note and hold it.  Not great, but better.  The repeats start off nice and I can dial out the runaway repeats (just barely) with my 500K T1.  Still needs some tweaking, as the repeats are turning to all static/noise too soon. 

What doesn't work well at all is a short, staccato note.  With the exact same settings, a note I play and hold sounds as described above.  If I cut that note off, the repeats sound bad (the "tocking" sound) and they get louder and continue forever.

[BillyBoy]

In playing around with this some more, I am more convinced that the repeats are being boosted too much, causing distortion of the repeat signal and the runaway repeats.  Another thing I noticed that supports this conclusion is that I have to turn the Mix almost all the way down before the first repeat isn't louder than the original note in the dry mix. 

I'm thinking that instead of increasing the resistance in the feedback loop, I need to somewhere lower the gain on the repeats before they get to both the Feedback and Mix pots.  Maybe make R33 and R36 smaller? 

[Scruffie]

Yes it sounds like you have too much gain in this case, sometimes people do need to increase the feedback resistance but it shouldn't be as dramatically as you've had to.

Q3 & 4 shouldn't add any gain if set up correctly, it's possible there's an issue with one of them having too higher Q (they're low pass filters) so perhaps R30 wasn't the only bad part.

I'd audio probe again to see where the gain is introduced... and you're certain all your values are correct?

[BillyBoy]

Ha, ha – you lucked out :^)   You won't be subjected to another lengthy post I was preparing!

I was probing around for the volume increase and had a long message all written up about that, with all my data.  I studied it carefully and figured out the problem was with volume changes in the compression and expansion.  After studying the 570 datasheet to figure out how that works internally, checking all the resistors and caps that control it, reflowing solder, and scratching my head, I decided my data didn't make sense unless the 570 was having problems internally. 

I was close.  I decided to pull the 570 and try another one.  But when I pulled it, pin 16 was bent up under the chip.  I couldn't tell if that happened when I was pulling it or when I put it in.  So I straightened it out and put it back in.  Bingo!  Apparently that pin was bent up the whole time.  It was making partial contact.  The result is that the expansion wasn't cutting the signal back down after it was boosted by the compression.  I could have caught that much sooner, since I see that in an earlier post containing my 570 voltages that pin 16 was getting almost 0 volts, consistent with my bent contact.  I suppose I was too preoccupied with my broken wet path problem at the time to notice the second issue looming.

The pedal already seems far better – repeats are better and I can manage the feedback.  That was just a very quick test, so I'll put it through its paces and make sure all is well.  I haven't really been using the modulation to this point, but earlier testing seemed to show it was working fine.  I could probably go back to the smaller T1 trimmer, but I've already desoldered and resoldered enough that I'm afraid I'll lift a trace or something.

Many thanks for your advice and help!  Some of what you provided I wouldn't have come up with anytime soon, if at all.  And sometimes just having a second brain reviewing my rambling helps keep me on track and moving forward.  I appreciate both!  I don't know where you're located, but if you find yourself in Dallas I owe you a beverage of choice :^)

[BillyBoy]

Yup, all good!

I had a little time to try it out and it sounds great.  Very quiet.  No clock noise.  No LFO noise.  No white noise.  Repeats are nice and clear, even on the long analog delays.  All the problems with muted notes, chords, runaways, tocking repeats, etc are completely gone.  100% analog satisfaction.

I have an off-brand single 3005 delay from the 80's, a Keeley modded AD9 (3208s, I think), and a much-loved dual 3005 DMM.  Never had a DM-2 (and now have no need to try one), so can't say anything about that.  This build is better than the first two, and gives my DMM a good show.  At about 1/4 the ebay price of a dual 3005 DMM or DM-2, I would build another one of these instead.  Haven't tried my Gilmour EMGs through a cranked Big Muff into a boost then into this delay yet to see if I can overload it, but my strat with Kinman MK-IIIs loves it in front of a Fender Twin.

[Scruffie]

That'll teach me for not checking the full voltages shoulda spotted pin 16.

Well done on taking the initiative to check the datasheet though, some good debugging.

Glad you got it going and are enjoying it anyway.