Boss OC-2 chopped distorted sound leakage problem

[plague.doc]

Hi all! First of all, I'm apologising for my english. I'm not a native speaker, and also my knowlege in electonics is mediocre. But I'll try my best to explain the situation.

I've been searching the forum and found out I have the very same issue explained in that topic by isildur100: https://www.diystompboxes.com/smfforum/index.php?topic=80547.0

"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 hear the very same distorted signal, especially with clean and oct1 knobs down. It seems he never found a solution to this problem, and concluded its just how the circuit works. Placing a buffer in front of a pedal didnt fix the issue for me. I want to fix this issue if possible, or at least confirm that it is just how this circuit works from other people who built this project

I also used schematic by Ben Milner, but buit it on veroboard. Layout I used is here: http://guitar-fx-layouts.42897.x6.nabble.com/Chopped-Boss-OC-2-compact-layout-td13851.html

I added several mods to the circuit. First one is a simple mod to lpf from oct1 section, I've got it from that video https://www.youtube.com/watch?v=mTCGaLGD1Ys. Second one is a variable 50k resistor instead of R7 to tweak the gain of a unit. Third one is switch which isolates the clean signal from the blend section of the circuit. I also used free holes in A row as a ground bus, to connect all of the grounds from in and out jacks, LED ground from 3pdt, etc. Apart from that, everything is exactly as diagram says, including the values. I double checked every component placement and cuts in a veroboard, so everything seems correct to me.

Either way, thanks for any and all help you throw at me!

[bushidov]

I had this same problem. Could never figure out the source. Hope you figure it out

[Marcvv]

If you search you will find that I encountered the same problem. I solved by putting a simple jfet buffer in front. For me that was also an extra solution as I wanted a greenringer in the same housing. From the Jfet buffer you can split the signal to the oc-2 and the greenringer and bring them back where the signal mixe just before the output. This is the same spot where you mix back other mod signals. (Look up mods for the OC-2)
I have this pedal still on my pedalboard.

[plague.doc]

to bushidov:
I've seen your threads about this circuit and learned some info about how it works. Thanks for the encouraging too!

to Marcvv:
Can you please post here the shematic of the exact buffer which fixed the issue for you? I tried to put a buffered boss pedal in front of it, but it didnt change much. I've found several buffer layouts on vero, I'm willing to try the smallest one so it could fit in the enclosure better. But I'm waiting for your advice about that. Thanks!

[bushidov]

The whole JFET buffer thing might explain some things. I did both the Chopped OC-2 and even reverse engineered a Behringer UO-300, which is more or less a full BOSS OC-2 clone. Like you had, I kept getting a faint distorted signal that seemingly was coming from my TL074 op-amp, which I eventually broke down into smaller TL072's. It appeared to be coming from the op-amp that mixed the dry and wet signals back together, but while swapping parts and re-organizing how it was laid out, could never get it to go away.

However, I didn't use a JFET buffer on its input. The UO-300 worked, but my clone of it didn't, and had the same problems. But, one thing I didn't carry over was the JFET switching system the UO-300 had (Boss's switching system), which in turn does have a bit of a buffer in it.

[plague.doc]

I built the simplest buffer from picture above and put it in front of the whole circuit. Maybe it helped a bit, its hard to tell. But overall the problem is still there. Maybe I'll try buffer circuit boss used in their pedals later

[duck_arse]

plague.doc - nobody has said it yet, so welcome to the forum.

I couldn't get the layout image to load for some reason, but - your Ge diode [D10?] appears to be backwards. it looks like a Russian type, which has the band at the Anode end. what's that you have in the J201 spot?

[plague.doc]

to duck_arse: Hi and thanks for the warm welcoming! You are absolutely right about the GE diode, it is a Russian D9B and it really has marking on the anode side. Pretty dumb mistake from me, considering only available datasheet for it is in Russian. Good thing I placed it on socket to experiment with a diode type. I'll try reversing it first thing when I'm home. About J201, its just a little PCB to place a SMD j201 to a regular thru-hole PCB. I'm pretty sure it is installed correctly, but I'll double check when I'm home

[bushidov]

Yeah, welcome to the forum. Forgot my manners.

[antonis]

Forgot my manners.

Sad but true.. 
(especially for a person certified for humane science studies..)

[plague.doc]

I've just flipped D10 ge diode (and also changed it to another russian ge diode, D9J, just to experiment. Sounds the same, as expected) and it changed the sound of the unit quite a bit. it seems D10 is a part of a filter applied to an octave signal. Now it sounds less loud and less distorted, more like an original pedal. But it sounds surprisingly good without a diode too (same as with diode in opposite direction, maybe I'll mod the pedal to have a switch to take off this diode out of the circuit). I've made more pictures of a zone duck_arse mentioned, here is j201 on a little pcb. Also I'm attaching the layout I used. I should have done this in a first post, sorry.

Back to the leaking signal issue, sadly it stays the same no matter type or direction of D10. I've recorded some audio examples too, just to demonstrate. First one is clean 50% and oct 100%, second one is oct 100% with the clean signal left out from the blend section, third one is both controls on 0%. I've boosted last clip to +15db so it is easier to hear

Sound clip: https://drive.google.com/file/d/1nbcsYPitfqRg31y5xAmo4-2MxCglI35U/view?usp=sharing

[Marcvv]

to Marcvv:
Can you please post here the shematic of the exact buffer which fixed the issue for you? I tried to put a buffered boss pedal in front of it, but it didnt change much. I've found several buffer layouts on vero, I'm willing to try the smallest one so it could fit in the enclosure better. But I'm waiting for your advice about that. Thanks!

Sorry for the late response.
This is the buffer I used http://www.muzique.com/lab/splitter.htm

[plague.doc]

I've tried several buffer layouts including this one, and it didnt fix the issue sadly. I'll get black lable oc-2 made in japan near the summer and do some tests. I'll post my findings, if there will be any

[plague.doc]

Finally I got my old black label japanese oc-2! And I can confirm that it has the very same issue with distorted sound leaking even when all knobs are down. So it is a flaw of an original circuit it seems, case closed. But it sounds pretty different from my clone too, which is suprising actually. Cant tell if it is about different schematic or components. Or maybe I did a mistake while making my clone, so I'll double check that too. Anyway thanks everyone for your help!

[bushidov]

Hi Plague.doc. I had the same problem with mine as well. I built a Chopped OC-2 as well. Same problems. Did it with a PCB, not a vero board. Same problems. However, I bought a Behringer UO300, and it didn't have the problem. I tore the unit apart and traced it, made a PCB based on it, though using through-hole components, and THAT version had the same problems. It appeared to be coming out of the quad-op-amp TL074. But not on the SMD Behringer version.

[Marcvv]

I had some trouble with a scratchy pot on my chopped oc-2 that I build a long time ago.
After solving the problem by replacing the pot I discovered I had that fuzzy sound at the tail of the signal again.
I have build the chopped oc2 twice. First on Taylor's board. On that one i solved that same  problem by using a buffer. The second one that I just worked on again was build on Diablochris's board. For that one I also used a buffer in front of it and thought that it had solved the problem. Apparently it did not.

I took the plunge and audiotraced the whole board and found out that a fuzzy signal bleeded in the signal just before the last mixing jfet (2n3904)
On the Diablochris board the output signal of the two pots, clean and octave, are blended with a 100k resistor that goes to the Base of the last jfet. The trace on the board from those resistors to the jfet apparently passes some parts of which signal is bleeding in the mix. I have bypassed that whole trace with a wire that I made sure it was not close to the position of that trace and now it is almost dead quiet.  No more fuzzy end on the octave or clean signal.

Be sure to cut the connection with that trace just after the junction of the two 100k resistors and also just before where the trace connects with the jfet and bridge the two points with a wire. If you do not cut it, signal might still be bleeding in.

Marc

[Processaurus]

Hi there, sorry to hear about your issues with the Chopped OC2.  I'm guessing from what has been reported back the bleed problems are similar to clock bleed problems in modulation pedals.  The problem could be electrostatic noise or noise on the power rails.  Electrostatic noise is often a sensitive, high impedance stage picking up bleed from a strong (often logic level/+5v or +9v) signal in other parts of the circuit through physical proximity.  Here the section with IC2 makes the big swinging logic level signal.  Using shielded cable on the input (output usually not necessary) can help if the input wires get close to that area.  Nuclear response, one could try a grounded tin shield over that area, like Boss has a tin shield over the input jack on the tuner pedal to keep robot noises from the processor out of the through signal.  Routing/layout makes a difference- I would suggest spacing the input and IC1 far away from IC2 and IC7.

Another way to help electrostatic noise is by lowering impedance between stages.  The input should be high, like >250K so it doesn't "tone suck" when guitar is plugged directly in.  The people that reported a buffer helping, could mean the noise was helped by lowering the impedance into the input.  One way to test if the input picking up the noise by virtue of it's high impedance is to to try a buffered pedal in front, which effectively lowers the impedance of the input to the output of the buffer.

One other place to lower impedance between stages is the mixing stage before the output.  It is significant you are getting the distorted noise with both dry and octave volume knobs down. The two volume pots for clean and octave, at 100K, don't need to be that high, an opamp output can drive lower.  10K (audio taper) on those pots should work fine.  Possibly the mixing resistors could be 10K as well.  I WOULD NOT short the mixing resistors, as suggested above, because if the volume pots are up all the way, you are shorting the opamp outputs together which isn't good electronics.

The other common noise in these circuits with logic level signals is power supply noise. IC2 is doing a comparator action and slamming the output to the power rails each audio cycle, this can cause little transients to get into the V+ rail, or especially the ground rail.  The ground, never being perfectly 0 resistance, can have some voltage drop from one point to another, causing the pulses of current from the logic part of the circuit (IC2, IC7) to corrupt the sensitive analog part of the circuit reference to ground, if it is wiggling around a little, rather than being 0 volts.

Star grounding is a good layout practice, where the IC's doing the noisy logic stuff are grounded with a separate wire/PCB trace as close to the power supply input as possible.

For noise on the V+, a couple small .1 bypass caps from V+ to ground, in parallel with the larger electrolytic caps, physically close to the IC's can help with little fast transients getting back onto the power rails.  A .1 cap across the V1/2 supply (that the volume pots are tied to) couldn't hurt either.

In conclusion, if the noise is audible with both dry and octave volume knobs down, I would try lowering the value of the pots to 10K and the mixing resistors to 10K.  Next I'd try star grounding IC2 and IC7 back to where the power supply comes into the board.  Make sure the offboard wiring isn't long loopy wires, short as can be while still being able to service the pedal is good, and twisting up the input and ground wires, and twisting the wires to the pots helps a little with electrostatic noise, as well as routing them away from the noisy IC's. Next, .1uF caps across V+ and ground, close to each IC, and a .1uF cap from V1/2 to ground.  Last, shielded cable for the input, or a tin shield over the logic section.  One other possible thing to try is a different quad opamp than the LM324, different opamps will have a likely different behavior when saturating, being abused as a comparator, as IC2 is being used.  Last possible thing would be try a different quad opamp for IC1, that isn't FET input, like TL084, that may be a little less sensitive to picking up noise from the other IC's.

Good luck!

[Processaurus]

I had some trouble with a scratchy pot on my chopped oc-2 that I build a long time ago.
After solving the problem by replacing the pot I discovered I had that fuzzy sound at the tail of the signal again.
I have build the chopped oc2 twice. First on Taylor's board. On that one i solved that same  problem by using a buffer. The second one that I just worked on again was build on Diablochris's board. For that one I also used a buffer in front of it and thought that it had solved the problem. Apparently it did not.

I took the plunge and audiotraced the whole board and found out that a fuzzy signal bleeded in the signal just before the last mixing jfet (2n3904)
On the Diablochris board the output signal of the two pots, clean and octave, are blended with a 100k resistor that goes to the Base of the last jfet. The trace on the board from those resistors to the jfet apparently passes some parts of which signal is bleeding in the mix. I have bypassed that whole trace with a wire that I made sure it was not close to the position of that trace and now it is almost dead quiet.  No more fuzzy end on the octave or clean signal.

Be sure to cut the connection with that trace just after the junction of the two 100k resistors and also just before where the trace connects with the jfet and bridge the two points with a wire. If you do not cut it, signal might still be bleeding in.

Marc

I read this closer and that's good work, the noise seems to be a layout issue (or at least worsened by layout), with the R60 and R59 connection to Q2 being sensitive to noise.  Alt layout as you did by rerouting, or lowering the impedance through 10K pots and 10K for R59, R60 could help the logic signal bleed.

For layout, getting the summing resistors R59 and R60 physically as close to Q2 as possible will shorten that trace that is prone to picking up noise, and locating Q2 away from IC2 and IC7 could help.

[Marcvv]

Hi Ben,

I had not thought of lowering the impedance by using 10k pots.
Makes me wonder now if that would have been enough to deal with bleeding noise.

Just for clarity, in case someone is working with a similar board: (I know you understood that Ben)
In my description I was maybe not clear. i did not mean to suggest to short the mixing resistors, which of course would not be a good idea. I basically replaced the trace on the board from the knode of the mixing resistors to the Jfet base by a wire that would be far enough away from the original trace to not pickup any bleeding.

[plague.doc]

Hello again and thanks for your replies! I want to put in my two cents

I tried to build chopped oc-2 again using this vero layout I posted above. I come across the same problem with high pitch fuzz tone leaking to clean signal path. Original Boss oc-2 I've bought to test had this problem too but now I can tell it was leaking a lot less, you totally can use it as is. Also fuzz volume is not changing when you turn clean volume knob up and down. Cant say this about my chopped oc-2 veroboard build, fuzzy tone can be clearly heard and it is annoying, also fuzz volume is changing when you turn the clean volume knob. First i thought problem is around mixing jfet just as Marcvv said, but soon I noticed that the sweepable LPF I added to clean signal path can filter out this fuzzy tone, so it is occurring before this LPF and surely before mixing fet. This lead me to believe that fuzz comes from the first opamp stage of the circuit. Not really sure why, maybe thats because of this veroboard too compact or how chopped schematic has tl074 opamps being used both in signal path and tracking part of the circuit. Anyway i turned first opamp stage to a non inverting buffer and then built another opamp stage on separate pcb using tl072, now it is quiet no matter if theres buffer in front or not.

Also this schematic is very sensitive for 3pdt clicking. My first build popped on switching as no other pedal did. First i tried to add pulldown resistors to in and out, ground the input while pedal is bypassed and other common solutions. It didn't help at all, and now I can tell it was just a clicky 3pdt. If switch pops by itself, octave effect makes it sound louder and bassy. Use quality switch and everything will be fine