Synthbox

[liquids]

No....the schematic actually has a error as mentioned above, that needs correction, but I update it later at some point, when I have access to edit those files...my computer had a meltdown this weekend, so that may not be right away, but hopefully will be soon.   I'll post here to make it clear when I've updated it.

[edd29]

No....the schematic actually has a error as mentioned above, that needs correction, but I update it later at some point, when I have access to edit those files...my computer had a meltdown this weekend, so that may not be right away, but hopefully will be soon.   I'll post here to make it clear when I've updated it.

thanks for reply! Coz I am really interested of this  and I'm hoping your update schematic.

[gigimarga]

It sounds fabulous....congratulations!

[liquids]

It's great hearing that people like how it sounds! I didn't expect this much response or interest at all.  Of course the credit goes to Joe Davisson, all I've done it tweak and experiment like the curious amateur DIY hack I am.   

I've corrected the errors on the schematic that were noted above.  Good catch Joe, thanks again!   

Has anyone breadboarded (etc) this yet?   I'm curious as to what people who are adventurous enough to try it sans a vero layout think of it in their own hands, if their results are consistent with mine, etc.   The buffer stages and low pass filtering is based on the filters in the Microsynth schematic, with a little flexibility (Added the tracking knob to control the last filter stage a bit, to each's personal taste).  But I did design it with my own rig, needs, and priorities in mind, so I'd love to hear about how it goes in other people's hands, and with other people's setups.

[isildur100]

It really sounds great! I am planning to breadboard this and eventually perf it. I recently finished a mod of Nocto Loco, but this one sounds less chaotic, and closer to what I was looking for

By the way, are the transistors critical? Can I use BC109, BC549 or 2SC1815 instead?

cheers

[oliphaunt]

I love the sound of this thing! Great job.  Looking forward to the corrected schematic and possible layout.

[Derringer]

sounds awesome man ... nice work!

[liquids]

By the way, are the transistors critical? Can I use BC109, BC549 or 2SC1815 instead?

It's worth a try, but 2N5089 and MPSA18 are cheap transistors that should be on any bench.  I would lean toward the BC549s if I had to pick one of the ones you mentioned.  But it's probably good that I explain some of the parts choices for those who are experimenting themselves--which is a good thing--rather than going direct from schematic to build.

MPSA18s are there for their low noise operation as input and output buffers, high hfe to keep the gain of the stage close to 1.  This may be overkill as, when the effect is not being played, it's kind of 'gated' silence anyway.   You could probably use darlingtons, or 2N5089s, or basically any NPN transistor I imagine without experiencing much difference.  Possibly any silicon transistor that you listed will be fine, if they're NPN.  

2nd and 3rdtransistors - I chose 2N5089s since are low noise, high HFE, cheap, relatively consistent, common, and I have them in abundance.  The second transistor is acting kind of like a buffer, but also is responsible for some of the 'gating' as its base is biased at ~2 volts above ground.  Likewise feeds a voltage to the base of the third transistor.  I'd pick any old BC549 for this stage out of the three types you listed, but again, the differences may be negligible, and BC109, 2N5088, 2N3094 will work probably quite similarly.  Don't quote me though.     The third transistors from the input is the one that is responsible for most of the gain/fuzzing, and so the higher hfe's are preferable, if I recall.  Again, of the three, BC549 is definitely best option you have as the others are lower hfe range, I think.

The 4th transistor in there with the 1N4148/1N914 diode...I don't think I did much swapping with, come to think of it;  I did mess with the diode, and the type chosen (1N4148/1N914) had the 'best' tracking I could find.  Germanium type diodes screwed up tracking, etc, and for whatever reason, 1N4001 was noticeably worse in tracking than the 1N4148, so stick with that if you aren't experimenting yourself.   That type are fairly common.

The flip flip is the most critical part.  I remember that I found that mismatching the caps affected tracking and tone, and in some cases eliminated the octaving.  Mismatching transistors  (2n3904 for one and 2N5088 in the other spot) really screwed with tracking and the octave, I found.  This made me think, "maybe if I match transistors, I'll get better tracking" and sure enough, it did.  I vaguely recall MPSA18 and darlingtons not working as well here for whatever reason, but I may be mistaken.  2N5089 have high hfe and were the easiest to match from the few I tried.  2N3904s do seem pretty consistent though, and I don't recall what the difference would have been here, actually....  2N5088s should work well too, but they are lower HFE (may not be an issue) and seemed harder to get two as close as with the 2N5089.   Note that the resistors should be the best you have, too.  On the breadboard I have 5% carbon films and that worked as you heard, but off values will likewise screw up the tracking or octave altogether.  If you can get 1% resistors, it surely wont hurt, closer is better.

If you are experimenting with transistors, set your DMM to test hfe across a batch of a certain kind of transistor.  For the flip flop, first pick two of the same kind (say 2N5089) that are the closest.  Than pick two of the same type transistor again from the said batch, but one with highest hfe in your batch and one with the lowest hfe in your batch.  Listen to the tracking differences between these two pairs--I happened to find it significant enough to note it on the schematic, as I wanted the best tracking I could muster.  I think you will too will that a matched pair will do better than 'any old' pair in the flip flop, if all else is in place for good tracking.

So in short, your transistors may work just fine, but it depends on 'where' in the circuit, and it will affect the circuit, anywhere from dramatically to negligibly.

[isildur100]

Ok thanks for all that information!

It seems tricky to fine tune but it looks like it is worth the trouble. Out of the 3 transistors I mentioned, the 2SC1815 has the lowest noise and highest hfe.

Anyways, I will experiment following your guidelines. Thanks again!

[earthtonesaudio]

I noticed that you made quite a few changes to the frequency-to-voltage converter section.  Is it just for additional lowpass filtering, or something else?

[liquids]

I noticed that you made quite a few changes to the frequency-to-voltage converter section.  Is it just for additional lowpass filtering, or something else?

I thought the F-to-V was the section with the diode and transistor combo? Be more specific, if you can..I'm just a hack.   

If you mean what I you mean, my schematic should in effect be the same as the shocktave, with the only difference in that area being an added 10k and the 68nF cap.  The DIYlc program I used to draw that schematic doesn't allow (or I haven't figured out how) to flip transistor symbols around, or how to more clearly indicate that lines crossing do not in reality intersect, etc. It's not that clear even to me on the schem, and as I said early on, the shocktave schematic should be referenced for clarity, especially in this area; it's much nicer looking.

I noticed that Joe calls the section I've been calling the 'flip flop' (and I don't even know why) the 'oscillator.'  He also says "the oscillator is tuned so that it can only keep up with every other cycle, resulting in the octave-down effect."  This is useful information.

That added R-C started as an idea for further filtering to help tracking. I think that is most of what it does.  It seems rather extreme, and it's interesting to tap the audio at this junction.  you'd think that no sound would get through...but it works somehow.  Most of what I added, other than the buffers, is a lot of filtering for tracking.  It still happens to sound good and aggressive despite that.

With the steep roll off, it filters out any sound of the 'unoctaved' signal well, to boot, so it sounds more synthesized as well as tracking better and having better decay characteristics.    The Roll off corner for 10k/.068uF is 234hz or lower!   I was baffled it even works...

Originally I added no resistor and just put a .1uF cap in there, and that worked to some degree.  With the 10k resistor in and a .1uF cap, everything works fine, until notes on the E string around 12th fret seem to be unable to 'keep up.' Yet rather than making no sound, it tracks those notes as '2 octaves down.'  AS counterpoint, too much filtering early on more so makes those notes inaudible or far too short in decay (so be careful with the tracking trim).  The bigger you make that cap, the more you get that affect of not being able to keep up with the high notes and reading them as 2 octaves down.   This sounds good, but it quickly becomes becomes a monotone bleep bleep bleep no matter what note you play. OH yeah, If you want two octaves down, you can run two in series, or an octaver into this circuit. I had an old prototype and ran it into this, and it indeed octaved 2 down.  I find 2 octaves down impractical, but to each their own.   

I digress.  In my mind, I only understand this part as lowpass filtering, but it may affect the 'oscillator' section too--I don't understand that part at all, so I'm not one to say.  This is certainly an area that could potentially be improved, by someone who understands what is happening more than I do--that is, someone who understands it at all.   

[Joe]

I wanted to replace the flip-flop with something that would divide by an adjustable amount, in hopes of using a rotary switch to select harmonies. If anyone has any ideas for that, I'm all ears.

[soggybag]

You've inspired me to give this one another try.

I tried to build a Shocktave once using an LM3086. This is a transistor array on a 14 pin dip. I thought it might work well since the transistors are matched. This would be good for the two that form the VCO that generates the octave. Not too mention it might make for a nice build since one part would replace all 5 transistors.

It didn't work on the breadboard so I gave up. It could have been that the transistors on the chip might not have enough gain for this application or just an error on my part. 

[isildur100]

Looking at the datasheet for the LM3086, the hfe is only around 100, which is probably too low for this circuit. This could explain your problems.

[~arph]

Hi,

Ive breadboarded this yesterday and I can confirm the schematics. It works like a charm. I love the sound and it tracks very nice. I just saw that I misread the 68nF value for 68pF (which I did not have, so I put in 51pF). I'll swap it tonight and see if the tracking/sustain gets better as you say. I did not find the tracking pot to be very useful, I set it to 0k for the best tracking. I was going to do a vero for this, but the component count is just a bit to high for vero, it will be a big board.

Now how about adding a mu-doubler in parallel, so we get into micro POG territory?

I'm definately going to play around with this and then box it up for a friend who will love this for sure.

[doitle]

Wow this thing sounds neato to the max. I've gotta put this on my list of pedals to build when I get better at all this stuff.

[solderman]

The schematic had some errors and a thing or two missing, so I updated it:

http://sites.google.com/site/liquidselectronics/synthbox

I found that the first link only goes to the original version, which will be removed.  This link should stay updated to whatever is the most recent version, if I ever do work on the schematic.  Once you're there, clicking on the image should isolate the page to view the image only, and make it clearer.

I did record a long sound sample as per Derringer's request, but I need to whittle it down a bit; that should be coming soon, for those who are interested!   

Hi
Really nice sound in this one. What about morph this one with the oct up/down capability a 4024 an a 4070. To take the signal from "Volume" then mix it with the oct UP and/or oct down like in the Slacktave MK II.   

http://www.diystompboxes.com/smfforum/index.php?topic=76695.0

[earthtonesaudio]

I wanted to replace the flip-flop with something that would divide by an adjustable amount, in hopes of using a rotary switch to select harmonies. If anyone has any ideas for that, I'm all ears.

The main route to do this analog is with PLLs.  Recently Nelson put up an unverified schematic for his "Glitchshifter" which might help you with ideas.

[liquids]

Ive breadboarded this yesterday and I can confirm the schematics. It works like a charm.

Great!  This helps my credibility.     Isn't breadboarding great?  (had to sneak that in there).

I love the sound and it tracks very nice. I just saw that I misread the 68nF value for 68pF (which I did not have, so I put in 51pF). I'll swap it tonight and see if the tracking/sustain gets better as you say.

You may find you don't need any cap there if you liked the tracking with '51pf' instead of 68uF and satisfied. Feel free to experiment.    Very easy with a breadboard!  If you don't find need that R-C filter stage, pull out that resistor, too.  I'm very interested to see what your findings are here.  How do you compare the decay/trail off compared to the same I recorded right now?  Is it more glitchy and more 'arpegiating' at the end of note decay/gate, or about the same?

I did not find the tracking pot to be very useful, I set it to 0k for the best tracking. I was going to do a vero for this, but the component count is just a bit to high for vero, it will be a big board.

Did you use the 5k? 0k as in not there, or full up? I originally designed it without the tracking pot. The values are based on the low pass filter design in the microsynth, based on the frequency of roll off, rather than exactly imitate the values. I wanted to likewise  keep the resistor values fairly low, and shoot for no more than 10k total out of the buffer, hence the three x 3.3k which worked rightly with .022uF; the roll off corner is around 2200hz. 

As I edited the schematic, I though to add a the tracking pot for others, realizing some people might not play so cleanly, and/or have brighter guitars than I do, and might need more high end roll off for better tracking.  I play this mostly with a bridge humbuckers, so it needs quite a bit of harmonic filtering--switch to a neck position and it's pretty effortless and easier to design around, but that was part of the fun. 

I worked it a bit more and found 10k to have a better range, as the 5K was still very subtle (there are 3 stages of filtering as it is!), but that is coming from someone who didn't need it in there anyway.  Anyone could just unsolder or socket that third 3.3k resistor on a final build anyhow and replace it with what tehy prefer--it's mostly set it and forget it pot. I may just leave the tracking trim out of mine, since I did design it with my rig in mind, without the tracking pot.     I did try it as a 25k pot on the breadboard for the heck of it (I had no 10k on hand that second), and it compromised the high notes when turned up near and past halfway (10Kish) which is starting to roll off around 800 or 400hz if  I recall...but it was otherwise subtle, and not necessary for me since I was getting fine tracking before.  The only benefit was the decay on the lower strings on the lower frets did smooth out and glitch less at the end, but losing the highest notes played was a compromise.  If you remove the trimmer, it will be a smaller board on vero.   

I started a rough vero middle of the week, and it's very tight, but I should be able to get it on a very messy 15/30 veroboard, or smaller...but it's going to take a few revisions, and time is limited in the coming weeks...ah well.  Maybe someone can do a small run of  PCBs?  I'd even buy one!   

Now how about adding a mu-doubler in parallel, so we get into micro POG territory?

I personally have no interest, especially since my Octron and microsynth have that octave down/straight/octave up mix.   So that is beyond the sound I'm going for with this particular design; that would also only make this design that much bigger, but go ahead!  that is what DIY is all about.   

I'm definitely going to play around with this and then box it up for a friend who will love this for sure.

Cheers to that!   One more - how did you find the 'shape'/tone control?  Too drastic, too subtle, or good?

Great to hear from people who have given it a try hands on...keep it coming!

[~arph]

I'll do some more testing this weekend. Yesterday was late so I can't recall all my exact settings.
I haven't really tried the shape control, I stuck a fixed 2k2 in there. Sounded the best. Not too sharp and not too dull.

Currently the decay is a bit more glitchy as in your sound sample. I'll do some testing around the tracking pt. BTW I used Single coils to test this, (mostly neck p/u) but my pu's are on the hot side (Lindy Fralin Vintage Hot's)