Problem with OC-2 chopped

[isildur100]

I have built the OC-2 octaver chopped version on my breadboard. Everything works as expected and I am quite satisfied except for one thing. It looks like there is some kind of signal bleed somewhere in the circuit. A faint distorted tone is heard in the background. This distorted signal comes from the octave divider part of the circuit and creeps up to the clean signal output. Even if I turn the octave pot all the way down, I can still hear that undesired distorted signal in the back. It is hard to describe but imagine that when playing with a clean guitar tone, you hear a distorted version of your notes far in the background.

I can't find a link to the schematic of the chopped oc-2 but if you want I can post it.

Any help would be appreciated

John

[oldrocker]

I am getting parts together to build this myself.  I found an old schematic and once I solder up the EH Base Balls circuit I was going to to try the chopped OC-2.   If I can I'll post my build report and see if it suffers from the same symtoms or not.

[isildur100]

Cool! Thanks!

[isildur100]

Anyone else build the OC-2 chopped?? I really would like to solve this issue

Or, can someone point me to causes and remedies when part of a signal bleeds through other parts of a circuit when it should not?

This circuit uses a lm324 quad opamp and the distorted signal is generated by it in order to send a squarish wave to the 4013 chip. However, this signal should not be mixed at all with the output. It is only used to drive the 4013. The undesired signal is still heard even if I disconnect the output of the 4013 to the jfet (at R47). So it looks like part of the signal bleeds backwards (from R8). Is that possible?

Here the schematic I used (by Ben Milner):

[Processaurus]

Hmmm, good approach disconnecting the FET that mangles the clean signal into an octave down, that means that the distortion (really bleedthrough from the squarewave that's part of the tracking circuit) is either coming through the power rails or from crosstalk from sensitive input lines running physically near the logic section (everything around IC2 and IC7, the LM324 and flip flop chip, respectively).

For power rails, try a physically close .1 uF cap from the power supply (+) pins to ground on IC2 (pin 4) and IC7 (pin 14).  If that doesn't work, you could more aggressively decouple the power to the logic section with, from the (+) supply for the rest of the pedal, a series 47 ohm resistor and a 100uF cap to ground.

You can also keep the ground for that logic section (sewer ground) separate from the rest of the pedal (signal ground) and connect it right at the power supply (-) to kind of star ground things.

Last thing to try, maybe a different quad opamp wouldn't bang against the rails as hard.

If it's routing, and it's the very first input stage, a good way to check that is by running a buffered pedal in front of the OC2 and see if that gets rid of it.  If it does, then shielding the input wire is a good idea, as is routing wires as far away from the logic section as is reasonably doable.  Also watch the routing on the wiper wire on the clean level pot. 

Hope that helps, interested to hear how it goes...

[isildur100]

Hi Ben, thanks for replying!

I will try all your suggestions tonight when I get home from work. The thing is on the breadboard and since it is not the simplest build, there are wires and components sticking out everywhere and crosstalk is certainly possible. So maybe that is the cause.

As for the LM324, what do you mean by "bang against the rails as hard" ?

Can I use a TL074 instead of a LM324, of course adapting the different pinout ?

John

[isildur100]

I tried all your suggestions except for the star grounding or separate ground. I tried a tl074 instead of the lm324 (i found out they have the same pinout . There was no difference so I put the lm324 back.

I tried a buffer in front of it and it did slightly reduce the volume of the undesired distorted tone but did not get rid of it completely.

At least it looks like it is not caused by the 4013 because even if I removed the 9v connection to it, I still had the problem.
So the last thing I need to try is the star grounding but I it will be hard to do on the breadboard right now, unless I start all over, so I will try that when I perf it.

Somehow, the signal from the lm324 finds its way to the output. Is it possible that it could creep up backwards from the tracking section? I mean, go back out at R8 and up to the output?

John

[oldrocker]

I'm currently building this on a breadboard.  I noticed that the original schem shows a 100k on the middle wiper of both pots.  The 100k's then tie together at the other end feeding a 1uf cap.  The 1uf cap goes to the base of Q2 and also from the base of Q2 to a 1M resistor to half voltage vr.  I don't know if it will solve it but it's worth a try.

[isildur100]

Hi there, thanks for your reply!

I think I have tried that before and it did not change anything. I will try it again just to be sure.

What did improve things a bit was to use the jfet buffer at the input of the pedal. Or using a buffered pedal in front of it. (Was suggested by Processaurus)

I believe that my problems are mainly caused by the fact that it is on a breadboard with wires and components going everywhere Once I perf this neatly, it will probably solve things.

Tell me if you hear that undesired thin distorted signal in the background when you play. To really hear it, open the clean pot all the way and turn off the octave pot.

I will post my version of the OC-2 chopped, with the input buffer and a couple of changes, such as fixed misnamed pin numbers on the TL074.

John

[oldrocker]

Well I breadboarded this and I have to say, it's the cleanest octave diy pedal I've heard.  An  Very versatile octave pedal.  As for the distortion problem you're having I'm starting to think it's the nature of the beast.  I did notice removing the .0022 cap around pins 12 and 13 on the TLO74 made mine sound better.  So you might try that.

[Processaurus]

Didn't realize I'd mixed up pins 8 and 10 on both quad opamps, they're corrected now.  Does anyone know how to make the layout galleries not store old copies of images?  It would be nice to be able to replace schematics in old posts so people aren't confused... I told it not to preserve the original photo when I re-uploaded it.
Old one's link is this
http://www.aronnelson.com/gallery/main.php?g2_view=core.DownloadItem&g2_itemId=9447&g2_serialNumber=2

New one is:
http://www.aronnelson.com/gallery/main.php?g2_view=core.DownloadItem&g2_itemId=9447&g2_serialNumber=3

weird

[Processaurus]

Putting the 100K's from the pots through a 1uF cap and then re-biasing it with the 1M resistor doesn't gain anything, that stuff was part of the FET bypass switching.

Your circuit will almost certainly behave better when it is on a circuit board and doesn't have wires going everywhere all over a bread board.  You shouldn't need an extra buffer, to get identical performance to the boss pedal, definitely try it without once you've built the circuit.  And when you do, star ground the logic section IC2, IC7 back by the power supply, and that should get you the best power supply noise rejection.  No need to star ground everything, just keep the signal and logic sections ground separate except the one place they meet at the power jack.

[isildur100]

to oldrocker:
thanks! yes it is surely the best clean diy octave down  project. I tried removing that 2.2n cap you were talking about and I did not see any difference in sound or tracking, so I left it there.

to Processaurus:
thanks for the suggestion! btw, you understand that I am not experiencing power supply noise, but more like a signal leak. Would separate grounding eliminate signal leak or just noise?

One other thing, on that new schemo, the pins for the lm324 (8, 9, 10) are wrong, they were ok in the old one though The pins for the tl074 are ok now in that new schemo. Just thought i'd let you know.

John

[Processaurus]

One other thing, on that new schemo, the pins for the lm324 (8, 9, 10) are wrong, they were ok in the old one though Smiley The pins for the tl074 are ok now in that new schemo. Just thought i'd let you know.

Ha! Good eye, guess the artist who did the tattoo of the LM324 on my arm got the pinout wrong.




Power supply noise can happen from the audio, I've fixed distorted sounding pedals where you can see little spikes on the power supply at the frequency of the audio.  The grounding thing is good layout practice for anything, you see it in digital guitar pedals, where there are two ground planes that meet up at one point, one for the digital section, and one for the audio.

But probably the problem in yours is physical closeness of the audio to the logic sections and maybe loopy wires/long component leads.

[liquids]

I have built the OC-2 octaver chopped version on my breadboard. Everything works as expected and I am quite satisfied except for one thing. It looks like there is some kind of signal bleed somewhere in the circuit. A faint distorted tone is heard in the background. This distorted signal comes from the octave divider part of the circuit and creeps up to the clean signal output. Even if I turn the octave pot all the way down, I can still hear that undesired distorted signal in the back. It is hard to describe but imagine that when playing with a clean guitar tone, you hear a distorted version of your notes far in the background.

I can't find a link to the schematic of the chopped oc-2 but if you want I can post it.

Any help would be appreciated

John

Just a question...just saw this....why the aversion to the input buffer?  

Anyhow, yeah my experience is, it's the breadboard.  Lately I use my volume pedal, and dial it down a bit to control the breadboard a bit.  Especially with high voltage designs.   I have a high voltage/variable voltage design on the breadboard recently, and when running at highest voltages, I had to keep the 'volume' on the breadboard pretty low for unity gain...seemed touchy...just a little signal and it was blasting.  I then noticed that EQ, etc, seemed less effective, etc only to eventually discover that even with some 'missing connections' I was still getting some signal at the output!   Haha.  It comes  the territory of breadboarding sometimes...

Try this... whatever you use to send your signal 'out' from the output of the breadboard...pull that wire out, crank up the gain on your amp a little more than usual and hover the connection wire it around/above the breadboard holes...you may not even need to crank the gain on the amp, with some designs...on some design's I've done, you can hear the volume fading out just by getting close to the breadboard!   If so, that may be your problem.  Sometimes, what is happening is, filters running to ground on the breadboard are seding signal to ground...and that grounded signal is bleeding back into the board somehow.

Bottom line, most of that should disappear if you set it in solder.

and PS, if the 4013 can take it, you're probably better off running that whole thing off a bipolar supply, and ground will be a more stable VREF.    OR, add a op-amp VREF.   Or a 5v regulator for VREF of the op amp stages, rather than that cheesy voltage divider.  Should be more stable.    

[liquids]

Well I breadboarded this and I have to say, it's the cleanest octave diy pedal I've heard.  An  Very versatile octave pedal.  As for the distortion problem you're having I'm starting to think it's the nature of the beast.  I did notice removing the .0022 cap around pins 12 and 13 on the TLO74 made mine sound better.  So you might try that.

I've not built it myself, but from a quick amateur glance, those caps you've removed look to be part a standard 3-pole active low pass filter stage you see on most of these octave designs for more active tracking on the octave down...they follow after the op amp booster stage but don't affect the clean, direct signal.

Think of those caps are being part of advanced R-C filters.   To those who read that and simply remove them especially on a soldered up build, note that it may increase the aggressiveness of the octave down, decrease the perceived 'cleanness' of it (which is probably one of the best features of this box), and compromise tracking all at the same time (again, another of it's noteworthy features).  That may, but may not be desirable.   Changing the values of the resistors or caps might be a better approach than just removing them.  You can reduce them, for example, etc, to taste...ideally maintaining the ratios...or work with the pair of 330k resistors.  I don't personally know why 330ks and 2.2n/220pf were chosen over 33k/22n/2.2n instead.  I'd think it'd be the same or possibly better, but I'm no EE.  Maybe those values were in part chosen due to cost given how many were built?

4013 octave downs are everywhere....this particular design is 'complex,' compared to the slacktave, 4ms loco, blue box, etc.  But if you have the time, space, and patience for it, I'd imagine the design of the 'fundamental extractor' your are tinkering with is a good 'upgrade' to most 'simpler' 4013-based octaves, in terms of tracking if you like the cleanness. If you want more filth, you can always add your preffered form of nastiness afterward.

There is always that compromise of striking a balance between good tracking and too-muted tone, however, and preference is a large part of where you draw that line....functionality the other.

...oh, and with a mention of a fundamental extractor,  let's not forget to give an honorable mention to the EH octave multiplexer fundamental extractor, or StephenGiles will be here to remind us all soon, indeed!        

[Processaurus]

Here's some more discussion of how the circuit works, though I haven't heard or been able to figure out the "magic" part of the fundamental extractor with the comparators.

I've not built it myself, but from a quick amateur glance, those caps you've removed look to be part a standard 3-pole active low pass filter stage you see on most of these octave designs for more active tracking on the octave down...they follow after the op amp booster stage but don't affect the clean, direct signal.

Yes, it's making the 3 pole Butterworth filter (don't worry, I really have no clue) into a two pole one.  Removing it would probably affect the tracking a little bit and that's it, but if people like it better, no smoke's going to come out.  Maybe it scoots the tracking window range a little higher.  The stock one has a cutoff around 800 Hz, here's a neat link that STM showed us in the linked thread above, that shows how to calculate parts for the desired cutoff in this filter (ours is on the upper right, if you look close), which boss uses a lot, they even use it again in this pedal to LP filter the octave sound.  Changing the values there would probably get some different, good sounds, as would changing values in the tracking section being a good mod for bass or getting it to stick to lower notes better, at the expense of weakening the tracking of the higher notes.

Just a question...just saw this....why the aversion to the input buffer?

Unneeded here, the TL07* opamp's FET input is a hi impedance buffer.

4013 octave downs are everywhere....this particular design is 'complex,' compared to the slacktave, 4ms loco, blue box, etc.  But if you have the time, space, and patience for it, I'd imagine the design of the 'fundamental extractor' your are tinkering with is a good 'upgrade' to most 'simpler' 4013-based octaves, in terms of tracking if you like the cleanness. If you want more filth, you can always add your preffered form of nastiness afterward.

Someone made a layout of just the fundamental extractor here, in the thread above I think.  You could always add a switch to make the tracking worse if you wanted that bluebox chaos.  I got my blue box to track extra bad by HP filtering what went into its extractor.

and PS, if the 4013 can take it, you're probably better off running that whole thing off a bipolar supply, and ground will be a more stable VREF.    OR, add a op-amp VREF.   Or a 5v regulator for VREF of the op amp stages, rather than that cheesy voltage divider.  Should be more stable.

I'd have to respectfully disagree, looking at the schematic, nothing's putting a serious DC load on the simple resistive divider, the logic section doesn't use it at all.  It also has a Big Freakin Cap to ground to suck AC signals to ground.  Maybe add a .1 cap in parallel with the BFC just to be safe from real high frequency transients and call it a day?

...oh, and with a mention of a fundamental extractor,  let's not forget to give an honorable mention to the EH octave multiplexer fundamental extractor, or StephenGiles will be here to remind us all soon, indeed!

Overrated... [ducks]  This one and the EBS are no worse than Howard Davis's design for EH (I have all three, somehow).  The Boss could trump the EH if it had a compressor IC feeding the tracking section, I'd wager...

[liquids]

Yes, it's making the 3 pole Butterworth filter (don't worry, I really have no clue) into a two pole one.  Removing it would probably affect the tracking a little bit and that's it, but if people like it better, no smoke's going to come out.  Maybe it scoots the tracking window range a little higher.  The stock one has a cutoff around 800 Hz, here's a neat link that STM showed us in the linked thread above, that shows how to calculate parts for the desired cutoff in this filter (ours is on the upper right, if you look close), which boss uses a lot, they even use it again in this pedal to LP filter the octave sound.  Changing the values there would probably get some different, good sounds, as would changing values in the tracking section being a good mod for bass or getting it to stick to lower notes better, at the expense of weakening the tracking of the higher notes.

Well, I think it would just make it a single pole filter with a lot of resistance, feeding an an op amp buffer, but, gotcha.     And yeah, I would say, if you want to mess with the tone of the octave, C26/C27/C28 might be better (post-filtering), though you make a great point in suggesting the values of C21, C5 and C4 (or R8, R9, R5) as mentioned might be adjusted to better suit bass, for example.  Hadn't thought of that.  Wasn't there a boss bass octave too, by the way?  Maybe that schem is useful. You mention the EBS - I've not seen the schem...though from what your saying it's a similar concept - what values are used for the pre-filter?

I'd have to respectfully disagree, looking at the schematic, nothing's putting a serious DC load on the simple resistive divider, the logic section doesn't use it at all.  It also has a Big Freakin Cap to ground to suck AC signals to ground.  Maybe add a .1 cap in parallel with the BFC just to be safe from real high frequency transients and call it a day?

Well, in part I mistakenly thought there were 8 op amp stages getting a VREF.     But then I realized not long after I posted...like you said, none of the logic is using it.    Oops. However, it seems the divider 4.7k/5.6k is either putting it just above Vref on purpose, or compensating for the losses.  So it could go either way...that 5v regulator might come in handy so long as you aren't using a battery...  But yeah, it may not make that much difference given that it actually doesn't have much load and that Big Cap.  I'm learning.    

One thing that seems odd is the pots are fed from the voltage divider rather than ground.  I assume this is to bias the base of the BJT buffer, and It's probably this way to avoid any low pass filtering that referencing ground, using DC caps, and making the BJT a standard buffer would probably require.  But it also occurs to me that if you are getting distortion, make sure those 100k R59/R60s are in there for the mixing, and that Q2s base is at or above 1/2V++, no?  Could be the buffer, potentially, if it doesn't get full swing....    But why wouldn't going to a higher voltage supply (say up to +/- 9v) be an advantage?  The datasheet I just checked said the 4013 it can handle 18v total maximum, with 15v recommended operating range.  Using a bipolar supply might make some of that much easier.  At that, you might be able to mod it for more clean output, and likewise send a stronger signal through the octave portion (which may or may not be an improvement, but you'd have more clean range)...and get more output overall if so desired.

Just some ideas for the experimenter accompanied by questions for those who understand this far more than I do.  

Overrated... [ducks]  This one and the EBS are no worse than Howard Davis's design for EH (I have all three, somehow).  The Boss could trump the EH if it had a compressor IC feeding the tracking section, I'd wager...

Now you've done it!  

[oldrocker]

.

[isildur100]

Hi liquids, maybe you knew this already:

In this circuit, the divided signal (a square wave) is not outputted directly but instead is used to trigger the jfet switch (Q8) on the clean signal. That's why it outputs a clean signal compared to the blue box or the shocktave where the resulting signal is distorted. In the OC-2, the output signal never went through the frequency division circuitry, it is an "indirect divider".

An very good explanation on how it works can be found in one Stomboxology issue: http://moosapotamus.net/IDEAS/stompboxology/FreqDiv_SubharmSynth.htm

As for the buffer at the input, I decided to put one because it helped the overall sound and reduced the leaked signal I was hearing. But I agree that it is maybe not necessary for everyone, it's just that I saw a little difference and didn't mind adding it.

I have also added a mid boost pot control on the octave'd output in order to tweak in different octave sound textures. It adds to the part count, which is already high but hey, it's useful, at least for me

I will post my mods once I confirm that my schemo is accurate.

John