Anyone ever built an MXR Blue Box?

[vfr800fiman]

Hi all-
Anyone have any links to a schematic for the MXR Blue Box?
Anyone ever built one? I'm not sure what's in the circuit, so I don't know how difficult it would be to clone.
Thanks!
Glenn

[LP Hovercraft]

I just finished building one a couple of days ago.  The project is posted on tonepad.com  I'll let you be the judge of how difficult it would be, but it is listed as "intermediate."  It worked for me the first time I fired it up.  A nice little build for such a freaky sound.  A rude little box, that's for sure!

[RickL]

I've built one too, using the tonepad layout. As far as I can tell it sounds and tracks the same as my original version and Dunlop reissue. If you build it a simple, usable mod is to use a spdt switch to choose between one or two octaves down. For one octave down simple tap the signal after the first flip flop instead of the second.

I suppose you could also add a mixer section and mix the two signals too.

[Jose]

Fransesco's (Tonepad) layout is great. Joe

[gmr1]

It's a pretty easy build, esp. if your using an etched board. You'll finish it, plug it in, it'll sound like a broken atari mixed with a hairdryer, and your either say: "whoa! this is awesome" or you'll be like me and say "Someone remind me, why the hell did I want to build this?"

[jmusser]

I believe this is the effect that someone told me the other day, that Page used on "Dancing in the Rain" . If so, I'll have to build it. I really liked the tone of it on that song anyway. Personally, I didn't feel that that fuzz was anything too extreme, just had a lot of guts. It's definitely more intense than a Fuzz Face or Big Muff, but not in the same league with the Tychobrahe or Tone Bender. Of course, that's only on that song. That may not bring out it's true nature.

[Danny G]

Mine never worked right.  Not to say it didn't sound cool.  Totally unusable except for making weird/cool noises.

Didn't realize it was an octave/DISTORTION pedal till I fired it up, heh.

[vfr800fiman]

Thanks for all the responses, I'm going to try to build one.
I just couldn't find a layout anywhere.
I used to own one of these back in '78, should never have sold it.
Thanks again everyone.

Glenn

[vfr800fiman]

Fransesco's (Tonepad) layout is great. Joe

Nice picture!
Maybe someone could answer this question for me.
Can you use a TL082A in place of the 4558?
It's just a dual op amp right?
Glenn

[Mark Hammer]

It is hard to imagine ANY analog octave box that will work flawlessly with every playing style and instrument without any modification.  Tracking optimization is the heart and soul of all such boxes.

I should be working, but here is a brief rundown of things you can change to alter the way the unit behaves.  I am using Francisco Pena's redraw and part numbering as the reference:

1) IC1A is the input and initial gain stage.  It provides enough gain to feed/drive the envelope follower formed by C7/8, D2/3, and R15.  The gain of IC1A is set for about 471, with a high end rolloff of just under 1130hz, and low end rolloff of about 16hz.  Making C7 or C8 smaller will raise either rolloff point, and making them bigger in value will lower it.  C4 also rolls off high end, but I couldn't tell you where.

2) The time constant of the envelope follower (i.e., how quickly it reacts) is set by R15/C8.  Increasing R15 will lengthen the time the gate is on, as will increasing C8.  Increasing C8 will slightly increase attack/onset time and lengthen decay time.  Sticking a small value (<1k) resistor between D2 and R15/C8 will also increase onset time, yielding a softer less sputtery attack.  (NOTE: This will make the follower a bit less sensitive so gain from IC1A may need to be increased to offset that)  The envelope follower serves to "gate" Q2 and Q3 such that sounds/notes which are not really sufficiently consistent to trigger octaves from IC2 tend to be squelched.  In such instances, the envelope signal would be too small to turn Q2/3 on and pass signal to the output.  Though R13/14 are equal values, making both fuzz and octave turn on simultaneously, I'm wondering if tinkering with the value of R13 might not be useful in permitting *some* fuzz to come through even when octaves don't (or vice versa).  C7 limits the low end sensitivity of the envelope follower.  Make it smaller and bass notes won't turn the gate on quite so easily.

3) IC1B forms another gain stage of around 101 (due to R6/R23 ratio) though the R7/R9 network attenuates that output considerably, presumably because Q1 needs them to be there.  Note that a gain of 471 times a gain of 101 is, well, a LOT of gain to be squeezed out of a puny 4558 and a 9v battery.  Naturally, the signal is very heavily clipped.  A cap across R6 would do wonders to help smooth out the resulting distortion.  For instance, a 47pf cap would roll off high end starting around just under 3.4khz, eliminating the buzziness.

4) The collector output of Q1 goes to flip-flop IC2B, which divides the signal it sees by 2, producing an octave down.  The output of IC2B goes to the other half, IC2A, and gets divided down by 2 again, yielding divide by 4 overall, and a note 2 octaves down.  As has been mentioned by others, there is no reaon why a person could not insert a switch to direct R11 to *either* pin 13 or pin 1 of IC2, to select between one or two octaves down.  I would imagine the biasing of Q1 is important here since the flip-flop does not divide everything it receives, just the content above a certain threshold.  I don't know how to do it, but it stands to reason that one could conceivably bias Q1 in a manner that results in triggering IC2 *without* requiring a fuzzy input signal from IC1B.

5) Q2/3 provide both buffering and gating of the fuzz and octave signals.  Their collector outputs feed the panning/blend control formed by R20 and R16/17.  Think of R16 and one leg of the pot as forming an attenuator, and R17 and the other leg of the pot as forming another.  As you attenuate one signal more, you attenuate the other one less.  The output of this attenuator/blend network is mixed passively via R18/19 and fed to the output pot.

6) There is NO tone control on this unit, other than what the blend pot achieves.  The flip-flop puts out a square wave which is not rounded off by much of anything except C9, which also rounds off the fuzz signal as well (global roundoff).  One of the things that people often complain about (and which they like about the PAiA Rocktave) is that you can't independantly vary the tone of the octave down.  Often, muter sounding octaves are a more pleasing and palatable addition to one's sound.  How could one achieve this in the BB?  Simple.  Take C9 out and move it so that it goes to ground from the junction of R17/R19.  If you want to get deluxe, get yourself a 3-way toggle (SPDT, centre-off), and have it select between .0068 (buzzy) and an additional .01 or .022uf cap in parallel for considerably "warmer" sounding octaves.  If you feel like it, you can do the same for the fuzz channel too atthe junction of R16/R18.  Being able to separately vary the tone of each channel AND their level balance will provide a great deal more flexibility in the unit, and likely a lot more pleasure and reduced frustration as a player.

All of this being said, it IS a monophonic unit and can only produce octaves of a note at a time.  Chords will confuse it.

[RickL]

Another wonderful post Mark. I don't believe I've boxed my Blue Box clone yet so I'll try the mods and report back the first chance I get.

[Jose]

I think I used a tl072, might even be a 82. I think I also used a 82k in place of a 220k somewhere to tame tracking. If i remember right someone suggested a pot wired variable in a spot  to do this and reffered the front end of John Hollis's Crash Sync? Worked good. But this isn't an Eventide h-4000 either! It does cool motorbike noises and octave jumps. Jose

[vfr800fiman]

Hi all-
OK, so I finish the clone using the layout from Tonepad.com, and it doesn't work first thing.    Now for the troubleshooting.

Using the component designators from the Tonepad layout, and using my trusty audio probe  :lol: , I'm not getting any signal at pin #7 of the 4558, I'm using a TL072BCP, and not the 4558. Is this what the problem is?
I do get signal from pins #1, 2, 3, 5 & 6, but not 7. This is the pin that supplies the signal to the next stage so that must be the problem.

Any suggestions?
Thanks for any suggestions anyone may give me.
Glenn

[vfr800fiman]

Anyone? Buhler? Buhler?

If I don't get any output from pin #7 of the TL072, does this mean the device is bad? I only get hiss from this pin using the audio probe.
Glenn

[vfr800fiman]

Here's some more information on transistor voltages:
Q1 E-0 B-0.21 C-9.03
Q2 E-0 B-0.55 C-0.04
Q3 E-0 B-0.55 C-0.03

Although with no signal from the output of the opamp, I'm not sure these are a problem.

Glenn

[vfr800fiman]

Take C9 out and move it so that it goes to ground from the junction of R17/R19.  If you want to get deluxe, get yourself a 3-way toggle (SPDT, centre-off), and have it select between .0068 (buzzy) and an additional .01 or .022uf cap in parallel for considerably "warmer" sounding octaves.  If you feel like it, you can do the same for the fuzz channel too atthe junction of R16/R18.

Thank you for this post Mark. I want to try this modification now that I have the Blue Box running (VERY small solder bridge was the problem).

Anyway, I had a question about this switching arrangement.
Does the switching at the junction of R17/R19 include the C9 cap, or in place of the C9 cap. I other words, the switch just selects between .0068 & .022uF right? No 0.01uF (C9) cap?

Or, do you leave the C9 cap in place and this is the center off position?

[vfr800fiman]

Is this correct?

[Mark Hammer]

Not "correct" and not "incorrect", but eminently workable.

The reason I'm noncommital is that there are a great many tone-varying options available at that junction, and you've opted for a few of them.

The essentially aspect linking them is that a cap to ground at the junction of either R17/R19 or R16/R18 will bleed treble off the signal coming from the corresponding source.  How MANY caps are involved, and their specific value, is a matter of taste and preference.

There is NO requirement that the caps be of identical value on each side.  For instance, it might make LOTS of sense to have a big-ass cap to ground on the octave-divided side because that's what it takes to mimic a "round" bass tone, but very little sense to have a similar value cap on the fuzz side because ....well.... who uses a fuzz for mellow tones?

My advice would be this.  You have it built and functionng without the original C9.  Take a bunch of caps between .001 and, say, .033.  Shift the blend over to fuzz only.  Now listen to the unaltered sound.  Do you like it?  Is it too buzzy?  If you do find it too buzzy then some sort of default cap to ground from R16/R18 may be in order.  Try out a .001 and just straddle the relevant contacts on the copper side.  If it's still too buzzy, up the value a bit.  Once you have a tone where you think "I never want to have it any buzzier than that...ever", that cap value becomes your default for the fuzz side.

The question to be determined at this point is whether you want any additional muting of the fuzz tone.  Keep in mind that the original location of C9 rounded off BOTH the bass and fuzz signals, so it was a compromise struck to balance rudeness and tameness.  Now that you've identified the maximum rudeness for the fuzz channel, how much more tameness do you want?  It may be that you only want one more level of tameness, in which case an SPST with one other parallel cap is in order for that side.  Alternatively, you may want 3 levels, so you can use a SPDT centre-off to select one of two additional parallel caps in the side positions.

Once you have some comfort with your cap choices on the fuzz side, shift the blend over the other way and do the same thing with the octave signal that you did with the fuzz, trying out different caps.  Here you will probably want 3 choices rather than 2, if only because you may want to include the divide-by-2/divide-by-4 option, and shifting up or down an octave alters the rolloff location of "round" sound.

Note that any drastic reduction of treble content will also reduce the apparent volume a bit.  So, if you seriously chop the octave high-end, you may find yourself needing to shift the blend over a little more towards the octave side. Not a problem, really, just something to plan around.

Of late, I have opted for series, rather than parallel switching of caps.  The reason is that caps left with one lead just hanging there tend to produce audible pops when that free end is engaged.  The likelihood of audible pop is a function of how much charge can be stored in the cap; you'd expect a 1uf cap to pop more than a 100pf one.

Series cap values behave like parallel resistance ones.  That is, the more caps you have in series, the smaller the overall series capacitance is.  If you use a switch to shunt/bypass one or more of a bunch of caps in series, you increase the effective capacitance value, but the beauty is that you've never allowed any cap to have a free end and not drain its stored charge.  I was "hipped" to this by John Hollis' "mode" switch on the Zombie Chorus.  It also lends itself to use of a 3-position toggle and shunting of combinations of caps.

The problem is that the math gets clumsy sometimes, and I find that it's eassier to zero in on needed capacitance values with parallel caps than series, sometimes.  Not always, but sometimes.  If you plan on using the box for gigging, "pop-lessness" trumps cap precision, but if you don't plan on gigging with it, the pops will not be huge and you may be able to arrive at the perfect tonal adjustment more readily.

[vfr800fiman]

Thanks again for the enlightenment Mark.
I never thought about the parallel switching being a problem, but it makes sense to me now. I will definitely play around with some values to see what I like for tones.

I'm off to Rat Shack now to get some SPDT switches!
Have a good holiday!
Glenn

[lightningfingers]

GLad you got it working