Nurse Quacky Problem

[Alex C]

Hello.  I breadboarded the Nurse Quacky after hearing a lot of good things about it and looking over the other similar designs.  This one seems simple yet effective.
So I got it all wired up, using a TL072 and the suggested 2n3904, and the LEDs light up as mentioned by others (one stays on and the other lights up and dims according to the signal strength-pretty neat just to play and watch that- by the way would it be possible to put that LED outside the box for a neat-o visual effect?) but when the signal does get strong and the LED lights up, the output is muted, then as the signal fades, the output gradually raises.  If I play softly I can keep it from cutting the signal off.  It's neat for a few minutes, but I want my quack!

I traced the schematic twice, checked the orientation of the transistor and polarized caps (and LEDs, although they wouldn't light if installed incorrectly), and I tried the pots in many combinations.
I only had a few deviations from the schem:  I used a 100k pot instead of the 25k range and a 100k instead of the 1k attack (I know that this is a big difference, so I tried to keep it at low resistance- later I might try adding resistors to the pot to lower the value) .  Also, I didn't have a 51 ohm resistor so I tried a 100 ohm and a direct connection, and none of these worked.  

I am easily frustrated by inanimate objects, and especially things like these circuits, where I just KNOW I'm right, so why doesn't the darn thing work?  But my frustration makes me want to work hard to fix it instead of giving up.  

So any ideas or help would be greatly appreciated.  Thanks!

Alex

[B Tremblay]

I only had a few deviations from the schem:  I used a 100k pot instead of the 25k range and a 100k instead of the 1k attack (I know that this is a big difference, so I tried to keep it at low resistance- later I might try adding resistors to the pot to lower the value) .  Also, I didn't have a 51 ohm resistor so I tried a 100 ohm and a direct connection, and none of these worked.

Replace your substituted parts with the correct values as indicated by the schematic.  The incorrect range pot is most likely the main cause of your troubles.  However, the attack pot is also way too big for any useful adjustment.

[petemoore]

Or twisting fixed resistors between 1k through less than 1k for instance.
Or use sockets, and pput in resistors of different value...
for the 25k pot you could try 20k pot or two ~12k resistors in parallel the middle of the pair being the 'wiper' and the outside legs of the pair being the outside lugs of a pot....or... a 50k pot with a 47k resistor across the outside lugs makes a nice werkalike [worked for me and I think just like] 25k pot.
 The best thing is probably to order some somponents from Mouser, and just set the ckt. aside till they come, then try it with the right parts.

[Mark Hammer]

You appear to have a severe case of secondary undersignalitis.

The status of the two LEDs means you have power and are detecting a signal of sufficient strength.  However, it does not guarantee that the rectified signal from the envelope follower is reaching its intended target or is appropriately conditioned once it reaches the target.

Although I'm largely responsible for the recommendation to use a 1k attack-time pot, 1k is probably still too big since it requires a change in sensitivity as you increase the resistance beyond about 500R.  You CAN go bigger than 1k, but the amount of current lost by that big resistance means the transistor undergoes very little sweep.  Perhaps more importantly, given the decay-cap value, most of the action with respect to meaningful attack-time changes is happening between about 33R and 330R.  You need to go much higher than 330-500R to get any increase in attack time that you can actually detect.   Since those sorts of resistance values require an adjustment of follower sensitivity, it starts to become a nuisance aiming for extended attack time.  How much of a nuisance it is to have to keep adjusting sensitivity for different attack times will depend on the user.  Certainly if the incoming signal is weak enough that the sensitivity control has to be dimed to hear any appreciable sweep, then installing an attack resistance of larger than 500R is probably unwise.  Just bear in mind that a 1k pot with a 470R resistor in parallel and a basic 47R fixed series resistor is a very nice compromise.

As for the tuning trimpot, if you mount it on the chassis and play with it, you'll quickly find out that maybe 1/3 of its rotation - if that much - is of any use.  What this tells me is that subbing a 10k pot and a pair of 6k8 resistors (one on each end) is probably a perfectly acceptable substitute.