Boss BD-2 Gain JFETs and BJT topology questions

[bushidov]

Hi Guys,

I got some more circuit topology questions. I am trying to understand the relative "why did they do this when designing the pedal" questions. Maths would be cool, but I am just trying to get some general understanding, as there are some things on here I just never seen before in most other pedal schematics.

Here is my Boss BD-2 Clone schematic


1. My first question is how the whole Q1 and Q2 butted up together and again with Q4 and Q5 butted up together. What is that? Is that like a Darlington pair for JFETs? Seriously, I have no clue.

2. I also see that this "gain" stage uses a PNP 2N3906 after the dual gang Gain pot. Then it does something that looks like a couple of RC networks for tone shaping into a set of hard clipping diodes and then does another one of these "gain" stages. Am I correct on this?

3. On the output stage, I see D5 and D6 forming another "hard clipping" set of diodes, but instead of going to ground, it is going into the negative feedback of op amp U1A. Why are they not just going to ground? What's the advantage of doing it this way?

4. On the output stage, I also see C18 decoupling stuff out of the negative feedback of U1A. What are C19, R20, R22, R23 and Q7 all doing? Is that some sort of tone shaping? Gain creation?

Sorry to be a man of millions of questions. Just trying to learn how things work.

[Kipper4]

1 I think is a differentail amp. as is the other

[swamphorn]

1. My first question is how the whole Q1 and Q2 butted up together and again with Q4 and Q5 butted up together. What is that? Is that like a Darlington pair for JFETs? Seriously, I have no clue.

The Blues Driver circuit has two discrete op amps formed by (Q1, Q2, Q3) and (Q4, Q5, Q6). In the first amplifier, Q1 and Q2 take the difference of the signal. This difference is amplified by a common-emitter amplifier formed by Q3 which feeds back into the other input of the amplifier stabilizing the gain. Any garden-variety op amp has this structure: a differential amplifier (Q1, Q2) and one or more gain stages (Q3); generally you provide your own negative feedback network (VR1 and surrounding components).

2. I also see that this "gain" stage uses a PNP 2N3906 after the dual gang Gain pot. Then it does something that looks like a couple of RC networks for tone shaping into a set of hard clipping diodes and then does another one of these "gain" stages. Am I correct on this?

Correct, there are two stages with linked gain controls. The tone network before the clipping stage is similar to a Fender tone stack with fixed values; the second tone network is a bit simpler cutting either treble or bass[1].

3. On the output stage, I see D5 and D6 forming another "hard clipping" set of diodes, but instead of going to ground, it is going into the negative feedback of op amp U1A. Why are they not just going to ground? What's the advantage of doing it this way?

These diodes keep the input pins of the TL072 within a diode drop of each other. I believe this is to prevent phase reversal, where when either of the TL072's pins go near the negative rail the output dramatically flips polarity[2]. Since the positive input will swing around 4.5V (due to R19) the negative input will be constrained to 4.5V±0.7V or whereabouts, keeping the inputs in the appropriate range.

4. On the output stage, I also see C18 decoupling stuff out of the negative feedback of U1A. What are C19, R20, R22, R23 and Q7 all doing? Is that some sort of tone shaping? Gain creation?

Yes, the Blues driver has quite a bit of tone shaping. C19 and the surrounding resistors and capacitors form a gyrator, or simulated inductor (technically a gyrator is a component which transforms a capacitor into an inductor and vice versa, but we really only use the former). This functions as if there were an inductor and resistor in series grounding one side of C18. According to Brian's article this produces a bass boost at about 120 Hz[1].

[1] https://www.premierguitar.com/articles/boss-bd-2-mods-1
[2] https://electronics.stackexchange.com/questions/234575/why-does-bringing-tl072-tl074-non-inverting-input-low-cause-output-to-go-high

[bushidov]

Wow, terrific info. I got a lot to read up on. Especially that discrete opamp part. I looked up a little already on it and found a bunch. I didn't realize you could do that, but that makes sense considering TL072 is a JFET based op amp.

[PRR]

> Q1 and Q2 butted up ... ... ... PNP 2N3906

This lash-up is an Operational Amplifier. Differential input, gain stage. What it does depends what you connect around it. 

See Fig 10 (whole page) here:
https://www.passdiy.com/project/amplifiers/diy-op-amps

Why a 3-Q opamp instead of a 19 cent chip? This opamp has maximum gain in the hundreds, not thousands. It won't act "perfect" at high closed-loop gain, it won't clip as harshly. It is a bad opamp and that can be good for musical distortion.



> output stage, I see D5 and D6

If the opamp is doing what an opamp is supposed to do, the voltage across the inputs is ZERO. The diodes do nothing.

We sometimes do the diodes when we know the inputs will be grossly over-driven, more than the input devices can stand. This plan is likely to overdrive the opamp. However IIRC the specific TL072 can stand 40V reverse bias on the gates, and if they do break-down, the drive through R18 and pots is unlikely to over-current the junction and blow it. So this may be excess conservatism. (The TL072 can do a wrong thing if an input is driven too close to a rail, but I don't think these diodes fix that, or that it matters in a distortion box.)

[amz-fx]

Boss really liked this discrete op amp and used it frequently in designs of the 1980s. The best example is the Boss ROD-10 in which there are multiple copies of the discrete op amp used to create the various distortion sounds, including a pair that are powered with 5.6vdc.

It's worth looking at the ROD-10 service manual to see how they are used, as well as the complex processing (and switching) that follows the distortion stages.

Best regards, Jack

[bushidov]

Thanks again, Paul and Jack! That is some great reading material for me.

It's worth looking at the ROD-10 service manual to see how they are used, as well as the complex processing (and switching) that follows the distortion stages.

Wow! No kidding. They really did like that discrete op amp.

Not to get off topic on that idea, but looking at Paul's link on PassDIY, but can you make that same discrete op amp without JFETs and instead use MOSFETs?

[Ben N]

OD-3 was another good Boss overdrive using a discrete op-amp.

[highwater]

Not to get off topic on that idea, but looking at Paul's link on PassDIY, but can you make that same discrete op amp without JFETs and instead use MOSFETs?

Yup. You can do it with vacuum tubes, too (see fig 12 on the same page).

A couple more links for DIY opamps:
Joe Davisson's diode-compression opamp
Rod Elliot's Project 07

[duck_arse]

does the discreet opamp also suffer from phase-reversal problems, and if not, why not use another one at the output stage, and go all opampless?

[highwater]

I played around with discrete opamps of varying complexity in a simulator a couple years ago, including a basic three-BJT circuit similar to the Boss design. IIRC, it reversed phase with overdriven *input* as a non-inverting stage, but not as an inverting stage. I honestly can't remember what it did with an in-range input but enough gain to overdrive the output.

I would expect the Boss design to be similar, unless there's some magic in the JFETs (I don't *think* there should be, but I could be missing/forgetting something).

I've been meaning to repeat those experiments, as I've learned a lot in the time since about what qualities (both good and bad) *we* want in an opamp.

[Ben N]

does the discreet opamp also suffer from phase-reversal problems, and if not, why not use another one at the output stage, and go all opampless?

(a) What's a "phase reversal problem"? There's no mixing going on in the circuit, so who cares if there is phase reversal? Besides, there are two identical discreet-OA stages in series, so either way the signal is in phase going to the output, and the output 072 is non-inverting, so no impact on this score.
(b) As noted above, simple discrete OAs have low gain and aren't very good at driving anything, so maybe the OP wanted something more efficient to interface with the outside world.

[teemuk]

They are refering to phase inversion that occurs only in specific condition when common mode range gets exceeded. Basically instead of peak clipping the waveform just 'folds' due to phase reverse. It generally happens at one lobe of the waveform only.

[antonis]

simple discrete OAs have low gain and aren't very good at driving anything..

Is it so, Ben..??

'Cause as far as I can see, each one of them drives a grounded Emitter amp..
(for a 2N3906 with h_FE about 80 at 2mA working current, driven load is slightly more than 1k..)

[duck_arse]

does the discreet opamp also suffer from phase-reversal problems, and if not, why not use another one at the output stage, and go all opampless?

(a) What's a "phase reversal problem"? There's no mixing going on in the circuit, so who cares if there is phase reversal? Besides, there are two identical discreet-OA stages in series, so either way the signal is in phase going to the output, and the output 072 is non-inverting, so no impact on this score.
(b) As noted above, simple discrete OAs have low gain and aren't very good at driving anything, so maybe the OP wanted something more efficient to interface with the outside world.

........

3. On the output stage, I see D5 and D6 forming another "hard clipping" set of diodes, but instead of going to ground, it is going into the negative feedback of op amp U1A. Why are they not just going to ground? What's the advantage of doing it this way?

These diodes keep the input pins of the TL072 within a diode drop of each other. I believe this is to prevent phase reversal, where when either of the TL072's pins go near the negative rail the output dramatically flips polarity[2]. Since the positive input will swing around 4.5V (due to R19) the negative input will be constrained to 4.5V±0.7V or whereabouts, keeping the inputs in the appropriate range.

[2] https://electronics.stackexchange.com/questions/234575/why-does-bringing-tl072-tl074-non-inverting-input-low-cause-output-to-go-high

[Ben N]

simple discrete OAs have low gain and aren't very good at driving anything..

Is it so, Ben..??

'Cause as far as I can see, each one of them drives a grounded Emitter amp..
(for a 2N3906 with h_FE about 80 at 2mA working current, driven load is slightly more than 1k..)

Fair point.

[PRR]

simple discrete OAs have low gain and aren't very good at driving anything..

Is it so, Ben..??
'Cause as far as I can see, each one of them drives a grounded Emitter amp..
(for a 2N3906 with h_FE about 80 at 2mA working current, driven load is slightly more than 1k..)

So. The forward voltage gain of this scheme is hundreds. 20-transistor chip-amps have voltage gain of thousands. And while here pull-up is by transistor, pull-down is a 2.2k resistor. If not lightly loaded, >20k, it won't pull close to the rail. Ah, in this case the worst load is ~~30k. But a 19-cent chip will confidently pull <1k in a 9V box.

[Steben]

I´m catching up on old threads....
Is there an analysis online to look for behaviour of this stage?
Symmetry, non linearity, ....