Chuck D. Bones
Circuit Wizard
I've been curious about this pedal for a while now. Lotta parts. I had to know why Roland used so many parts. So I built one up and played it. I'd describe it as a dirty boost. It plays clean at low GAIN settings. The tone is smooth throughout the range of control settings. The distortion fades out pretty abruptly as the note decays. There is nothing amp-like about this pedal in my mind. I did not try stacking it with other pedals. Seems like it could work well for overdriving an amp or another pedal. However, the low output impedance may not play nice with certain pedals such as the FuzzFace. Max end-to-end gain is a modest 52dB. More gain than a Klon, less than a Muff.
The Cobalt schematic is shown below. The only difference between this and the Boss BD-2 is the Boss pedal has buffered bypass and soft switching. The basic building blocks in this circuit are a pair of discrete opamps and three tone shapers. Everything between C1 and C9 makes up the first opamp. This circuit features low noise, high input impedance, rail-to-rail output swing and well-controlled overload behavior. R5 & C4 roll-off the bass starting at 700Hz. Maximum gain is a little over 40dB (100x) and peaks between 2KHz and 3KHz. C5 rolls off the top end above that. I took a few liberties on my breadboard. I didn't have a dual A250K for the GAIN control and they are hard to come by. I have plenty of dual A100K pots, so I scaled R8-R11 down by 2.5x and scaled C3-C6 up by 2.5x. The circuit works the same. I could have just as easily gone up by 2x and used an A500K dual. I don't have any J201's, so I subbed some 2SK193's for Q1 & Q3. Their Vp is a little higher than a J201, so I increased R6 to 8.2K to bring their drain currents into balance. Had I used J201's, I still would have had to tweak R6 if I wanted their currents to balance. Not absolutely necessary to balance the JFETs, I was just "blueprinting" the circuit. I also used a 2N5087 for Q5 to minimize noise. Q5 makes most of the gain and is responsible for the overload characteristics. This circuit does not exhibit the classic FET distortion you'd hear in something like the SoB or Covert.
Now here's where it gets really interesting. The tone-shaping network consisting of C9, C11, C12, R14 and R16-R18 is a Fender tone stack with Treble at 0, MID at 6 and Bass at 10. It mostly compensates for the bass that is lost due to C4 & R8. Then there are the diodes. I don't know what these are supposed to do. There is so much signal lost in the tone-shaping network that the first stage is driven to saturation before there is enough signal to crack the diodes on. I listened to this with the diodes connected and disconnected. I looked at it on the scope. There is no discernible difference. I tried jumpering D2 & D4, still no difference. I even tried germanium diodes but the circuit impedance is too high and their leakage loaded the signal down. Two schottky diodes: barely a difference. When you build this, you can put the diodes in or leave them out, it's Mox Nix.
Q2, Q4 & Q6 form the second gain stage. The Rs & Cs in the feedback loop are slightly different, resulting in a max gain just under 40dB and flat freq response from 100Hz to around 6KHz. The GAIN controls on the two stages are ganged so that their gains go up and down together as the GAIN control is rotated.
After that comes another tone shaping network: C10, C13 & R15. They reduce the treble above 1KHz by 6dB.
The TONE control (C14, C15 and the TONE pot) is next. It affects frequencies above 1KHz. The freq response is flat at noon. A popular mod is to increase C14 to 33nF or more to fatten up the tone. More on that later.
Next comes the LEVEL control and then a (strange) bass boost network. C18, R20, R22, R23 and Q7 form a gyrator that synthesizes an inductance of 32H. Willikers! That inductance resonates with C17 and creates a 6dB gain bump focused at 120Hz. Not sure what Roland had in mind there. The only thing I know of that's going on at 120Hz is hum.
Did I mention that D5 & D6 are useless? Well, they are. Some people say they are there to protect IC1's input from large signals. Thing is, TL071s don't need that kind of protection. Maybe Roland was using a different opamp in the beginning. For sure we can drive IC1 into saturation if we dime the LEVEL control. Not sure why anyone would do that. Q7 should be a low-noise transistor, like MPSA18 or BC549C, because gyrators are noise multipliers. Q7's noise is multiplied by R22 / R20 = 392x or 52dB.
At the start of this discussion, I mentioned that the Cobalt does not have soft switching or buffered bypass. As a result of removing the buffering, one little difference, I hesitate to call it an error, crept into the input circuit that has a tiny effect on the tone. The BD-2's input buffer has a 1Meg input impedance. That's what the guitar sees if the BD-2 is first in the chain. The Cobalt's input impedance is 200K (R1 in parallel with R2). This is easily fixed in the Cobalt by changing R2 to 2Meg (or 2.2Meg if that's what you have). On my breadboard, I made R2 1Meg and omitted R1.
That's it for part 1. In part 2, I'll discuss what's different in the Waza Craft BD-2w and some mods.
The Cobalt schematic is shown below. The only difference between this and the Boss BD-2 is the Boss pedal has buffered bypass and soft switching. The basic building blocks in this circuit are a pair of discrete opamps and three tone shapers. Everything between C1 and C9 makes up the first opamp. This circuit features low noise, high input impedance, rail-to-rail output swing and well-controlled overload behavior. R5 & C4 roll-off the bass starting at 700Hz. Maximum gain is a little over 40dB (100x) and peaks between 2KHz and 3KHz. C5 rolls off the top end above that. I took a few liberties on my breadboard. I didn't have a dual A250K for the GAIN control and they are hard to come by. I have plenty of dual A100K pots, so I scaled R8-R11 down by 2.5x and scaled C3-C6 up by 2.5x. The circuit works the same. I could have just as easily gone up by 2x and used an A500K dual. I don't have any J201's, so I subbed some 2SK193's for Q1 & Q3. Their Vp is a little higher than a J201, so I increased R6 to 8.2K to bring their drain currents into balance. Had I used J201's, I still would have had to tweak R6 if I wanted their currents to balance. Not absolutely necessary to balance the JFETs, I was just "blueprinting" the circuit. I also used a 2N5087 for Q5 to minimize noise. Q5 makes most of the gain and is responsible for the overload characteristics. This circuit does not exhibit the classic FET distortion you'd hear in something like the SoB or Covert.
Now here's where it gets really interesting. The tone-shaping network consisting of C9, C11, C12, R14 and R16-R18 is a Fender tone stack with Treble at 0, MID at 6 and Bass at 10. It mostly compensates for the bass that is lost due to C4 & R8. Then there are the diodes. I don't know what these are supposed to do. There is so much signal lost in the tone-shaping network that the first stage is driven to saturation before there is enough signal to crack the diodes on. I listened to this with the diodes connected and disconnected. I looked at it on the scope. There is no discernible difference. I tried jumpering D2 & D4, still no difference. I even tried germanium diodes but the circuit impedance is too high and their leakage loaded the signal down. Two schottky diodes: barely a difference. When you build this, you can put the diodes in or leave them out, it's Mox Nix.
Q2, Q4 & Q6 form the second gain stage. The Rs & Cs in the feedback loop are slightly different, resulting in a max gain just under 40dB and flat freq response from 100Hz to around 6KHz. The GAIN controls on the two stages are ganged so that their gains go up and down together as the GAIN control is rotated.
After that comes another tone shaping network: C10, C13 & R15. They reduce the treble above 1KHz by 6dB.
The TONE control (C14, C15 and the TONE pot) is next. It affects frequencies above 1KHz. The freq response is flat at noon. A popular mod is to increase C14 to 33nF or more to fatten up the tone. More on that later.
Next comes the LEVEL control and then a (strange) bass boost network. C18, R20, R22, R23 and Q7 form a gyrator that synthesizes an inductance of 32H. Willikers! That inductance resonates with C17 and creates a 6dB gain bump focused at 120Hz. Not sure what Roland had in mind there. The only thing I know of that's going on at 120Hz is hum.
Did I mention that D5 & D6 are useless? Well, they are. Some people say they are there to protect IC1's input from large signals. Thing is, TL071s don't need that kind of protection. Maybe Roland was using a different opamp in the beginning. For sure we can drive IC1 into saturation if we dime the LEVEL control. Not sure why anyone would do that. Q7 should be a low-noise transistor, like MPSA18 or BC549C, because gyrators are noise multipliers. Q7's noise is multiplied by R22 / R20 = 392x or 52dB. At the start of this discussion, I mentioned that the Cobalt does not have soft switching or buffered bypass. As a result of removing the buffering, one little difference, I hesitate to call it an error, crept into the input circuit that has a tiny effect on the tone. The BD-2's input buffer has a 1Meg input impedance. That's what the guitar sees if the BD-2 is first in the chain. The Cobalt's input impedance is 200K (R1 in parallel with R2). This is easily fixed in the Cobalt by changing R2 to 2Meg (or 2.2Meg if that's what you have). On my breadboard, I made R2 1Meg and omitted R1.
That's it for part 1. In part 2, I'll discuss what's different in the Waza Craft BD-2w and some mods.
I just printed the schematic and went to grab it and there were 2 pages (it was dark)....I'm thinkin' no shit! I'd better plan REAL carefully. Got to the light and the other page was the fuzz I had printed earlier..maybe I'll just play guitar awhile....
.
