Nurse Quacky debugging

[jimladladlooklike]

Hey folks.

Having trouble with my Nurse Quacky build. There's no change in sound when the effect is turned on. when the range pot is turned anti-clockwise a hiss becomes audible.

Here's the schematic and layout I'm using:





And here are the voltages. I've included the reading for when the range pot is both turned clockwise all the way then anti-clockwise all the way in the following format: clockwisev/anti-clockwisev

IC1 TL072
1: 1.6v/1.6v
2: 1.6v/1.6v
3: 1.6v/1.6v
4: 0v/0v
5: 2.2v/1.6v
6: 2.2v/1.6v
7: 2.2v/1.6v
8: 9v/9v

Q1 2n3904
C: 0v/0v
B: 0v/0.5v
E: 0v/0v

LED1
A: 1.6v/1.6v
K: 0v/0v

LED2
A: 1.6v/1.6v
K: 2.2v/1.6v

It looks as though there may be a problem with the transistor voltages... There's also no signal at the collector of Q1. Any guesses?

[jimladladlooklike]

Okay so now there is sound but it's super crackly and there is none of the desired effect

[Mark Hammer]

First thing to do is assess the two halves of the circuit separately.  Lift the connection between the transistor collector and the junction with the two 4n7 caps and 47k resistor to enable this.

You should be able to measure a DC voltage at the junction of the Attack pot and 22uf cap, in the range of a couple of volts (max) when you strum.  If so, that indicates the envelope-follower part is working properly.  You should also be able to hear a low-range bandpass-filter sound at the audio output when you strum.  If either of those does not occur, that indicates a problem in that half of the circuit.

[jimladladlooklike]

Thanks for the reply!

The voltage reading at the junction of the attack pot and 22uF cap was 0.02v when strumming...

Do you mean basically a low pass filter at the output? Or some kind of filter effect similar to the one we're aiming for? Either way, it doesn't seem to be doing either of those things I'm afraid.

[reddesert]

One of of the op-amp + inputs, pin 3, is held an LED diode drop above ground (so about 1.6-2 V), and the other op-amp +input is wired directly to the +4.5 V bias rail, which is the 4th row from top on your vero layout. The DC values of the first op-amp inputs, pins 2 and 3, should be about 1.6 V, which is okay, and the DC values of the second op-amp's inputs, pins 5 and 6, should be 4.5 V and not change when the pot is turned. So there is clearly a problem with the wiring around the second op-amp and bias rail, whether it be the bias rail voltage or something else.

[Mark Hammer]

Thanks for the reply!

The voltage reading at the junction of the attack pot and 22uF cap was 0.02v when strumming...

Do you mean basically a low pass filter at the output? Or some kind of filter effect similar to the one we're aiming for? Either way, it doesn't seem to be doing either of those things I'm afraid.

The filter is a bandpass, as far as I know, but when the resistance to ground from the junction of the two 4n7 caps is high, then the passband is low.  The envelope voltage fed to the transistor base makes the transistor operate like a voltage-controlled resistor in parallel with the 47k fixed resistor.  It makes the emitter-collector resistance lower, which raises the range of the passband.  The 22uf cap "hangs onto" the envelope voltage for a little while, such that the decay is slow enough to hear the filter sweep beck down again.

If you temporarily wire up a variable resistor (e.g., a 100k pot) in parallel with that 47k resistor, you should be able to make the filter move up and down by adjusting that variable resistance.

As for the rectifier/follower portion, 20mv seems pretty low to me.  Are you measuring that with the Sensitivity control up full?  If so, verify the signal path by measuring the AC voltage at the wiper of the Sensitivity pot, after the 47k, then after the .01uf cap, and finally at the output of that op-amp section.  Then switch to measuring DC, and measure after the LED on the output of that op-amp section.  All of this is while you feed an audio signal to the input jack (ideally strumming your guitar, but you can use other sources).  You should see something at each of those points that corresponds to the intensity/loudness of the audio input signal.

[thermionix]

Thanks for bringing this up, I had no idea an envelope filter could be so simple.  I'm gonna breadboard this one (I like simple).  I'll report back if I have any troubles.

[jimladladlooklike]

One of of the op-amp + inputs, pin 3, is held an LED diode drop above ground (so about 1.6-2 V), and the other op-amp +input is wired directly to the +4.5 V bias rail, which is the 4th row from top on your vero layout. The DC values of the first op-amp inputs, pins 2 and 3, should be about 1.6 V, which is okay, and the DC values of the second op-amp's inputs, pins 5 and 6, should be 4.5 V and not change when the pot is turned. So there is clearly a problem with the wiring around the second op-amp and bias rail, whether it be the bias rail voltage or something else.

I see. Thanks for pointing that out, makes sense! I have re measured the voltages a pins 5,6 + 7 since removing the collector of the tranny from the circuit and they are no all reading as 4.5v. So could the tranny or close to it be the problem?

[jimladladlooklike]

The filter is a bandpass, as far as I know, but when the resistance to ground from the junction of the two 4n7 caps is high, then the passband is low.  The envelope voltage fed to the transistor base makes the transistor operate like a voltage-controlled resistor in parallel with the 47k fixed resistor.  It makes the emitter-collector resistance lower, which raises the range of the passband.  The 22uf cap "hangs onto" the envelope voltage for a little while, such that the decay is slow enough to hear the filter sweep beck down again.

If you temporarily wire up a variable resistor (e.g., a 100k pot) in parallel with that 47k resistor, you should be able to make the filter move up and down by adjusting that variable resistance.

As for the rectifier/follower portion, 20mv seems pretty low to me.  Are you measuring that with the Sensitivity control up full?  If so, verify the signal path by measuring the AC voltage at the wiper of the Sensitivity pot, after the 47k, then after the .01uf cap, and finally at the output of that op-amp section.  Then switch to measuring DC, and measure after the LED on the output of that op-amp section.  All of this is while you feed an audio signal to the input jack (ideally strumming your guitar, but you can use other sources).  You should see something at each of those points that corresponds to the intensity/loudness of the audio input signal.
[/quote]

Thanks for explaining. The readings were taken with the pot up full indeed. I will try out what you've suggested when I get chance!

[jimladladlooklike]

Clearly unable to use the quote function! ^

[reddesert]

One of of the op-amp + inputs, pin 3, is held an LED diode drop above ground (so about 1.6-2 V), and the other op-amp +input is wired directly to the +4.5 V bias rail, which is the 4th row from top on your vero layout. The DC values of the first op-amp inputs, pins 2 and 3, should be about 1.6 V, which is okay, and the DC values of the second op-amp's inputs, pins 5 and 6, should be 4.5 V and not change when the pot is turned. So there is clearly a problem with the wiring around the second op-amp and bias rail, whether it be the bias rail voltage or something else.

I see. Thanks for pointing that out, makes sense! I have re measured the voltages a pins 5,6 + 7 since removing the collector of the tranny from the circuit and they are no all reading as 4.5v. So could the tranny or close to it be the problem?

The transistor doesn't have a connection to the +4.5v rail, or a DC connection to the - input of op-amp number 2. So it shouldn't really have any effect on those DC voltages, if the circuit is wired correctly. Start by
visually checking areas around the 4.5v rail and the transistor to make sure everything is where it is supposed to be. Use the resistance or continuity function on a multimeter to verify all the connections.

IME, problems, especially in vero builds, typically have to do with a part that is not soldered, or a solder bridge between rows, or a part that is misplaced one hole over. People often focus on possible faulty components when debugging, but faulty connections are more common than bad components.

[thermionix]

Couple questions.  Is that 100r in the power supply in the right place?  Should it be on the battery side of the filter cap?

Would a 4558 perform better than a 1458?  I had a DOD FX25 a couple years ago, different circuit I know, but I replaced the 1458 in it to a 4558 after seeing the recommendation here, and it got rid of the distortion it was prone to when I'd hit the strings hard.

[thermionix]

Well it's working on the breadboard, per the ROG schematic above, only sub was 47r for 51r (big deal).  It sounds good, but it's way more subtle than I was expecting.  Too subtle IMO.  Should try a different 2N3904?  I'm not really sure if I want more sensitivity or range, but I think sensitivity.  Pot value?

I'm also picking up radio pretty bad, I'm sure that would be reduced or eliminated in a metal box, but it comes through most when the sweep goes to treble, and at times sounds like distortion.  But at least I think it's not distorting.  I used a 1458, I don't have an LM358.

[reddesert]

There is a vero layout for the ROG Nurse Quacky (descendant of Jack Orman's mods to the Dr. Q) at: https://tagboardeffects.blogspot.com/2013/01/nurse-quacky_12.html.

I like this vero layout better than the one posted upthread because no standing resistors and generally cleaner, which makes it a lot easier to check voltages, continuity, and debug.

I built the tagboard vero layout and it works, but IIRC the effect strength varies considerably depending on where on the fretboard you're playing vs the knob settings.  There is an old DIYSB thread where people tried different op-amps, and I found that a 1458 worked, an NE5532 worked well, and a TL072 really didn't quack unless all the knobs were turned up. That is a little weird since normally op-amp type doesn't matter that much; maybe it's sensitive to the input impedance.  Anyway, I also built a Madbean Kraken (on Bean's PCB) which I like better, but it's a more complex circuit of course.

[thermionix]

Cool, I'll try a 5532.  Tomorrow though, amp's off for the day.

[thermionix]

I messed around with this some more.  I put in a 5532, didn't notice much difference.  Couldn't find an error, maybe a halfass connection somewhere as my breadboard has been (literally) collecting dust for the last few months.

What happens is, with the range pot around halfway, it works semi-okay.  With the range pot maxed, it seems to stick at a full treble tone, little or no reaction to picking dynamics, slight swirly sound like a slow phaser is on.  Maybe this is normal for this circuit, but not what I would expect.  I tested my range pot and it seems fine.

After looking into the AMZ Dr Quack, I decided to try the JFET buffer in front.  It maybe sounded a little better overall, but still had the range control weirdness.  I swapped a few more values over to Dr Quack specs, and wound up with distortion.

So I pulled it off my breadboad, screw it.  An envelope filter isn't exactly a priority for me, but thought it would be fun to try.  I have a Dist+ that I'm not using at all, wouldn't mind gutting it for some other 2-knob build that I would actually use.  But not this one, not today anyway.

[thermionix]

I just watched a demo of the EHX Dr Q, and apparently what I was hearing is actually normal behavior for this circuit.  Very different from the DOD FX-25, where (IIRC) a high Range knob setting yielded a very large sweep from bassy to trebly.  Here, the most actual range seems to be in about the middle of the Range control.  Counterintuitive.

[Mark Hammer]

What is labelled as "Range" is intended to be a trimpot, not a control.  It can be adapted to be a control if you add fixed resistors of suitable value to each side of a 10k pot since, as you found out, only a bit of the pot/trimmer's rotation actually does something useful.  But in general it is made to dial in optimum sweep for each transistor.

The Quack was designed by Jack Orman to sidestep the requirement to use an LM1458,  The stock Dr, Q circuit will not work properly with any other dual op-amp.  I know because I've tried a bunch of different ones.

I recommend tinkering with the decay time,and leaving the Attack time resistor at 68-100R.  Also, you will note the Nurse omits the 1N914 diode.  Use a Schottky of some type (1N5817, BAT 41, etc.) instead of a 1N914, and place a resistor/resistance in parallel with the 22uf cap (which the Nurse shows as 10uf).  A small resistance, like 22k will give a very fast more synth-like decay, while a larger value, like 470k-1M will give a much longer decay time.

Finally, the sweep will be a function of the change in resistance from ground to the junction of the two 4n7 caps.  That includes the 47k fixed resistor and the transistor's emitter/collector resistance.  You can change how wide the sweep is, independent of the sensitivity, by playing with that total resistance.  The sensitivity does dictate the width of sweep, but you could set the min/max resistance to have a filter that triggers easily and always, but only sweeps over a narrower range, instead of something that goes higher than you want.

[thermionix]

Ah, thanks Mark!  That explains a lot.  I have this thread bookmarked for if/when I have another go at it.

[jimladladlooklike]

Thanks for the reply!

The voltage reading at the junction of the attack pot and 22uF cap was 0.02v when strumming...

Do you mean basically a low pass filter at the output? Or some kind of filter effect similar to the one we're aiming for? Either way, it doesn't seem to be doing either of those things I'm afraid.

The filter is a bandpass, as far as I know, but when the resistance to ground from the junction of the two 4n7 caps is high, then the passband is low.  The envelope voltage fed to the transistor base makes the transistor operate like a voltage-controlled resistor in parallel with the 47k fixed resistor.  It makes the emitter-collector resistance lower, which raises the range of the passband.  The 22uf cap "hangs onto" the envelope voltage for a little while, such that the decay is slow enough to hear the filter sweep beck down again.

If you temporarily wire up a variable resistor (e.g., a 100k pot) in parallel with that 47k resistor, you should be able to make the filter move up and down by adjusting that variable resistance.

As for the rectifier/follower portion, 20mv seems pretty low to me.  Are you measuring that with the Sensitivity control up full?  If so, verify the signal path by measuring the AC voltage at the wiper of the Sensitivity pot, after the 47k, then after the .01uf cap, and finally at the output of that op-amp section.  Then switch to measuring DC, and measure after the LED on the output of that op-amp section.  All of this is while you feed an audio signal to the input jack (ideally strumming your guitar, but you can use other sources).  You should see something at each of those points that corresponds to the intensity/loudness of the audio input signal.

Okay so the AC voltages at the wiper (the middle pin, right?) and after the 47K resistor are both 0, but the after the .01uF cap it's 1.9v and the output (pin 1) it's 2.9v... Weirdly, strumming the guitar has no effect on this voltage. It's constant.

The DC voltage at the cathode of LED2 is 0.01v but strumming also has  no effect on this.