AquaBoy DX build problem caused by diodes?

[BillyBoy]

Hi,

My Aquaboy DX build is having problems.  I think the problem is I missed a couple of points that weren't clear to me in the instructions.  I want to verify those points, as well as get some opinion on whether my misunderstandings are the source of my problem.

Some of my build options, FYI: 
    15v RoadRage with LT1054
    MN3005 BBDs
    3101 clock
    jumper R27, R55, and R58
    R56 is socketed with a 470K in it
    R13 and R14 socketed both with 33K
    R50 and R51 both 8.2K
    NE570 compander
    VTL5C3 vactrol

Everything else is as listed in the BOM.  I put in the extra LED and it adjusts as expected with the Rate knob.  Switch wiring works as expected.  Have triple checked soldering.  Used a meter to measure each component value before soldering it, so pretty confident about part values/placement.  I've built quite a few pedals and am pretty careful, since I've found it's much easier to take the time and do it right than debug and fix it :^)

The symptom is found when trying to set the trimpots.  I set the controls as described in the doc with the 2nd BBD removed.  I used my DMM to "preset" the clock trimmer to 6.5K, but it wouldn't adjust down quite that far.  So I put it back in the middle, which is around 8.5K.  Using an audio probe, I can follow my input signal through to pin 7 on the first BBD.  But I get loud whining and maybe some very, very faint buzz that might possibly be some input signal coming through on pins 3 and 4.  But there is no easily recognizable signal coming through from the input on pin 3 or 4, and no signs of any delay - loud whining instead.  The bias for that BBD does nothing to the sound on those pins.  Adjusting the clock trimmer does change the sound/pitch of the whining.  Voltages on the first BBD are:

pin1 - 8.6v
pin2 - 4.3 v (loud whining sound)
pin 3 - 5.9v (loud whining and maybe very faint buzz/signal)
pin 4 - 5.9v (loud whining and maybe very faint buzz/signal)
pin 5 - 0
pin 6 - 4.3v (whining)
pin 7 - 4.2v (good guitar sound, altered by compander a bit)
pin 8 - 7.9v

I'm obviously not getting my 15v to the BBD.  I know my RoadRage is putting out the 15v (measured it).  The whining frequency matches the clock at around 8.5K.

In searching the forum, it appears that D2 and D3 should both be left out for a 15v MN3005 build.  I think that is my mistake in reading the instructions.  I have both D2 and D3 in place.  I believe that is chopping the power to the BBDs, clock, etc. 

But before I remove those two diodes, I want to verify that I should leave out both D2 and D3 and leave those spots on the board both open (no jumpers).  I believe that is what I should have done.

Secondly, any guesses on whether those diodes (low power) could cause the symptoms I'm seeing?  I tried several MN3005 and MN3008 BBDs and they all behaved the same way.  I don't have another way to test the BBDs, but I think they are OK, particularly the MN3008s.  All 12 BBDs I have could be bad, but I'm discounting that for the moment, particularly if the low voltage could cause my problem.  Maybe a better question is whether anyone happens to know if bad BBDs might behave like this?  I suppose before I remove those diodes, I could test with an MN3205, BL3208, etc, but I'm not sure that's really safe for those other BBDs, or even useful.

That clock trimmer would only get me down to about 6.7K when I turned it all the way.  Not sure if that'll be a problem, or if maybe it'll change once the diodes are removed.

Thanks for any words of experience!

[Scruffie]

Yes you should have left out D2 & D3, remove those and then start debugging if it doesn't work after that, it wont work with them.

[BillyBoy]

Thanks for the quick reply!

I lifted an end of each diode (D2, D3) off the board, leaving them attached at only one end for now.  On the first 3005 that gives me:

pin 1:  14.5v
pin 2: 7.3v
pin 3: 8.3v
pin 4: 8.4v
pin 5: 0
pin 6: 7.3v
pin 7: 6.2v
pin 8: 1.0v

I can now hear signal passing through pins 3 and 4 on my BBDs (on all 4 of my 3005s and all 8 of my 3008s).  I found that I could not hear signal through the first BBD unless I also installed the second BBD.  Then I could adjust the bias pots for each of the BBDs.  But I could only get signal to go through even then after also adjusting the clock trimmer nearly all the way counterclockwise.  If I don't do that, then I can't find a spot on the bias trimmer where I can hear signal.  My clock frequency is something like 140K, measured on pin2 of the first BBD.  Even if I ignore the horrible noise and adjust the clock all the way clockwise, I still can't get a frequency as low as 6.5.  The lowest is about 6.7, but it is all noise and I cannot hear signal through the BBDs anywhere near that setting.  Hmmm... that doesn't sound right.

I cannot hear the delay signal with my probe any further than the cancel pot, or the "in" side of R30.  After that, I have no sound going back to Q3, Q4, or the compander.  I have only dry signal coming out of the pedal output.

I don't get far by following the instructions for setting the trimpots.  I have to crank the clock way up before I can get a small range on the bias pots where I can hear the signal on pins 3/4.  When the clock trimmer is set to sound best, it has a very narrow band of adjustment.  A hair clockwise, and I start hearing clock whine, which gets progressively worse as you continue turning that direction.  A hair counterclockwise starts getting spluttery.  There's little travel on the pot counterclockwise from the sweet spot.  The sweet spot is nice and clear though.  This is listening on pins 3/4 of the BBDs. 

I have to toggle the wave switch off center, up the depth and rate, etc to hear any effect on the pin 3/4 output of the BBDs.  T1 and Feedback adjustments don't appear to do anything.  Neither does Cancel.  But that's probably because my delay signal is being lost, I guess.  It's hard to tell, but I don't think I have any delay that is as long as I would expect with two 3005s - I imagine that's because the clock is off.

Not sure if there is more than one problem, but it seems like the clock is messed up and I'm losing my delay signal somewhere after the Cancel pot, maybe at C22 since R30 is the last place I hear it.  I suppose those are the places I'll start looking for a bad connection or wrong value component - any other suggestions?  Other than goofing up the diodes, did my other build options look OK?  Will take a break and let my head clear first :^)

[BillyBoy]

Follow up...

I checked labeled values of all caps, resistors, etc and made sure they were correct - all good.  Went over both sides of the board with a magnifying glass (again) looking for extra connections, bad connections, etc - didn't see anything.  Reflowed all the solder, just 'cause.  Reseated all ICs.  Went through the calibration process again.  It still isn't working correctly, but there are a couple of changes.  I think the changes are due more to following the calibration instructions differently, not likely due to any physical changes. 

I can now get in the 6.5K range on the clock (measured on BBD pin 2).  I'm between 6.5K and 8.5K in the first half of the pot.  I think this was true previously, but I wasn't setting the trimmer there, because I wasn't following the instructions correctly previously.  So I put it on 6.5K (about two thirds of its full range, clockwise).  When I listen on pins 3/4 of both 3005s, I have a narrow range of adjustment near center on the trimmers, where I can hear guitar signal.  The clock noise is piercingly loud, so it is a little difficult to dial in the point of least distortion in the signal with all that clock noise on top of it.  The signal is strong, but the clock is loud, too.  I have the trimmers set to about the middle of those ranges where I can hear the guitar.  Previously, I used the clock trimmer to kill the clock noise on pins 3/4 of the BBDs so I could bias them.  I'm guessing that's not what I should have done.  I should get the least distortion, but with whatever clock noise there is when the clock is at about 6.5K (see below).

I have a pretty decent amount of delay in the first 3005, and about twice as much coming out from the second BBD.  They seem to be working OK.

The next step is to adjust the clock trimmer.  Previously I was doing that while listening on pins 3/4 of the 3005s.  To make the clock noise go away when doing that, I have to crank the clock trimmer nearly all the way counterclockwise, increasing the frequency to around 140K.  But maybe at that point, I'm supposed to be listening to the sound at the output jack, adjusting out any clock noise that is present after the compander and filters do their things, which include getting rid of a bunch of clock noise.

But that's where I'm still stuck.  The delay signal coming out of the second 3005 is still stopping at R30.  As a result, none of the knobs or the wave switch have any effect on the sound when I listen at the output jack.  I just have a nice clean signal with no delay or repeats (and no clock noise).  I also don't hear any difference when I adjust the Cancel or T1 trimmers, presumably also due to losing the signal.

What should I hear on pins 1, 2, and 3 of the Cancel trimmer, particularly as I adjust it?  There is sound getting to R30 (pin 2 of the Cancel trimmer).  I have the Cancel trimmer set to midway and have sound coming into R30, so I think that trimmer is OK (or at least not yet my problem).

I measured the value of each cap and resistor before I soldered it.  They obviously were all working OK at that time.  Should I be suspicious of C22 since it goes to ground (perhaps taking my signal with it) just after R30?  I've never fried a cap or resistor from soldering, but there's always a first time.  About all I know to do at this point is start trying to desolder and replace stuff, starting with R30.  With a board this full, that is an option I'm not crazy about.  I'm more likely to cause a new problem than solve one doing that.

I don't think twiddling values on R13/14 and R56 would fix this symptom.

When I listen on BBD pins 3/4, I don't get any repeats.  But that seems in line with my signal dropping out before it gets to the point where it feeds back.

Voltages on both BBDs are about the same as what I last reported for BBD 1.  There are slight differences on pins 3/4 between the two BBDs because the trimmers are set slightly differently.  BBD 2 also gets a higher reading (8.5v) on pin 7 than BBD 1 (6.5 v).  I think that should all be OK, just caused by bias trimmer settings.

I think I'm just grounding out or have a broken path at or after R30, since that's the last place I can hear anything. 

Any suggestions before I start replacing components? 

[Scruffie]

I don't mean to be rude and lots of info is very helpful but any chance you could just bullet point what the problem is? That's a lot to read through.

I don't know what you're on about with the 6.5k on the clock trim too... it doesn't need setting by resistance, it needs setting by frequency or just ear.

From the last line of your post though, don't just start replacing parts, you'll likely make things worse.

Post your full voltages readings up and where you have signal up until, that should be enough to start with.

[BillyBoy]

No rudeness taken! 

Problem 1 - My lack of confidence in properly interpreting/following the instructions for adjusting the trimmers.  I'm guessing what is meant, leading to trying stuff, leading to lots of words describing what I find.  In those steps, it says to set the clock trimmer to 6.5K (hz) to start with, if you have a DMM with a frequency function.  My DMM does that, and that's what I'm referring to.  This following-the-instructions problem is made more difficult because of Problem 2.  I've adjusted analog delays before, but I knew those delays worked and only needed to be adjusted.  I could tune by listening to the output from the output jack - no probing needed.

Problem 2 - I have delayed, modulated signal coming out of both 3005s and the Cancel trimmer, going into R30.  (Since my last post I figured out my Cancel trimmer works OK.  I was having dumb issues with getting good probe contact on it.)  But the signal stops at R30.  I cannot hear anything with my probe past the Cancel trimmer, pin 2 (going to R30). 

Since I believe I should be able to hear something past R30, I'm guessing that R30 is failed open, breaking the circuit.  Or C22 is failed closed, sending everything to ground.  If those were both OK, I should be able to hear something on R31, but I can't.  If I can't get signal from the BBDs going back to the filters, Q3, Q4, compander, feedback loop, etc, I'm left with nothing to hear at the output jack except clean dry signal - no repeats, no delay, no modulation, no clock noise, etc - because that all requires the signal to get past the point where I'm losing it.

So if my logic is correct, I think I need to try replacing R30 and/or C22.

[Scruffie]

Right, gotchya, so the delay is working but you're loosing signal in the wet path.

Could you post the voltages up for Q3 and while you're there Q4 and the compander.

[BillyBoy]

Yup, that's what I'm chasing now.

Not sure if knob positions or anything matters, but here's what I've got:

Q3 and Q4 both have 14.8v on one pin.  Both have 0 volts on the other two pins.

Compander pins
1: 1.15v
2: 1.83v
3: 1.83v
4: 0v
5: 1.83v
6: 1.83v
7: 5.78v
8: 1.83v
9: 1.83v
10: 4.95v
11: 4.95v
12: 1.83v
13: 14.9v
14: 1.83v
15: 1.83v
16: 350mv (yes mv)

[BillyBoy]

Also, on the end of R30 connected to the Cancel trimmer, I have about 9v.  On the other end of R30 I have 0v.

[Scruffie]

Certainly sounds like you have a short somewhere, Q3 & 4 base & emitter should certainly not both be 0V.

Might just be a hair of solder, if you hold a light through the board sometimes it can help spot it, otherwise a simple re-flow of the solder joints in the schematic area might solve it.

[LaceSensor]

dont wanna get too involved, but one issue I had one time building a delay based on the DM-2 was simply a dodgy transistor

this thread is quite long, I havent read it all, but replace all parts that are (hopefully) socketed such as transistors, op amp, after the delay stages.
Seems the BBDs and LFO work; thats more than half the battle won!

[Scruffie]

dont wanna get too involved, but one issue I had one time building a delay based on the DM-2 was simply a dodgy transistor

this thread is quite long, I havent read it all, but replace all parts that are (hopefully) socketed such as transistors, op amp, after the delay stages.
Seems the BBDs and LFO work; thats more than half the battle won!

You're on the right track, I haven't read the whole thing either but he says he has 0V on Q3 & Q4 Base & Emitter so it could very well be dodgy transistors, although first i'm trying to rule out a simple short.

[BillyBoy]

Thanks!  One problem solved, new one uncovered.

R30 was bad.  I replaced it and that fixed my wet path problem.  Delayed/Modulated signal now comes out my output jack.

The remaining problem is with Feedback.  I have the bias pots (near noon on both), Cancel (about 10 o'clock), and Clock (about 6KHz) all set pretty well.  I have the T1 trimmer set all the way counterclockwise.  Modulation off.  If I turn the feedback knob past about 8 or 9 o'clock, I get runaway feedback.  That's at about the point where I can start to hear a couple of repeats.  Once it starts running away, the repeats quickly get totally distorted.

Getting close!

[Scruffie]

Okay, now fiddle with R13 & 14 in equal increments until you achieve half supply on Pin 7 of the 570 and then adjust R37 down until you get half supply on pins 10 & 11.

Once that's done, if you're sure every value is correct and you still get runaway feedback you can try increasing R39 to reduce the amount of signal feeding back through the compander.

[BillyBoy]

OK.

I changed R13 and R14 to 47K, which gives me around 7.4v on pin 7.

I changed R37 to 8.8K, which gives me around 7.4v on pins 10 and 11.

That made my repeats sound better. 

But it didn't seem to have any impact on the runaway feedback.  It still happens with T1 all the way counterclockwise and the Feedback knob at any point past about 8 o'clock.

I will increase R39 next.  In my build, R39 is surprisingly difficult to get to because of the size and proximity of other components in that area.  How would it be to make T1 a 100K, 150K, or 250K trimmer?  I'll almost certainly have to remove T1 to change R39, so it would be easy to just swap T1 instead.  Not sure how much resistance I need to add to move the runaway feedback to the very end of the Feedback adjustment.