Debugging a Neutron

[taeagan]

So my Neutron is very blatty/farty sounding.  In bandpass it sounds OK.  In highpass or lowpass it's just about un-usable.  Sometimes it has no effect at all.  Sometimes you have to hit the strings really hard to get anything.  Other times it starts to get really distorted and farty as the note rings out.  Also, when the gain is cranked up, it gets very distorted.  I used the H11F3 option and did not use the power converter - I'm using two 9 Volts.  I've quadruple checked the polarities, resistor values, wiring etc...  It would seem to me that an opamp or opamps are not biased properly.  I'm getting +9V and -9V to all the right places (pins 4 and , but I don't know what voltages I should be looking for on the other pins.  Any advice?

[mlabbee]

What kind of optocouplers are you using?  I used NSL32's and found that the range given for Rx was way off - the build instructions say it should be no greater than 4k7, but I have a 4k7 and a 5k pot in series, with the 5k trimpot cranked - so I'm pretty close to 10k for Rx and it barely keeps the sweep under control for most of the settings.  Try putting in a fairly large trim pot (maybe 25k?) and cranking the resistance at Rx to get the sweep under control.

[taeagan]

I'm using the H11F3 photocoupler option.  Also, I have a 2.5k trim pot for Rx, but cranking it one way or the other doesn't change things.  

But I really doubt that's the problem.  It really sounds like a problem with the way that the opamps are biased.  With the gain turned even 2/3 of the way up it sounds like I'm playing through a cranked tube screamer.

[R.G.]

"fartiness" is nearly always a bias problem. Can you post all the voltages on all the opamp pins?

[taeagan]

Is that fartiness or fartyness?  I never could figure that one out...

Can't post the voltages from work, I'll do it from home tonight.  One thing is for sure, +9 and -9 are definitely being delivered properly.  If I remember correctly, most of the 1-3 and 5-7 pins on U1 and U2 showed up with 0V or in the 50mV range.  U3 had higher voltages (1.5V and up) that also varied when the switches were thrown.

[taeagan]

Here are the voltages:

U1
1: 0V
2: -50mV
3: -50mV
4: -8.9V
5: 0V
6: 0V
7: 0V
8: 8.8V

U2
1: -95mV
2: 0V
3: 0V
4: -8.9V
5: 0V
6: 0V
7: -1.9V

U3 Direction Switch Down
1: 6.1V
2: 1.8V
3: 3.3V
4: -8.9V
5: 0V
6: 0V
7: .5V
8: 8.8V

U3 Direction Switch Up
1: -7.3V
2: 0V
3: 0V
4: -8.9V
5: 0V
6: 0V
7: 0V
8: 8.8V

Any ideas?  A lot of zeroes there, but from what I can tell from the schematic, all of those points appear to be isolated from any DC voltage.

[taeagan]

I spent several more hours on this last night and I'm completely stumped.  Everything is grounded properly.  I checked the leads running from each pin of each opamp and verified with my meter that they are properly connected to the appropriate components.  No disruptions in any of the traces and no traces shorted together.  I also tried substituting opamps.

[R.G.]

That one is tougher than the normal bug.

The zeros you see are OK, I think. This circuit runs from + and - power supplies, so the correct bias voltage is at 0V, where most of them are. The under-50mv ones are caused by DC offsets in the opamps, and are not problems - I think.

Pull out the H11F3's. Now try it in bandpass, hipass and low pass. Still farty?

[taeagan]

I'll try this tonight.  I did get some NSL32's yesterday and tried them in place of the H11F3's.  It was worse, still a little farty and much quieter/dead sounding.  

What should I look for if it's still farty after pulling out the H11F3's?  And what should I do if it's not farty when I pull them out?

Thanks for all of your help!

[Karmasound]

Anyone built this with the cm6000.


i'll do a search later but i'm getting ready to build this and this thread is right on time.


this place rocks

[R.G.]

I'll try this tonight. I did get some NSL32's yesterday and tried them in place of the H11F3's. It was worse, still a little farty and much quieter/dead sounding.

What should I look for if it's still farty after pulling out the H11F3's? And what should I do if it's not farty when I pull them out?

Pulling the H11F3s is purely diagnostic. That way, the envelope detector can't affect the filter.

What I'm after is why it sounds farty if the voltages at the pins of the audio path seem OK - as they do. You say that it changes nature from high- to band- to low-pass, so pull the H11s, then try it for fartiness in all three positions. There will not be any sweep, as pulling the H11s will disable that. So we see if it's still farty, and on which settings, and then I contemplate my navel for why that should be so.

[taeagan]

RG - we're on to something.  Pulling the H11F3's eliminates the fartiness.  It sounds basically like a wah in a fixed position on all three settings - as I imagine it should.  

I really don't know what to look for next.

[R.G.]

RG - we're on to something. Pulling the H11F3's eliminates the fartiness. It sounds basically like a wah in a fixed position on all three settings - as I imagine it should.

Yep, that's what I hoped would happen.

I really don't know what to look for next.

Mother Nature is telling us that voltage and/or current is getting from the envelope circuit through the optoisolators into the audio filters. Being her normal cryptic self, she's letting us figure out for ourselves just **how**.

The most likely thing is a solder or copper trace short from the LED side of the H11 to the FET side. If they're socketed, maybe a fault in the socket. An unlikely but possible explanation is one or two bad H11F3s, but that's probably a last choice.

I think it's ohmmeter time - try the ohmmeter to figure out whether traces that are supposed to connect do and ones that are not supposed to don't, in the area under the H11's. If that fails to turn it up, we'll have to dig deeper.

I looked back at the thread, and maybe just missed it, but is this one of my board layouts, or did you perf or vero this one?

[taeagan]

I started checking all of the pins on the socket with the meter as soon as I found that it was the problem area.  All of the pins on the socket are connecting where they should and none of them are shorted to one another.  I tested continuity to each component on the board that each pin connects to.  And it's not bad H11F3's, I've tried swapping them out already.

I'm using your board from the geofex article.  I've cloned quite a few pedals now and this is definitely the biggest challenge I've ever faced.

[R.G.]

I started checking all of the pins on the socket with the meter as soon as I found that it was the problem area. All of the pins on the socket are connecting where they should and none of them are shorted to one another. I tested continuity to each component on the board that each pin connects to. And it's not bad H11F3's, I've tried swapping them out already.

I'm using your board from the geofex article. I've cloned quite a few pedals now and this is definitely the biggest challenge I've ever faced.

Subtler problems are always the hardest.

So now we have to think.
Removing the H11s removes the fartiness. The printing is not bad on the PCB, and you don't (presumably) have a solder short. There are a couple of things that come to mind.
1) the H11s are much faster than an LDR. It's possible that they are following an LED transient and making it sound "farty" when it's actually working correctly in the audio path and the problem is filtering in the envelope detector. Rare, but possible.
2) The H11 could be leaking some voltage or current in a way that I haven't thought of yet.

With the H11 out of the board and power on, can you measure the voltage across the H11's? That is, from the LED pins to the FET pins?

Do you have a large value (250K-1M) dual pot? If so, try subbing it in for the FET side of the H11's. Do you get the filter sweep as you (or more realistically, someone else) turns the pot?

The next debugging steps will get more intrusive.

Since you have spare H11s, bend the leads out on the FET side so they don't go in the socket, only the LEDs go in the socket.

Now, try the pedal. We expect the same result as if the H11s are pulled out entirely, but if the fartiness is now there, we have to run down why the envelope detector is interfering with the audio path.

[R.G.]

Hey, wait a minute. I was looking at my layout files and wondering where the traces were too close, that kind of thing, and I notice that I used a pin on one of the H11s that is supposed to be a NC (see the thread on this recently in the forum). It's possible that it bit me.

If you read my comments on NC's in the forum, you know this is bad practice. I remember the thought process from doing the layout now - I didn't want to use that pin, tried to find a way around it, and finally did when I probed several H11F3s and found that they were the open variety of NC. That was a particularly tough part of the layout, no room for that ground trace and no room for a jumper, and I was NOT doggonit going to make the board bigger and have to relay 90% of it.

If you got H11s from another source, they may NOT be NCs on the middle pin of the FET side. Can you try the H11s but bend the middle pin of the FET side out so it does not go into the socket? If all is well then, I get to do some remedial remarks involving bending or clipping off that lead. If all is not well, we go back to slogging through the circuit and finding what's wrong.

H11F3s *have* been used in this thing, and not just in my proto, so this revelation was slow in coming; had to consider a lot of alternatives.

Try it and let's see.

[taeagan]

I was thinking about it this morning and that's what I was going to try next - try it with just one H11F3 or the other, just the LED side, just the FET site, etc...

I got the H11F3's from Mouser, so I imagine they're the same ones that you used.  I doubt it's the NC pin since I did try it with those NSL32's that time and they didn't solve the problem.  I'll let you know tonight.  And thanks again!!!

[taeagan]

Here's where I am right now:

- Bent the NC pins out of the way on the H11F3's.
- With just the LED sides (or just the FET sides) plugged in, it's the same as if they're not plugged in at all.
- Plugging in just one FET side (while both LED sides are plugged in) gives a slightly better result.  In the lowpass position, it's mostly normal sound and you can hear a little bit of the envelope filter trying to kick in.  In highpass, there's no or little sound when hitting the strings hard.  As the note rings out, you start to hear a farty sound.

I'll measure the voltages and try using a pot in place of the FET/LDR in a few minutes.

[taeagan]

Measuring voltages from the LED side pins to the FET side pins (with no H11F3's inserted) the only place where any voltage shows up is the first pin of the LED side (un-banded end, connects to pin 1 of U3).  7V from there to all pins on the FET side.  No other LED side pins show any voltage.

Sounds like a Mutron when I plug in two pots where the FET/LDR should go!!!!

[R.G.]

Sounds like a Mutron when I plug in two pots where the FET/LDR should go!!!!

That's good - we know that the audio part is OK.

- Plugging in just one FET side (while both LED sides are plugged in) gives a slightly better result. In the lowpass position, it's mostly normal sound and you can hear a little bit of the envelope filter trying to kick in. In highpass, there's no or little sound when hitting the strings hard. As the note rings out, you start to hear a farty sound.

OK, that's something to work on. Lemme go introspect with the schemo a bit.