Zombie Chorus -- LFO doesn't Low frequency oscillate.

[morcey2]

Hi all.

I built a zombie chorus from the schem/layout on geofex:  http://www.geofex.com/PCB_layouts/Layouts/zombie.pdf

At the moment, it is a fine and dandy fixed-very-short delay pedal.  I hooked it up to the scope and the MN3007 and CD4046 look like they're doing their thing just fine.  The depth pot changes the frequency coming out of the 4046.  The clone-mode switch works great.  it cuts that frequency in half.  But ... the LFO doesn't seem to be oscillating at all.  Here's the voltages for U3:

Vb is 4.97V and the battery voltage is 8.25.  I tried with a new battery and the behavior is the same, but I didn't write down the voltages from that. 

pin 1: 1.36V
2: 4.97
3: 3.96
4: 0.00
5: 4.98
6: 1.34
7: 7.57
8: 8.25

Here's something odd.  Pins 1,2,3,4,5, and 8 have a very high frequency spike type pattern on them matching the frequency of the output of the 4046.  pins 6 and 7 have a semi-sawtooth pattern on them at the same frequency.  Why I'm seeing anything on pins 4 and 8 has me quite worried, being that they are ground and +9 respectively.  I've checked the grounds all the way around the board and they seem to be fine.  I assume that if the LFO was LFO'ing, the frequency output from the 4046 would change as the voltage from the LFO changes, and that's not happening.  The Speed pot also does nothing. 

It does have a pretty cool sound right now with the really short delay.  The delay part is very obvious on the scope. 

I'm using TL072s and have tried several, not the recommended TL062s, but I think the TL062 is specific to U1 and the LFO should work with pretty much any opamp, if I've read correctly. 

Any ideas or similar experiences? 

Thanks,
Matt

[oldrocker]

I don't have any answers for you.  I do have a question.  I'm assuming this was a PCB built from the Geo site.  If so I posted a pic of where I thought there might be issues for this board.  I built mine on perfboard so this was not an issue for me.  Did see any problems with your etched PCB like what I posted?  Just a thought.  This may not be your problem at all.  I'm just curious how your PCB came out.

http://www.logmonster.com/images/full/?pic=Zombie_1185718672-422-15307.jpg

[morcey2]

Oldrocker,
On the spot where the pad is too close to the other trace, I scraped off a wee bit of the PNP before etching and there's no continuity between the traces.  With the 2 dead-end traces, I think they're there for a reason, but I'm not sure why.  Looking at the schematic, they don't look like they're supposed to be connected to anything else.   The PCB turned out quite good, considering it's the first batch of PCB's I've etched.  At least until I drilled it.  I scoured off 2 of the pads with the chuck of the drill press by getting a little over-zealous when drilling.  Fixed that by running the lead a little long and soldering it to the trace.  Not the best solution, but I've had to do it on production boards that have had pads pull off or otherwise disintegrate. 

I'm pretty sure it's going to be something really simple and I'll smack myself in the head after finding it.  I'll try to get pictures posted tonight for a little more insight.

Thanks,
Matt

[oskar]

Vb is 4.97V and the battery voltage is 8.25.  I tried with a new battery and the behavior is the same, but I didn't write down the voltages from that. 

Vb should be 60% of battery R12 is 10k and R13 is 15k   5v is spot on...

oskar

[morcey2]

Vb is 4.97V and the battery voltage is 8.25.  I tried with a new battery and the behavior is the same, but I didn't write down the voltages from that. 

Vb should be 60% of battery R12 is 10k and R13 is 15k   5v is spot on...

oskar

I probably should have seperated those 2 sentences.  They didn't have anything to do with each other. .  The behavior is that the LFO doesn't work.  I've taken some pictures, but haven't had a chance to upload them.  Should to that tonight if I can find the USB cable to the camera.

Matt

[R.G.]

Your problem is that there is some difficulty with U3a not flipping state.

Here's the analysis:
a. Section U3a is sitting with its output at 1.36V on pin 1.
b. This pulls current out of pin 6 through the speed pot and pin 17. Pin 6 is lower than pin 5, so this forces pin 7 as high as the U3b output can go, 7.57V.
c. With pin 7 high, this should pull pin 3 higher than pin 2 to force the comparator formed by U3a to flip states and start U3b integrating down.
d. But with pin 7 fully high, pin 3 is only being pulled up to 3.96V. That's why the LFO won't oscillate.
e. We can use Ohm's law to figure out more. With 7.57V on pin 7 and 3.96V on pin 3, the current through R15 is (7.57-3.96)/47000 = 76.8uA. Since opamp inputs can't source or sink current on the pins themselves, this current is going somewhere. Maybe it's going through R14; pin 1 is 1.36V and pin3 is 3.96V, so the current through R14 is (3.96-1.36)/100k = 26uA. Nope, not there. So we have a place to look. Either or both of R14 and R15 may be the wrong value; if they're correct, something is stealing current from pin 3 that shouldn't be, and that's keeping pin 3 from rising high enough to flip the comparator and make the LFO work. It might even be a bad pin 3, sucking current in. Not likely, but possible. The usual suspects also weigh in here: bad solder, shorts, etc.

Find out which of these it is and the LFO will work. Mother Nature says so.

[morcey2]

The usual suspects also weigh in here: bad solder, shorts, etc.

Find out which of these it is and the LFO will work. Mother Nature says so.

R.G.
It couldn't be shorts.  I don't wear them.  My legs are far too white. 

I'll check that out when I get home.  Probably my soldering job.  I'm currently using a 30 watt RatShack iron that only has the huge tips.  I'm getting a better iron today that can use much finer tips.  The big tips work well on tube amps, but it's really hard to solder just 1 pad on a pcb with something that big. 

Thanks,
Matt

[oskar]

Vb should be 60% of battery R12 is 10k and R13 is 15k   5v is spot on...
oskar

I probably should have seperated those 2 sentences.  They didn't have anything to do with each other. . 
Matt

And I should have my left and right brain halves separated.... I don't remember this post at all and I can't see what motivated it    but hey, math is fun!
oskar

[morcey2]

Here's some pics.  I need to clean the flux off the board because it's reflecting light in the picture and looks a little like solder.  Like I said before, the tip that I've been soldering with is a little too large for this resulting in the messy soldering.  I've been over it and haven't been able to find any bridges. 

trace side:


component side:


Thanks,
Matt

[R.G.]

Your problem is that there is some difficulty with U3a not flipping state.

Here's the analysis:
a. Section U3a is sitting with its output at 1.36V on pin 1.
b. This pulls current out of pin 6 through the speed pot and pin 17. Pin 6 is lower than pin 5, so this forces pin 7 as high as the U3b output can go, 7.57V.
c. With pin 7 high, this should pull pin 3 higher than pin 2 to force the comparator formed by U3a to flip states and start U3b integrating down.
d. But with pin 7 fully high, pin 3 is only being pulled up to 3.96V. That's why the LFO won't oscillate.
e. We can use Ohm's law to figure out more. With 7.57V on pin 7 and 3.96V on pin 3, the current through R15 is (7.57-3.96)/47000 = 76.8uA. Since opamp inputs can't source or sink current on the pins themselves, this current is going somewhere. Maybe it's going through R14; pin 1 is 1.36V and pin3 is 3.96V, so the current through R14 is (3.96-1.36)/100k = 26uA. Nope, not there. So we have a place to look. Either or both of R14 and R15 may be the wrong value; if they're correct, something is stealing current from pin 3 that shouldn't be, and that's keeping pin 3 from rising high enough to flip the comparator and make the LFO work. It might even be a bad pin 3, sucking current in. Not likely, but possible. The usual suspects also weigh in here: bad solder, shorts, etc.

Find out which of these it is and the LFO will work. Mother Nature says so.

[gez]

I've been over it and haven't been able to find any bridges. 

RG has isolated the area where the problem lies.  Concentrate your efforts there.

Even microscopic bridges can have low enough resistance to bugger up things, so just because you can't see them doesn't mean they're not there and causing a problem. Try scraping between traces with a scalpel.  Quite often, this is all it takes to get a circuit up and running.  The other usual suspect is dry joints: just reflow them.  These two problems account for around 95% of my builds that never fire up first time.  The other 5%?  Main reasons are: a trace is broken (check all traces with a continuity meter before you solder - it saves so much time in the long run); or I've missread a component value; or left off a trace when designing the PCB /something equally as dumb.

[gez]

(check all traces with a continuity meter before you solder - it saves so much time in the long run)

The other thing I do before I start soldering, is check for resistance between unconnected traces (it should be infinite).  This tells you if there are any bridges as a result of a poor etch.  I used to check for continuity, but I've learned the hard way that not all bridges have low resistance (back to those microscopic buggers again) so they slip under the radar.

Although this might seem time consuming, it can save hours spent on debugging (if a circuit is quite complicated).

[Mark Hammer]

I have also found that sometimes what appears to be a properly seated IC in a socket is but mere illusion.  Something as simple as using a better quality socket, or scraping the tarnish off the IC pins, or (circumstances permitting) tinning the IC pins a bit to "thicken" them and provide better contact, can sometimes turn a non-working thing into a working one.

That advisory about pin-to-socket contact goes for the LFO AND the 4046 whose clock frequency is supposed to be responding to the LFO.

[morcey2]

Finally got a chance to work on it.  It was a mis-labeled resistor.  I ran out of 47k resistors, so I made some out of 38K + 8.1k.  Both of the 8.1k were off the same strip, but when I pulled them and measured them, one was 81k.    Made a 44k out of 2x 22k resistors (I measured them this time) and it works wonderfully.   OK, mostly wonderfully.  It does have the "tick" that the zombie is famous for, which I will fix soon.  I don't want to mess with it right now though.  Gotta start etching the VibraTone and get what I've built so far in enclosures.

Thanks everyone and specifically R.G.

Matt.

[Paul Perry (Frostwave)]

Thanks for letting us know how it turned out Matt!
Nothing worse than being left hanging after a saga like this.

[R.G.]

Kewl dood.