Chuck D. Bones
Circuit Wizard
The Cobalt Blues Driver - Boneyard Edition
I've shown you all of the pieces of the Boneyard Blues Driver in Part 3. Here in Part 4 I put it all together and make some final updates. All ref des in Part 4 refer to the Cobalt Build Docs. If you can't hit a part value dead-on, anything within ±20% is acceptable.
First Stage
The first stage is everything from the board's IN pad up to and including diodes D1-D4. What I wanted to accomplish here was to raise the input impedance up to match the factory pedal, use JFETs I had in my stash, balance the JFET currents, sub an A100K dual pot for the A250K dual and add a TIGHT switch that performs the same function as the first part of the BD-2w S/C switch. If you have an A250K dual pot, then you don't need to sub in an A100K dual pot and you will leave R8, R10, C4 & C5 at the original values.
The changes to the first stage are listed below.
+SIT means Select-In-Test
*Diodes D1-D4 lightly clip the very lowest freq signals when GAIN is above noon. Then the second stage strongly clips the signal. I can't hear a difference when the diodes are in or out. Install them or not at your discretion.
The TIGHT switch is part of the first stage mod. It mimics the first part of the -2w S/C switch. When the TIGHT switch is on, R18A is connected in parallel with R18. I used 3.3K for R18A; Roland used 10K in the -2w. Either will work. 3.3K cuts the bass about 2dB more than 10K. The plot below shows the freq response at the input to the second stage. GAIN is at noon. Red trace: TIGHT sw off. Blue trace: TIGHT sw on, R18A = 10K. Green trace: TIGHT sw on, R18A = 3.3K.
Balancing
The balancing procedure is not complicated. We balance the JFETs in this circuit by adjusting R6 until the voltage on Q5's collector is equal to Vref. It helps to start with JFETs that have similar Vp. IMO this should be done no matter which JFETs you use because we're not just matching the JFETs to each other, we're matching the circuit to the JFETs. You need a good DMM that can resolve 1mV. Temporarily install the stock 4.7K value for R6. You don't need to solder R6, just wedge the leads in the holes with toothpicks. Set GAIN to minimum. Connect the DMM + lead to Q5-C. Connect the DMM - lead to Vref. Vref can be found on the end of R101 closest to C2. Power up the pedal. We want the difference between Q5-C and Vref to be less than ±10mV. If the reading is too high on the positive side, then we increase R6. If the reading is too far negative, then decrease R6. Don't touch any parts or traces when reading the DMM because skin resistance can upset the measurement. On my breadboard, I used 2SK193 JFETs and ended up with R6 = 8.2K. The difference between Q5-C and Vref was -7.5mV.
Second Stage
The second stage is everything from C2 to Q6. The main things we're doing here is changing the GAIN pot to A100K (if req'd), adjusting the associated parts to keep the same gain range and tone and balancing the JFETs.
The balancing procedure is the same as with the first stage, except this time we're measure from Q6-C to Vref and adjusting R7. I used 2SK193s for Q2 & Q4, R7 ended up at 8.2K. Q6-C was within 0.5mV of Vref. I got lucky; you don't need to hit it that close.
Tone Control
The Tone control consists of R15, C10, C13-C15 and the TONE pot. This is where we add the FAT switch. The FAT switch does the exact same thing as the second part of the BD-2w S/C switch. The FAT switch is DPDT ON-ON. We need a two pole switch because we're switching two capacitors. When the FAT switch is on, C13A is connected in parallel with C13 and C14A is connected in parallel with C14.
Third Stage
The third stage is everything from pin 2 of the TONE pot to the board's OUT pad. This is where I deviated from the BD-2 and BD-2w. These changes retune the bass boost and are optional. The easy way to do this is to keep the opamp and change a few parts.
The Boneyard bass mod extends the bass boost down to 60Hz. The ultra-bass mod extends the bass boost down to 40Hz. In clean mode, the overall freq response is pretty flat below 500Hz. The bass boost happens after the second stage. If you drive the gain stages to clipping, the second stage limits all frequencies more-or-less equally. As a result, the bass boost that happens in the third stage makes the frequencies below 200Hz louder than everything else. I'm running the ultra-bass mod and it is not too boomy in my rig.
The hard way to do the stage three mod is to toss IC1 and Q7 overboard and kludge the MOSFET circuit from Part 3 onto the Cobalt board. I think I'm the only one crazy enough to do that, so I won't bother detailing how it's done here. If someone intends to route a board, I'll provide the necessary input.
To recap:
The Boneyard Mod...
I've shown you all of the pieces of the Boneyard Blues Driver in Part 3. Here in Part 4 I put it all together and make some final updates. All ref des in Part 4 refer to the Cobalt Build Docs. If you can't hit a part value dead-on, anything within ±20% is acceptable.
First Stage
The first stage is everything from the board's IN pad up to and including diodes D1-D4. What I wanted to accomplish here was to raise the input impedance up to match the factory pedal, use JFETs I had in my stash, balance the JFET currents, sub an A100K dual pot for the A250K dual and add a TIGHT switch that performs the same function as the first part of the BD-2w S/C switch. If you have an A250K dual pot, then you don't need to sub in an A100K dual pot and you will leave R8, R10, C4 & C5 at the original values.
The changes to the first stage are listed below.
Ref | Cobalt | Boneyard | Comments |
| C1 | 100nF | 22nF | Optional. At 22nF, C1 does not limit the bass. |
| R2 | 220K | 2.2M | Increases input impedance. |
| R6 | 4.7K | SIT+ | Balances Q1 & Q3. |
| R8 | 1.5K | 620 | Req'd if changing GAIN pot to 100K. |
| R10 | 22K | 9.1K | Req'd if changing GAIN pot to 100K. |
| C4 | 150nF | 390nF | Req'd if changing GAIN pot to 100K. |
| C5 | 47pF | 120pF | Req'd if changing GAIN pot to 100K. |
| C9 | 220pF | 390pF | Adds a little treble, optional. |
| R17 | 15K | 10K | Enhances the TIGHT switch, optional. |
| R18A | Not there | 3.3K | Switched in parallel with R18 by TIGHT sw. |
| S1 | Not there | SPDT | TIGHT switch. ON = TIGHT. |
| D1-D4 | 1N914 | delete | See note* below. |
| Q5 | 2N3906 | 2N5087 | Or BC560C. Lower noise, optional. |
+SIT means Select-In-Test
*Diodes D1-D4 lightly clip the very lowest freq signals when GAIN is above noon. Then the second stage strongly clips the signal. I can't hear a difference when the diodes are in or out. Install them or not at your discretion.
The TIGHT switch is part of the first stage mod. It mimics the first part of the -2w S/C switch. When the TIGHT switch is on, R18A is connected in parallel with R18. I used 3.3K for R18A; Roland used 10K in the -2w. Either will work. 3.3K cuts the bass about 2dB more than 10K. The plot below shows the freq response at the input to the second stage. GAIN is at noon. Red trace: TIGHT sw off. Blue trace: TIGHT sw on, R18A = 10K. Green trace: TIGHT sw on, R18A = 3.3K.
Balancing
The balancing procedure is not complicated. We balance the JFETs in this circuit by adjusting R6 until the voltage on Q5's collector is equal to Vref. It helps to start with JFETs that have similar Vp. IMO this should be done no matter which JFETs you use because we're not just matching the JFETs to each other, we're matching the circuit to the JFETs. You need a good DMM that can resolve 1mV. Temporarily install the stock 4.7K value for R6. You don't need to solder R6, just wedge the leads in the holes with toothpicks. Set GAIN to minimum. Connect the DMM + lead to Q5-C. Connect the DMM - lead to Vref. Vref can be found on the end of R101 closest to C2. Power up the pedal. We want the difference between Q5-C and Vref to be less than ±10mV. If the reading is too high on the positive side, then we increase R6. If the reading is too far negative, then decrease R6. Don't touch any parts or traces when reading the DMM because skin resistance can upset the measurement. On my breadboard, I used 2SK193 JFETs and ended up with R6 = 8.2K. The difference between Q5-C and Vref was -7.5mV.
Second Stage
The second stage is everything from C2 to Q6. The main things we're doing here is changing the GAIN pot to A100K (if req'd), adjusting the associated parts to keep the same gain range and tone and balancing the JFETs.
Ref | Cobalt | Boneyard | Comments |
| C2 | 2.2nF | 4.7nF | Part of the TIGHT switch mod. |
| R7 | 4.7K | SIT | Balances Q2 & Q4. |
| R9 | 2.2K | 910 | Req'd if changing GAIN pot to 100K. |
| R11 | 33K | 15K | Req'd if changing GAIN pot to 100K. |
| C3 | 1uF | 2.2uF tant. | Req'd if changing GAIN pot to 100K. |
| C6 | 100pF | 220pF | Req'd if changing GAIN pot to 100K. |
| C8 | 100pF | 47pF | Boneyard is same as BD-2w. |
| C9 | 220pF | 390pF | Adds a little treble, optional. |
| Q6 | 2N3906 | 2N5087 | Or BC560C. Lower noise, optional. |
The balancing procedure is the same as with the first stage, except this time we're measure from Q6-C to Vref and adjusting R7. I used 2SK193s for Q2 & Q4, R7 ended up at 8.2K. Q6-C was within 0.5mV of Vref. I got lucky; you don't need to hit it that close.
Tone Control
The Tone control consists of R15, C10, C13-C15 and the TONE pot. This is where we add the FAT switch. The FAT switch does the exact same thing as the second part of the BD-2w S/C switch. The FAT switch is DPDT ON-ON. We need a two pole switch because we're switching two capacitors. When the FAT switch is on, C13A is connected in parallel with C13 and C14A is connected in parallel with C14.
Ref | Cobalt | Boneyard | Comments |
| C10 | 5.6nF | 6.8nF | Same as -2w. |
| C13A | Not there | 4.7nF | Switched in parallel with C13 by the FAT sw. |
| C14A | Not there | 22nF | Switched in parallel with C14 by the FAT sw. |
Third Stage
The third stage is everything from pin 2 of the TONE pot to the board's OUT pad. This is where I deviated from the BD-2 and BD-2w. These changes retune the bass boost and are optional. The easy way to do this is to keep the opamp and change a few parts.
Ref | Cobalt | Boneyard | Comments |
| C17 | 56nF | 390nF | For ultra-bass use 1uF. |
| C18 | 56nF | 6.8nF | For ultra-bass use 18nF. |
| R20 | 1.2K | 5.6K | For ultra-bass use 2.2K. |
| D5-D6 | 1N914 | delete | They don't do anything, optional. |
| Q7 | 2N3094 | 2N5089 | Or BC549C, MPSA18, etc. Lower noise, optional. |
| C20 | 10uF | 10uF tant. | Optional. |
| R24 | 10K | 100K | Optional. C20 should be tantalum if you do this. |
The Boneyard bass mod extends the bass boost down to 60Hz. The ultra-bass mod extends the bass boost down to 40Hz. In clean mode, the overall freq response is pretty flat below 500Hz. The bass boost happens after the second stage. If you drive the gain stages to clipping, the second stage limits all frequencies more-or-less equally. As a result, the bass boost that happens in the third stage makes the frequencies below 200Hz louder than everything else. I'm running the ultra-bass mod and it is not too boomy in my rig.
The hard way to do the stage three mod is to toss IC1 and Q7 overboard and kludge the MOSFET circuit from Part 3 onto the Cobalt board. I think I'm the only one crazy enough to do that, so I won't bother detailing how it's done here. If someone intends to route a board, I'll provide the necessary input.
To recap:
The Boneyard Mod...
- Adds two switches that mimic what the BD-2w S/C switch does, but now you can switch TIGHT and FAT on and off independently.
- Balances the JFETs and allows the use of alternate JFETs.
- Subs the more common A100K dual pot.
- Deletes some or all of the diodes that do nothing.
- Retunes the bass boost in the third stage to smooth out the frequency response and increase the bass over a wider frequency range.
DoJzeh!
