Chuck D. Bones
Circuit Wizard
The first thing one notices when comparing the BD-2 and BD-2w is that the BD-2w is all surface mount. Here's the BD-2w schematic. It's believed to be correct. My breadboard based on this schematic sounds like the real thing. Note that this schematic includes the buffering and soft switching, which are not on my breadboard. All of the electrolytic caps in the audio path are tantalum in the -2w. This is as much about packaging as it is about tone because SMD tantalum caps are smaller. IC2 - IC5 are analog switches controlled by the S/C switch. All of the analog switches are switched in tandem. I'll walk you thru the differences between The BD-2w and the BD-2. The ref designators are different between the two schematics, it couldn't be helped. In this discussion, the ref des refer to the BD-2w schematic.
The -2w has a fatter input cap (C8 & C40 are 1uF film) and a current-sink load on Q5's source (Q23, D13, etc.). Very "boutique," but I doubt this "upgrade" is audible.
R15 is 1.2K vs. 1.5K in the BD-2. Raises the max gain by about 2dB. Probably not audible.
The first set of analog switches change the Fender tone stack tuning. IC2-1 & IC2-2 connect a 10K resistor in parallel with the Fender bass control (R20). This effectively turns the bass from 10 down to below 1. It tightens up the bottom end between the two gain stages. C7, C29 and R43 are there to keep the analog switches happy and do not affect the tone. IC3-2 connects C41 in parallel with C16, which restores some of the bass lost by switching in R42. That's right, "C" mode both cuts and boosts the bass. Go figure. The diodes still do nothing. Nada. Zilch. Bupkess.
This is the freq response from Q9-C to Q10-G with the S/C switch in both positions. Green = S; blue = C. In the S position, the signal at the input to the 2nd stage is about 4.5dB stronger at 50Hz and about 1dB weaker at 500Hz. Not a huge difference. Basically, the C setting flattens the freq response of the Fender tone stack. I call this the TIGHT switch.
The BD-2w 2nd stage amplifier is the same as the BD-2 except C18 is 100pF on the -2 and 47pF on the -2w. Not audible.
The second tone-shaping network has two capacitors (C43 & C44) switched by IC4. In S mode, the switches are open. C19 is slightly larger in the -2w, 6.8nF vs. 5.6nF. C20 is slightly smaller in the -2w, 4.7nF vs. 5.6nF. These are subtle changes that work with the mode switching. IC4-2 connects C44 in parallel with C21, increasing the capacitance from 18nF to 40nF which fattens the upper bass and midrange. Increasing C21 was a popular mod before the -2w came out. IC4-1 connects C43 in parallel with C20, which lowers the treble a little bit. The TONE and LEVEL controls are unchanged.
This is the freq response from Q12-C to the top of the TONE pot. TONE and LEVEL are both set at noon. The difference is not huge, but it is noticeable. As before, green = S; blue = C. I call this the FAT switch.
Roland engineers replaced the TL071 opamp with another 3-transistor discrete opamp, similar to the 1st & 2nd stages. The 3rd stage opamp is tweaked for stability at low gain. Q15 gets a current-sink collector load for maximum output swing. Then Roland did something that completely baffles me. They built two gyrators with nearly identical freq response and then switch between them.
This is the 3rd stage freq response for the two switch settings.
Both switch positions make a 6dB bump. One at 125Hz and one at 135Hz. I don't get it.
Below is the overall freq response for the S and C switch positions. All controls are at noon. Not a huge difference between S & C. This is born out in listening tests. Note that this is what we get when the signal is small enough that none of the amplifier stages saturate. The effect of the S/C switch is slightly different when GAIN is cranked, but it's still pretty subtle. The best way to describe the S/C switch is it gets a little louder in the C position.
OK, so how do we get a BD-2 to sound like a BD-2w? The easy way is to say "I only want the C setting." Then we tweak the component values to match the C switch position and we're done. Changing the TL071 to a discrete opamp is icing on the cake and not absolutely necessary to achieve the tone. Leaving the TL071 alone is much easier than building the discrete opamp on Vero and then hay-wiring it into the circuit.
I like some flexibility, so I installed two DPDT toggle switches on my breadboard. One for the TIGHT switch and one for the FAT switch. I didn't bother with the gyrator switch. Because I didn't use analog switches, I didn't need C7, C29 or R43 and I only needed one switch for R42, not two. I use 1/2 of the TIGHT switch to connect R42 in parallel with R20. The other 1/2 of the TIGHT switch connects C41 in parallel with C16. This can all be done quite easily on a Cobalt build. But before you do that, you might want to read part 3.
Next time: more mods.
The -2w has a fatter input cap (C8 & C40 are 1uF film) and a current-sink load on Q5's source (Q23, D13, etc.). Very "boutique," but I doubt this "upgrade" is audible.
R15 is 1.2K vs. 1.5K in the BD-2. Raises the max gain by about 2dB. Probably not audible.
The first set of analog switches change the Fender tone stack tuning. IC2-1 & IC2-2 connect a 10K resistor in parallel with the Fender bass control (R20). This effectively turns the bass from 10 down to below 1. It tightens up the bottom end between the two gain stages. C7, C29 and R43 are there to keep the analog switches happy and do not affect the tone. IC3-2 connects C41 in parallel with C16, which restores some of the bass lost by switching in R42. That's right, "C" mode both cuts and boosts the bass. Go figure. The diodes still do nothing. Nada. Zilch. Bupkess.
This is the freq response from Q9-C to Q10-G with the S/C switch in both positions. Green = S; blue = C. In the S position, the signal at the input to the 2nd stage is about 4.5dB stronger at 50Hz and about 1dB weaker at 500Hz. Not a huge difference. Basically, the C setting flattens the freq response of the Fender tone stack. I call this the TIGHT switch.
The BD-2w 2nd stage amplifier is the same as the BD-2 except C18 is 100pF on the -2 and 47pF on the -2w. Not audible.
The second tone-shaping network has two capacitors (C43 & C44) switched by IC4. In S mode, the switches are open. C19 is slightly larger in the -2w, 6.8nF vs. 5.6nF. C20 is slightly smaller in the -2w, 4.7nF vs. 5.6nF. These are subtle changes that work with the mode switching. IC4-2 connects C44 in parallel with C21, increasing the capacitance from 18nF to 40nF which fattens the upper bass and midrange. Increasing C21 was a popular mod before the -2w came out. IC4-1 connects C43 in parallel with C20, which lowers the treble a little bit. The TONE and LEVEL controls are unchanged.
This is the freq response from Q12-C to the top of the TONE pot. TONE and LEVEL are both set at noon. The difference is not huge, but it is noticeable. As before, green = S; blue = C. I call this the FAT switch.
Roland engineers replaced the TL071 opamp with another 3-transistor discrete opamp, similar to the 1st & 2nd stages. The 3rd stage opamp is tweaked for stability at low gain. Q15 gets a current-sink collector load for maximum output swing. Then Roland did something that completely baffles me. They built two gyrators with nearly identical freq response and then switch between them.
This is the 3rd stage freq response for the two switch settings.
Both switch positions make a 6dB bump. One at 125Hz and one at 135Hz. I don't get it.
Below is the overall freq response for the S and C switch positions. All controls are at noon. Not a huge difference between S & C. This is born out in listening tests. Note that this is what we get when the signal is small enough that none of the amplifier stages saturate. The effect of the S/C switch is slightly different when GAIN is cranked, but it's still pretty subtle. The best way to describe the S/C switch is it gets a little louder in the C position.
OK, so how do we get a BD-2 to sound like a BD-2w? The easy way is to say "I only want the C setting." Then we tweak the component values to match the C switch position and we're done. Changing the TL071 to a discrete opamp is icing on the cake and not absolutely necessary to achieve the tone. Leaving the TL071 alone is much easier than building the discrete opamp on Vero and then hay-wiring it into the circuit.
I like some flexibility, so I installed two DPDT toggle switches on my breadboard. One for the TIGHT switch and one for the FAT switch. I didn't bother with the gyrator switch. Because I didn't use analog switches, I didn't need C7, C29 or R43 and I only needed one switch for R42, not two. I use 1/2 of the TIGHT switch to connect R42 in parallel with R20. The other 1/2 of the TIGHT switch connects C41 in parallel with C16. This can all be done quite easily on a Cobalt build. But before you do that, you might want to read part 3.
Next time: more mods.
