Call for comments: Blues Driver clone layout at gaussmarkov.net

[gaussmarkov]

so ... the diodes across the op amp inputs are for protection ... see R.G.'s What are all those parts for??.  also, this has been asked before: 

http://i41.photobucket.com/albums/e270/gaussmarkov/Forum%20Posts/opampinps.png

...  is this probably another example of the sort of precautions that commercial designers take that we can generally omit?

it has been explained on this board many a time, and... 

and what?

[MartyMart]

Great work guy's, I think the BD-2 is "worthy" of a DIY version, it's one of my most
used Boss products !
Reminds me a bit of Mark's "The Crank" with more crunch on tap, a friend of mine uses
2 x BD-2's that I modified for him, one set clean and one set for grind/solo tone
Marty.

[gaussmarkov]

i vacillate between thinking that the Vref supply needs an op amp and doesn't.      still unable to find a discussion of this (though i saw a similar question in a thread about the electric mistress), i decided to sim the two discrete op amp stages and look at the Vref current requirements.  this is my SwitcherCadIII schem:



the input is an AC signal with 100mV amplitude.  the current through the 4V supply has an amplitude of 600nA which hardly seems worth worrying about.  this suggests to me that the op amp is unnecessary.

OTOH, Boss doesn't put this op amp into all of its power sections.  so maybe it matters.  OTOH^2, Boss doesn't protect all of its op amps with diode clamps either.  so that kind of logic is not compelling to me.

any help much appreciated.

[MR COFFEE]

I think the buffering of the midpoint voltage reference with the op amp is probably there because the discrete op amp using the JFETs and BJT have such poor supply rejection and cmv rejection. Which only matters if somebody puts it on a pedalboard with a cheesy power supply ... that'll never happen, huh?

[puretube]

a wise man once said: "A circuit is only as good as its power supply"...

The power supply "wants" to be a low impedance source/drain, signalwise (ideally a shortcircuit AC-wise, between positive and negative pole),
and be able to supply sufficient current at any time.
(Important with LFOs, for example  - think: ticking...).

Large caps help this effort.

When working with virtual ground circuits ( bias = V_b/2 ),
this "ground" "wants" to be of low impedance too,
and also needs to supply enough current into changing load conditions,
to keep the DC voltages constant.

The voltage follower IC serves those goals better than just a cap across a 2x 10k voltage divider...

[sfx1999]

Is the BD-2 basically a discrete version of another pedal? Has anyone tried subbing opamps or the diode compression opamp for the FET OPAs?

[WGTP]

http://aronnelson.com/gallery/WGTP/OD_3_SubClone?full=1

Here is slice of the Boss OD-3 that also uses that type circuit.  I haven't tried it yet.  Also check out the Boss ROD-10 that has multiple examples of it.  There is also some info in the separate Diode Compression section.   

[gaussmarkov]

thanks for these comments!

I think the buffering of the midpoint voltage reference with the op amp is probably there because the discrete op amp using the JFETs and BJT have such poor supply rejection and cmv rejection. Which only matters if somebody puts it on a pedalboard with a cheesy power supply ... that'll never happen, huh?

i will have to think about this.  i haven't found an explanation of "cmv rejection yet."  also, this makes me think of a concern i had when i first encountered this op amp role:  what comes out of the op amp is only as good as what goes in.  if there is a ripple in the power supply, then the op amp will make a perfect copy of it and we are back to relying upon the filter capacitors.

a wise man once said: "A circuit is only as good as its power supply"...

The power supply "wants" to be a low impedance source/drain, signalwise (ideally a shortcircuit AC-wise, between positive and negative pole),
and be able to supply sufficient current at any time.
(Important with LFOs, for example  - think: ticking...).

Large caps help this effort.

When working with virtual ground circuits ( bias = V_b/2 ),
this "ground" "wants" to be of low impedance too,
and also needs to supply enough current into changing load conditions,
to keep the DC voltages constant.

The voltage follower IC serves those goals better than just a cap across a 2x 10k voltage divider...

thanks, ton, for confirming that the op amp is even better decoupling.  i am hoping to learn how to decide when such additional decoupling is needed.  i thought that if the simulation showed a lot of current variation in the constant voltage supply that would be such an indication.  but i found almost none on the virtual ground.

i can imagine that with an LFO this could really help.     i may try this! 

and subscripts!  i had not noticed those before.     e = m c²

Is the BD-2 basically a discrete version of another pedal? Has anyone tried subbing opamps or the diode compression opamp for the FET OPAs?

i don't think so.  the idea (as i have seen it explained by R.G. and others) is to get some character by using a discrete version of an op amp.  if you look at the threads mentioned by WGTP, you will see some examples of the exact opposite:  taking op amps and subbing discrete versions for them.  or am i misunderstanding your question?

http://aronnelson.com/gallery/WGTP/OD_3_SubClone?full=1

Here is slice of the Boss OD-3 that also uses that type circuit.  I haven't tried it yet.  Also check out the Boss ROD-10 that has multiple examples of it.  There is also some info in the separate Diode Compression section.   

thanks for this WGTP!  i wish i had started with your OD-3 and the Diode Compression section is a good connection.  do you happen to have a good reference on tweaking the resistors for putting in other transistor types, like replacing the JFETs with MOSFETs?  the Nelson Pass article has that example, but doesn't explain the resistor value selection process.  i am thinking that the next pass will have the compression diodes as an option and extra resistors so that the same pcb can be used for different transistor trios.

again, many thanks all!

[WGTP]

From my experience with transitor substitutions, you will probbaly have just try some different values until it sounds the best.   

[MR COFFEE]

Hi Gauss,

i haven't found an explanation of "cmv rejection" yet./quote]

Try here

http://www.ecircuitcenter.com/OpModels/CMR/Op3_CMR.htm

Since the cmv rejection depends on the matching of the input devices, and since in discrete op amp circuits the input devices are hand-matched at best (prolly not) and are therefore less well-matched than virtually any monolythic (i.e., IC chip) op amp, the JFET input discrete op amps tend to have mediocre (inferieor relative to monolythic implementations) CMV rejection. Most monlythic implementations (all?) have additional circuitry designed to improve the CMR rejection ratio (CMRR).

Doesn't mean discrete JFET-based op amps don't have other virtues (low noise, less "throwaway" gain, pleasing overdrive sonic signature characteristics, etc.). It just means the power rails and pseudo-ground better be well-filtered, especially if you run it off a cheesy wall-wart power supply, or it may have a bit more baseline hum than you would like. The op amp buffer for the divider makes the pseudo-ground less susceptible to power supply funnies.

mr coffee

[deviance2250]

Cant seem to get this one to work.  Am a little confused on the transistors now.  Q2, 3 and 5 have four holes, one square.  Is the transistor lead suppoed to solder on the square and then jumper across to the fourth lead?  Or is is suppoed to go directly in the fourth hole only, skipping the square hole??

Hope this hasnt been asked already - this is only the fourth effect I've tried building myself.

Thanks!

I looked at the layout, and I'm a bit confused.  There are 4 pads for the transistors, and one of them is supposed to be labeled as a square, but I don't see that.  Also, is one of the pots dual-ganged?  I just wanted to see if I could understand what was going on your clone.  Still learning the ropes in terms of circuit mapping, schematics, and layouts. 
Thanks.

huh!?!  you are right.  and i see what happened.  i changed only the second version of the layout.  so if you look in the pdf project file and compare the two you can see where the square should be.  but i will fix that as soon as i can so that it is consistent everywhere.

and yes, the gain pot is dual-ganged.

ok.  i made corrections.  i am hopeful that in the process i did not introduce errors.  just for more clarity, the issue is that we laid out several transistors so that one lead straddles an underlying trace.  and we put in drill holes (and associate pads) so that socketing the transistors with inline sockets would be easier:  4 holes in a row instead of 3.  so we also put a square pad for the socket that is not supposed to be used.  here's a pic showing Q1 normal and Q2 with the extra square pad:



the project file explains all of this--but there is bound to be some confusion.  here is another colour palette where the difference between the leads (grey) and the pads (green) is clearer:

[gaussmarkov]

skip the square hole.  the idea is that the one odd-shaped hole is the one to avoid. 

[oomnelson]

Hi everybody !, and thanks for this interesting project. I´m working on it, but in my country I can´t get the J201 fet, I think in replace it with a 2N5458, but it´s very expensive, so, I´m thinking in a BF245A to replace the J201. Would it be ok ?
Thanks again, I hope someone can help me with this item.

[caspercody]

That should work.

Or if you can get 2SK117. I have been reading that that one is better than a J201, but hard to get here.

[caspercody]

Got home and tried to replace existing J201, with BF254A and it works. The pin out is different than the J201, backwards. But they seem to be quiter than the J201. I think I will leave them in.

[oomnelson]

Thank you so much caspercody!!!... I´ll do it that way.

[oomnelson]

Hi, I just finished the project, I turn it on, but the only thing I heard is a big noise... after checked everything, several times, and as I see it´s all right, I changed the IC TL072 for a TL081, and now I can hear the sound of the guitar, with a continous noise, and when I put some gain from the pot, the distortion is disgusting. The tone pot add noise, and the volume pot at high volumes make a terrible noise.
Could it be the diodes???...
I changed 1N914 diodes for 1N4007 diodes, but didn´t change anything.
I thought the problem could be the BF245 (I put them in place of the J201, the pin out is different so I put them backwards), but for caspercody it works fine.
I´m thinking in start all over again, I´m really interesting in this project. If someone had the same troubles and want to help me, or someone have some idea, I´ll be very grateful.
Thanks!

[caspercody]

I quickly looked at the project, and thereare (2) versions. Depending on which one you did, one is with a single op amp and the other is with a dual op amp. There are (4) Jfets, and the BF245 should work, just keep in mind the pin out. The diodes are for clipping, and Te 1N914 Will work just fine. Now I see there are other transistors in thisproject, what transistors did you use?

You are getting sound so I wouldsay things are in their right place, but sounds like crap. You can make a audio probe and start looking for where it starts to sound like crap. Maybe have a wrong value somewhere?

[oomnelson]

Thank you again caspercody !... you´re great.
I´ll take it with patience and time.
I did the "Vref op amp" version, I used 2N3906 (Q3 and Q6 transistors), and 2N3904 (Q7 transistor).
Thank you for your time and dedication.

[caspercody]

Oomnelson, check your Jfet legs. I have a pedal I just built, that uses a 2SK117 but I can not get them. So I put a J201 in its place (backwards, which is correct), but had to bend the source leg so that it is in the gate leg position. And bend the gate leg so it is in the source position. I hope that makes sense. But now my pedal works.

Sorry if I gave you wrong info that these would swap out. They will, but with a little more work.