Cgs76 Envelope Generator Help

From circuitbending to homebrew stompboxes & synths, keep the DIY spirit alive!

Moderators: luketeaford, Joe., lisa, Kent

User avatar
yan6
Ultra Wiggler
Posts: 953
Joined: Sat Jun 15, 2013 7:14 pm
Location: Ontario, Canada

Cgs76 Envelope Generator Help

Post by yan6 » Mon Sep 02, 2019 6:52 pm

I'm working on a cgs76 for the first time and having some issues with the rise portion. I dont seem to getting any slope on the rise. If I turn it all the way down, the generation of wave will stop completely, once its turned back up 1/3 or 1/4 of the way it starts to cycle again but never any slope adjust on it, just straight up as per the pics. I have also include a shot of the pcb in case you see anything.

For the 2n3565 I have pn3565 in backwards
For the 2n3638 I have bc557 in normal
For the 2n4248 I have bc557 in normal
For the 2n3391a I have one bought from mouser in backwards. This seemed correct according to the data sheet.

Image

Image

User avatar
guest
Super Deluxe Wiggler
Posts: 4318
Joined: Mon Aug 19, 2013 11:49 am

Post by guest » Mon Sep 02, 2019 8:56 pm

measure the voltage on the fall and rise pot wipers, and make sure they do the same thing.
openmusiclabs.com

User avatar
guest
Super Deluxe Wiggler
Posts: 4318
Joined: Mon Aug 19, 2013 11:49 am

Post by guest » Mon Sep 02, 2019 9:01 pm

shouldnt one of thoe ICs be an MC3401? or maybe that doesnt matter because its the same thing as an LM3900?
openmusiclabs.com

User avatar
guest
Super Deluxe Wiggler
Posts: 4318
Joined: Mon Aug 19, 2013 11:49 am

Post by guest » Mon Sep 02, 2019 10:39 pm

check that the resistor going to pin13 of the "MC4301" is 82k. also, you can try swapping the LM3900s to see if it makes any difference.
openmusiclabs.com

User avatar
yan6
Ultra Wiggler
Posts: 953
Joined: Sat Jun 15, 2013 7:14 pm
Location: Ontario, Canada

Post by yan6 » Sat Sep 07, 2019 10:34 pm

Okay worst response time ever :oops: I finally got some time to spend on this and I haven't yet made any progress. I've tried swapping both lm3900's numerous times, double check the 82k.

What I can say for sure, is there isnt any change with the sweep of the rise pot. I'm always sitting at 5.8v at the wiper :hmm:

I changed the pot and have checked the wiring

User avatar
guest
Super Deluxe Wiggler
Posts: 4318
Joined: Mon Aug 19, 2013 11:49 am

Post by guest » Sun Sep 08, 2019 12:34 am

sounds like the pot isnt grounded. turn off the power, and measure the resistance to ground from the pot pins. one should be at 0ohm, another at 100k, and the middle should go between those two as it turns. be sure to connect your negative probe lead to some ground point far away on the circuit, and not right at the pot or its wiring.
openmusiclabs.com

User avatar
yan6
Ultra Wiggler
Posts: 953
Joined: Sat Jun 15, 2013 7:14 pm
Location: Ontario, Canada

Post by yan6 » Sun Sep 08, 2019 9:06 am

I seem good with ground on the pot. I took my readings from the pot and 0v at the power header. One side is 100k and the wiper sweeps from 4ohm to 90k. I also confirmed the same range with the fall side.

User avatar
guest
Super Deluxe Wiggler
Posts: 4318
Joined: Mon Aug 19, 2013 11:49 am

Post by guest » Sun Sep 08, 2019 9:16 am

and you still cant get it to go to 0V on the wiper? if its 4ohms to ground, it has to be pretty close to 0V. can you try your wiper voltage test again, but this time using the same ground and wiper points you used for your resistance tests?
openmusiclabs.com

User avatar
yan6
Ultra Wiggler
Posts: 953
Joined: Sat Jun 15, 2013 7:14 pm
Location: Ontario, Canada

Post by yan6 » Sun Sep 08, 2019 9:31 am

Tried the voltage test as suggested and I couldnt see any change, my meter was 2-3mv. One thing I find strange is pin 9 of the lm3900 is wave that follows the output (yellow trace is pin 9 and blue trace is the output)

Image

But the other side of the 15k, so at the 100k and 15k junction also the rise emitter junction, I get a solid voltage.

Image

User avatar
guest
Super Deluxe Wiggler
Posts: 4318
Joined: Mon Aug 19, 2013 11:49 am

Post by guest » Sun Sep 08, 2019 10:12 am

ok, so its 3mV at the wiper when full down? that sounds correct.

the new scope shots help a lot. the voltage at the 15k/100k juncture will only toggle when output goes above 6V, so i think youre getting closer to the problem. the output is not going high enough, and is resetting early. this is most likely a problem up in the "cycle" circuitry. a backwards or faulty diode would be likely. check to see if the "end" pulse is happening at this reset point.
openmusiclabs.com

User avatar
wackelpeter
Veteran Wiggler
Posts: 515
Joined: Fri Apr 17, 2015 5:56 pm
Location: Germany, MD

Post by wackelpeter » Sun Sep 08, 2019 2:10 pm

okay, maybe not of much help, but i tried to wrap my head around your problem.

That's my wild assumption i came off with: I suspect a short somewhere at the side of the 470k resistor (which goes to pin 3 of the LM3900) on the side that's going to the wiper to the 6V line ... as when you turn the pot, you always measure those roundabout 6V and when you crank the pot up you would short this to gnd, hence it totally stops cycling then and only starts again if there is enough resistance between gnd and the 6V when you turn the pot back...

Have you disconnected the side of the pot which goes to the 2N4248 for the rise and measured at it's wiper? Do you still have 6V there then?

Edit: well doesn't seem to make sense as you said it always hooks at the 6V and my theory with shorting it wouldn't make sense and also you would have already smelled it...
I can perhaps tomorrow try to take some measurements on my stripboard build... i hvae btw. successfully used some regular BC547 and BC557 as supposed alternative on Ken's site.

User avatar
guest
Super Deluxe Wiggler
Posts: 4318
Joined: Mon Aug 19, 2013 11:49 am

Post by guest » Sun Sep 08, 2019 4:20 pm

the 470k short is a good idea. the transistor is current limited by the 15k, so it wont smoke anything, but will be in heavy saturation at the 3900. and turning it down to ground also forces the output to saturate, as it prevents any current from going into the inverting pin.

set up your scope to on the rising edge of the output, and then look at the voltage on pin9. zoom in to 1us per division or so, and see if it spikes up to 10V for about 1us at that transition.
openmusiclabs.com

User avatar
yan6
Ultra Wiggler
Posts: 953
Joined: Sat Jun 15, 2013 7:14 pm
Location: Ontario, Canada

Post by yan6 » Sun Sep 08, 2019 4:59 pm

As described, the yellow trace ( ch1 ) is pin 9 of the lm3900 and the blue trace ( ch2) is the envelope output, trigger is set to the rising edge of ch2.



Image

Also at pin 9 with ch1 I see these random spikes which I'm guessing show up as below on the end output.

Image

I dont seem to get any end pulses when its cycling, except for these small random spikes.

Image

If it isnt cycling and I hit start with a trigger the end output dips low

Image

User avatar
guest
Super Deluxe Wiggler
Posts: 4318
Joined: Mon Aug 19, 2013 11:49 am

Post by guest » Sun Sep 08, 2019 5:12 pm

with the end cycle and pin9 voltages, try looking at them at 10us per division, with the trigger point set at the middle of the screen horizontally. pin9 looks like its going up to 10V for the rise portion, which would indicate a bad 470k on the pot wiper. turn the power off and measure the resistance from the pot wiper to pin3 of the 3900.
openmusiclabs.com

User avatar
yan6
Ultra Wiggler
Posts: 953
Joined: Sat Jun 15, 2013 7:14 pm
Location: Ontario, Canada

Post by yan6 » Sun Sep 08, 2019 5:47 pm

I'm hoping this is the scope shot you were asking for? Yellow is pin9 blue is output

Image

I changed the wiring from the rise pot, just to be sure it's not there...no change.

I measure from the wiper to pin 3 with the wires out and have 469k

With the wires in, from the wiper to pin 3, I see a range from 169k - 206k, depending on the pot. I confirmed this with the fall pot.

User avatar
guest
Super Deluxe Wiggler
Posts: 4318
Joined: Mon Aug 19, 2013 11:49 am

Post by guest » Sun Sep 08, 2019 5:50 pm

what happens if you disconnect the rise pot wiper entirely? does it rise at all?

also, is the above scope shot pin9 and the output? if so, can you go out to 100us? i want to see the start of the rising edge on pin9, to see what makes it go high.

also, how do you have it setup so it cycles?
openmusiclabs.com

User avatar
yan6
Ultra Wiggler
Posts: 953
Joined: Sat Jun 15, 2013 7:14 pm
Location: Ontario, Canada

Post by yan6 » Sun Sep 08, 2019 6:23 pm

Disconnected the rise wiper with no change

You are correct the above scope shot was pin 9 ( yellow) and output (blue)

This is the same shot at 100us


Image

The cycle switch takes a copy of the end jack and a copy of the cycle jack and shorts them together via a spst on-off switch. [/video]

User avatar
guest
Super Deluxe Wiggler
Posts: 4318
Joined: Mon Aug 19, 2013 11:49 am

Post by guest » Sun Sep 08, 2019 7:31 pm

does the fall wiper work as it should? what does the end pulse look like? does it line up with the pin9 pulse above?
openmusiclabs.com

User avatar
yan6
Ultra Wiggler
Posts: 953
Joined: Sat Jun 15, 2013 7:14 pm
Location: Ontario, Canada

Post by yan6 » Sun Sep 08, 2019 8:50 pm

Yes the fall wiper seems to works as it should.

This first shot is with an ext trigger on the start input ch1 is yellow and is pin 9. Ch 2 is blue and is the end output

Image

On the next shot, ch1 (yellow) is pin 9 and ch1 (blue) is the end output. This is with the cycle switch on. Note that I dont really see the blue spikes unless the fall rate is quite high.

Image

User avatar
guest
Super Deluxe Wiggler
Posts: 4318
Joined: Mon Aug 19, 2013 11:49 am

Post by guest » Sun Sep 08, 2019 9:17 pm

what does pin5 of the 3900 look like with respect to pin9? the 100us per division scope shots give the best information. im curious if it is resetting fully.
openmusiclabs.com

User avatar
yan6
Ultra Wiggler
Posts: 953
Joined: Sat Jun 15, 2013 7:14 pm
Location: Ontario, Canada

Post by yan6 » Sun Sep 08, 2019 9:30 pm

Ch1 (yellow) is pin9 and ch1 (blue) is pin5

Image

User avatar
guest
Super Deluxe Wiggler
Posts: 4318
Joined: Mon Aug 19, 2013 11:49 am

Post by guest » Sun Sep 08, 2019 9:35 pm

ok, theres the problem, its not resetting. so now to figure out why. take a look at those transistors on the reset circuit. ill see if i can guess something from the shcematic.
openmusiclabs.com

User avatar
guest
Super Deluxe Wiggler
Posts: 4318
Joined: Mon Aug 19, 2013 11:49 am

Post by guest » Sun Sep 08, 2019 10:05 pm

there are 2 possible reasons its not resetting: the transistors arent drawing enough current or they are not on for enough time. if the transistor switched on instantaneously and was a dead short, that 33ohm resistor would limit the discharge time to 60us or so, which is about the length of the reset pulse. this would require a peak current of 200mA to go through that transistor.

check the 220pF and 470k on the MC3401 and make sure they are the right value. then take a look at the timing between the end cycle output and pin5.
openmusiclabs.com

User avatar
yan6
Ultra Wiggler
Posts: 953
Joined: Sat Jun 15, 2013 7:14 pm
Location: Ontario, Canada

Post by yan6 » Sun Sep 08, 2019 10:08 pm

:bananaguitar: :bananaguitar: :bananaguitar: :nana: :nana: :bananaguitar: :bananaguitar:

I replaced the pn3565 with 2n3904 (backwards) and it works

Image

I think for good measure I best replace the other 2n3565's as well

Guest, I cant send you enough praise. Your knowledge and support has been phenomenal :tu:

After I replace the rest I'm going to have a second look at the rest of its functionality, tomorrow probably

User avatar
guest
Super Deluxe Wiggler
Posts: 4318
Joined: Mon Aug 19, 2013 11:49 am

Post by guest » Sun Sep 08, 2019 10:34 pm

the other 2N3565s dont need to carry much current, so it probably isnt necessary to replace them. that transistor was supposed to a 2N3566, which im assuming is a higher current device. the 3904 is a pretty decent transistor for current handling.

im glad you got it fixed. thats a crazy circuit, i have no idea why it was designed that way. its sort of a rube goldberg machine of envelope generators. it probably has something to do with the devotion to the 3900.
openmusiclabs.com

Post Reply

Return to “Music Tech DIY”