Curing the HIGH END BUZZ of an OCTAVIA

[analog kid]

I have now built two Octave effects , starting with the Tycobrahe that is soooo popular in this community and next the RM version Octavia. I immediately scrapped the Tyco because although it DID produce the Octave (stronger in different neck positions) I just could NOT stomach the fizzy sounding buzz that was present in the High frequency end of the Fuzz that the effect produced. SO
after listening to sound samps of the RM pedal. I went on to build it, After two days building and some voltage and misbiasing problems i got it fired up and immediately noticed a fuller sounding fuzz than the TYCO, the octave is much less noticeable and the pickups/volume have to be played with to hear it, BUT MOST NOTICEABLY again there it was!!! That SH#@Y sounding high end "Fizzle" that the tyco had    :evil:  .
THOUGH the effect(s) Does have entirely too much High end to it , that just doesnt seem to be what attributes to this characteristic that I'm hearing. (maybe some of you here know what I'm speaking of and know of ways to improve upon it?) Very annoying as notes decay and when picking with a softer attack or cleaning up the volume.  
I've heard about matching the DIODES closely in Octave as well as good Tranny selection , although I don't know how to "select" Darlingtons for gain,etc..  SO I don't know if this type of "tuning" is what needs to be done to quell this bad characteristic OR if it's just purely inherent in the OCTAVIA . but I'd definitely Hope Not!!!! And I sure hate to just keep buying parts and building more Octave/Fuzz boxes til I find one that sound pretty good. ie; the FOXX Tone Machine (tempting though) and the SOU or sickbox, just to keep getting the same results when I could just TUNE the ones I have built to get the best from
Anyone with experience with this chime in please
The Kid

[jmusser]

I built the Tycho, and didn't have the type of fizz with it that you're talking about. I mentioned that I was slightly disappointed because I couldn't get the "Tobias" tone out of it, but I'd say my solid state amp is the problem there. Now, I built the Whisker Biscuit, and it has the addition of the 47pf caps. This was supposed to be for a "more velvet tone". I didn't experience that at all, (I'm going somewhere with this btw), but I did get rid of the upper end fizz that you're talking about, by adding them. The upper end of that effect, transistions into the Bazz Fuzz from the BMP, and is really harsh without those caps in there. I have no idea if you'll need both, or just the output end one, but check out the schematic, because that addition just may work for your Tycho. It's worth a shot anyway to just tack them on to see what happens.

[analog kid]

thanks I may try that soon, though I've went on from the tyco because of the problem. It seemed to be worse than the Mayer Octavia I just finished.  almost an Oscillation as much as high end fizziness(or are they the same thing?) anyway now the RM has the same problem AND doesn't sound biased right !  It sounds great at full on GAIN but cut the volume knob back OR pick really lightly and it's very weak and misbiased sounding. I just went over the whole shcematic and everything looks ACES. So this worries me that it's just nature of the way these things sound. Which I really hope not and I don't think people would build and play them if they sounded like that.
Ideas on either of these Octavias as far as the Trannies needing matched  or Diodes (MPSA13s for the RMayer and a 3906 PNP for Q1...weird!)  and  MPSA18 and two 2N6519's for Q2and Q3 . Anything here have to be spot on like a FF does??
and the diodes in the RM are NOT germanium , i used 4148's , I assume just any pair of them should work good. I did match the Ge's on the TYCO identical voltage drops and didn't notice much difference.
thanks

[brett]

Hi.
Similar experience to mine.
I like the octavia lots more than the tyco.  The tyco makes my ears wwant to bleed.
Why?  RM is a smart dude.  The octavia forward-biases the diodes and filters the signal post-rectification.  IMO this is the way to do it.  

The others (Neo, Bob, Rambler, ...) all really benefit from a pre-boost (to 1V or more), because they are subject to heaps of rectification loss/noise generation due to the diode forward voltage requirements.

cheers

[analog kid]

Great Brett!
That's exactly what I thought the difference was(though I wasn't sure exactly) and just how the Tyco makes me feel as well. the whole reason I wanted to build this one! So since you might know a bit about this circuit maybe you could take a look at my voltages. As I am pretty sure that there's a bit of misbiasing in there somewhere. (It sounds great except when the Fuzz isn't up way high AND OR the volume is cleaned up on the guitar or I pick lighty. Then it gets weak and {"misbiased" sounding
Anyway:
FIRST ...Does it really use a 3906PNP tranny in combo with those MPSA13 in a Neg ground cicuit??
Voltages
BATTERY 8.8v
Q1   C 2.9
      B 2.9
      E 4.19    (I know this doesn't look right and I've got the 3906 in there. and NO emmitter and Base are not "shorted" together
Q2   C 5.5
      B 2.9
      E 2.0

Q3  C 6v
     B 3.0v
     E 2.5

Q4  C 8.7
     B 2.0
     E 3.3

built on etched copper pcb from the Phillip Bryant Schematic and Layout
ONLY deviations are 39k resistors on Q2's Collector to +9v and Emitter to ground are both 33k instead.
10k reverse log pot is 50k audio with tapering resistors to 9.5k reverse
thanks

[Doug_H]

FWIW, Roger Mayer essentially designed the tyco too. It was designed for a higher voltage than 9v though, which he claims is one of the things wrong with the tyco's adaptation of his circuit.

Don't know about that, but IME any of these octave up fuzzes generate a certain amt of "gritchiness" in the high freqs, due to the rectification. One thing I found that helps that is to drive a fuzz box or tube amp with the octave box. That tends to smooth it out pretty nicely.

Doug

[Fret Wire]

built on etched copper pcb from the Phillip Bryant Schematic and Layout
ONLY deviations are 39k resistors on Q2's Collector to +9v and Emitter to ground are both 33k instead.
10k reverse log pot is 50k audio with tapering resistors to 9.5k reverse
thanks

I can't say for sure, Phillip doesn't have the schematic up right now, but something doesn't seem right. Your 33k sub for the stock 39k's are definately going to affect the voltages. That's 15.4 % off the specified value. Without the schematic it's hard to tell, but some ckts do not tolerate parts value substitutions.

Tapering a log pot to a rev. log pot doesn't work to well. Also, depending on the ckt, and how the pot is being used, tapering a lin. pot to log or rev. log doesn't work well either. This could be part of the problem.

Brett's right, RM is one smart dude. :wink:  Unlike tubescreamer's, Dist.+'s, Big Muff's, etc., where it's easy to improve the original sound, alot of people seem happy enough when their Octavia or Axis Fuzz just sounds as good as the factory model.

[brett]

MPSA13

Hi.
Change that to an MPSA18/2N3904/PN100/2N5089, and see what difference it makes.  Should fix things considerably.
MPSA13 is a darlington pair, and either the high hFE (10,000) or double base-emitter bias will have had some consequences for sure.
cheers

[analog kid]

OK , just to see, I did replace those two 33k res. to 39k as called for and all it did was what I figured.. it just dropped the related voltages(on Q1 and Q2) a very small fraction of a volt each.  Here's a link to the schematic that should be up:
http://bryant02.home.att.net/schematics/mayeroctaviaschem.gif
Brett....Quote:
" Change that to an MPSA18/2N3904/PN100/2N5089..."
REALLY? I thought those Q2-Q4 had to be the Darlingtons for some reason. and as I said on the schematic Q1 is a PNP 2N3906 and I don't get it!!!  anyway  I do have all but one of those , PN100 , what could I use besides that?? and would the MPSA18 be Q1 in this example?
Knew I should've socketed those darlingtons anyway!! :twisted:

[brett]

Hi.
Almost any garden-variety NPN tranny will work in place of those wrong MPSA13s.  But quiet ones are going to be best (except for the last one, which is just an output buffer with gain =1).
 
The most important position for replacing the MPSA13 with a quality transistor (ie any of those I listed above or a BC549C) is Q2, where low noise is needed because the gain is large.

2N2222A, BC107, BC108, BC109 and BC549 are all ok small signall NPNs.

cheers

[brett]

PS As long as the emitter is connected to at a higher voltage than the collector, a PNP transistor can work.  So in a neg ground circuit, the emitter just goes to the supply instead of earth.

The reason why RM made Q1 a PNP is that he wanted to match it to an NPN device.  In this circuit they work in opposition, and gain is low UNLESS the negative feedback from the collector of Q2 to the emitter of Q1 is bypassed to earth via the 22uF cap and "fuzz" pot.  With low or no negative feedback the gain is high the fuzz is heavenly.
cheers

[Fret Wire]

http://diystompboxes.com/sboxforum/viewtopic.php?t=33734&postdays=0&postorder=asc&highlight=rm+octavia&start=15

There 3 MPSA13's sit in a real Octavia. I'd love to pick RM's brain about some of his Q choices. I get the feeling he has some specific hfe ranges he selects within his choice of Q's.

[analog kid]

Ok I just socketed them and I[m gonna try it. I still am not sure if you are saying  (goin in order from Q1-Q4) that the MPSA18 needs to in Q1 spot , where the PNP3906 was at??
and if any garden variety NPN small sig. would've worked , I stilll don't see why mister Mayer used the darlingtons in those spots rather than some npn THAT'S common?
BTW i did stick a MPSA18 in for the 3906 in Q1 while all the MPSA13s were still in and it really dindn't make an improvement. Is that just because of the Darlingtons?
I hope it works.
Oh , I want to NOTE something,
Is it most likely the voltage drop between Collector and Base of Q1 (2.8 to 2.77) being so small MIGHT be causing the bit of bias problem I 'm vaving??? and If so should changing to these trannies really be able to change that?

[Fret Wire]

Did you do all the standard debugging? I'd double check the traces and soldee joints with a magnifying glass. A minute strand of copper trace or solder can be just enough where the pedal still works, but drives you mad trying to get it all the way home. Same with a cold solder joint. Transistor subs may make an improvement, but it still should work fine with the same trannys RM uses.

[analog kid]

Yes I DID. Or course  because I never ask for help until after I've done all my debugging. Double checking all resistor values twice and then AGAIN. then polarities of caps and such.. Continuity through the circuit traces working off the schematic from input to output. Everything looks good on it. I usually go to the audioprobe method last OR first if I'm not getting signal to the output. I will find where it's stopping.  the answer is YES
NOTE though that I did do this on a rough etch where the traces were not all thick and shiny. I used a different etch resist and it started to eat into my resist a bit. a few spots where the traces have some line cracks in them. BUT I checked the entire board , every trace not just for continuity but for point to point resistance as well to make sure  I wasn't getting contin. but at a higher resistance. I found no weaknesses.
SO based on this , this still couldn't cause SAID problem could it?

[Fret Wire]

Sounds like you checked the board good. If a crack goes all the way through a trace, it's dead for sure. There should be the same reading consistantly through out the board. Post soldering, I always ckeck the board with a glass to make sure there are no cold or bridged joints. Magnifying glass, audio probe, and dmm are your three best tools for debugging.

It sounds like the pedal is 3/4 working, it can't be anything major. If all else fails, you might consider temporarily testing it with a plain 10k lin. or log pot.

[analog kid]

thanks, but I think the "bias" symptom is still way bad enough a prob that there DOES have to be something pretty major wrong. And I would think those voltages I posted would be enough to tell the story for someone here(?)Especially that Q1 b/e. I am fixing to sit down again with the probe and schematic and see if I can find exactly where it 's happening. BUt if anyone thinks they know PLEASE!
also BRETT, I tried that MPSA18 for Q1 and it causes the Gain;Fuzz pot to give NO EFFECT at all. Put the 3906 back in and its norma(if you call it that) >??

[Doug_H]

I would think Q4 b should be sitting closer to 4v, given the 2M2/2M2 divider.

Doug

[analog kid]

Ok that's a start! ( I have thought the prob was around either Q1 or Q4.
NOW,  a stupid audio probe question..
All the places within the circuit I should get affected signal. Are Ground and +9v included?? I don't remember if I usually get signal touching ground/+9.    Q4 is the only transistor that I don't get any signal though the probe when I touch (just the emitter) BUT that emitter is connected to the +9v.   SO should I NOT be getting anything there anyway?? I just get a bit of HUM . WHICH is also all I get when probing anything touching ground.]
thanks; Keep it rolling , I know this thing will sound good if I can get this
!

[analog kid]

OK doug I think your on it. I just went over that area with the probe again, and Q4 'B and C are the only trans.legs that are not as STRONG as the rest. AND that I can actually hear a "touch" of that gate sound if I stum really light and probe them. I would figure that Q4signal should be just as loud OR louder than the rest being closest to OUTPUT?  SO now is a matter of fixing it I guess. and I guess we also know that Q4 is recieving the proper voltage since I'm getting BATT vltg at it's Emitter.
How's this all sound to you?