Synthbox

[gigimarga]

Thx a lot liquids!
I've read all that you send to me and I understood that's not possible to change only the caps from the oscillator to obtain an octave up.
Anyway, it's a very good soundinf effect....thx again for it!

[liquids]

News:  I've finally updated my version of the the schematic per gigimarga's suggestions.   Still no vero layout for it.  Maybe someday. 

I've learned a bit even since I first worked on this project.   I've finally gotten to reading up and experimenting with the basics of op amps (basics only).   Now I understanding that the Microsynths early filtering is active and the advantages of that etc, which is what finally brought me back to working with this project.  I still haven't built one myself other than whats been on the breadboard!    Anyhow, duplicating and tweaking around the microsynths ACTIVE filter more accurately afforded some other little tweaks, which have all yielded a little better tracking.   

I'm also toying with the idea using your basic active EQ filtering at the end (likely based on what's on the LM833 datasheet) using a second dual op amp instead of the passive filtering options, and eliminating the BJT buffers for op amps....  It's still a work in progress.   And if/when I get that version of the circuit where like it I'll probably post a schematic of that version, too.   

[gigimarga]

News:  I've finally updated my version of the the schematic per gigimarga's suggestions.   Still no vero layout for it.  Maybe someday. 

I've learned a bit even since I first worked on this project.   I've finally gotten to reading up and experimenting with the basics of op amps (basics only).   Now I understanding that the Microsynths early filtering is active and the advantages of that etc, which is what finally brought me back to working with this project.  I still haven't built one myself other than whats been on the breadboard!    Anyhow, duplicating and tweaking around the microsynths ACTIVE filter more accurately afforded some other little tweaks, which have all yielded a little better tracking.   

I'm also toying with the idea using your basic active EQ filtering at the end (likely based on what's on the LM833 datasheet) using a second dual op amp instead of the passive filtering options, and eliminating the BJT buffers for op amps....  It's still a work in progress.   And if/when I get that version of the circuit where like it I'll probably post a schematic of that version, too.   

I'm feeling very honoured to be mentioned here (you're very kindly liquids, because I think that my contribution is almost zero...) and I'm very glad to hear that you will try to improve it (and impatient!).

Best regards!

[isildur100]

Hi guys,

I finally found some time to put this on my breadboard  I've been playing around with it for almost a week now. Here are my observations.

First, I built it exactly like the schem (the version I made a perf layout for). Unfortunately, I was not able to obtain any good sounds mainly because of bad tracking. The octave down could only be heard once in a while. The tracking trim pot did not change anything for me so I removed it. I did use the transistors recommended and tried different 2N5089's for finding the best match possible for the oscillator section. The matching made a difference but not enough to be satisfying.

Also, like gigmarga, I found the switchable 10nf cap at the end sounded very thin, so I left it with a fixed 1uf cap. But that's my taste

Just as I was going to quit on it I decided to try a bit of tweaking Here's an interesting tweak I found:

I was able to make it track almost perfectly by changing the biasing at the base of the second transistor (the 680k and 330k voltage divider). This solved it !!

At first I replaced the 2 resistors with a 1M pot and and found where the sweet spot was. There is a range of around 100k in there that is track-able and depending if you want to play very low notes or higher notes, you can set it where it tracks better at those note frequencies. So I replaced the 1M pot with a 100k pot in between resistors in order to map the usable bias range to the full range of the pot. It looks like this:



The resistor values look strange but that's what it took to properly spread the usable bias range on the 100k pot. What is really cool is that you can obtain different sound textures when turning that pot.

Also, I found that it tracks better when using more than 9V. I used a 9v wall wart outputing around 14v and it sounded better than when I used a 9v battery...

Anyways, I thought I might share my experience and give hope to those having tracking problems with this cool effect.

cheers

[~arph]

Cool, will try this too.. I'll start with wiring up a voltage doubler 

[liquids]

Isildur - your mods are interesting and make a lot of sense in some way.  Thanks for your input!  You're right that the base biasing of that BJT stage is critical to the tracking, gating, etc.   It works heavily in conjunction with the collector and emitter biasing, keep in mind.   While  I like the 680k/330k (2:1), which is in the range of what you are doing, for us tweakers who don't mind trimpots even on sizable builds(!), why not?   But it could be improved....

However, I've breadboarded my schematic quite a few times over, and had none of the issues mentioned.  The octave down not sounding sounds like an issue in the oscillator section, though it could be the gain section, as your fix would indicate. Still, check your connections and values.  You shouldn't have that issue even slightly, matched parts or not so long as you have the same parts and values, so keep with it....the trimpot may help and even improve things, but it shouldn't be 'needed' to fix the issues you mentioned.

I also find the 680k/330k to have had the best tracking after lots of resistor swaps, (or 820k/390k or 560k/270k much the same) though your trimpot is a nice addition to get it perfect by ear.  The original ratio is (by ratio) within the trimpots range.  However, you may want to 'double all your values' if nothing else.     That is, the input impedance of that stage is affected by these values.  With the schematic it is already low (lets say it's 200k now).  With a 73k (?) resistor, the impedance is probably ~50k, which means you are probably losing a fair amount of signal at that end of the trimpot...combined withh the 20-30k of resistance  feeding it = no good.

Maybe make your 74k  resistor ~330k again, and do the rest of the math from there...something like 1M, 330k, and 500k trimpot, or 470k/220k and 250k at lowest.   Worry less about the extremes end of the trimpot range functioning exactly correctly...it's all about compromise.    At that point, maybe a pot is even merrited on your build...try and see.   Just some thought.

Along with that, I tried it with higher voltages myself.  I didn't notice any significant difference - in fact I noticed it was a bit worse, because the higher voltage makes the BJT stage drive the rest harder....  You need a fair amount of gain to drive the F-V and oscillator, however, too much and it goes wacky.  Note that you may have liked the higher voltage either because of a weakened battery by comparison, or because of lack of designed amount of gain, in one or multiple places...as mentioned above.

So the need for the trimpot and prefering higher voltage  leads me to believe that your issues may come surrounding the gain stage, if not more...  Be sure the emitter bypass cap (47uF) is connected and properly sized (at least 22uF or biffer).

Last but not least, the bass trim.  I tried it again and found 10n a subtle change...but my amp is bass heavy...and there may be other factors.  I think that was bad design anyhow.   A better place for a cap switch would be between the oscillator and last buffer - far more effective.   Or none at all...

That being said,if you're up for it, I've found following the circuit with an op amp buffer feeding a Baxandall like this (http://sound.westhost.com/dwopa2.htm#baxandall), which would give more interactive control over boosting/cutting both sub bass and treble than passive treble cut controls and a 2 position bass switch.   Give it a shot....

All that said, again, thanks for your input!  I'll tinker with your ideas in short time I'm sure.     And more great tweaks coming! Arnoud (~arph) is helping me out....and I found another synthy TRUMPET sound in here...to be continued!  

[isildur100]

Hi liquids,

I also find the 680k/330k to have had the best tracking after lots of resistor swaps, (or 820k/390k or 560k/270k much the same) though your trimpot is a nice addition to get it perfect by ear.  The original ratio is (by ratio) within the trimpots range.  However, you may want to 'double all your values' if nothing else.     That is, the input impedance of that stage is affected by these values.  With the schematic it is already low (lets say it's 200k now).  With a 73k (?) resistor, the impedance is probably ~50k, which means you are probably losing a fair amount of signal at that end of the trimpot...combined withh the 20-30k of resistance  feeding it = no good.

Maybe make your 74k  resistor ~330k again, and do the rest of the math from there...something like 1M, 330k, and 500k trimpot, or 470k/220k and 250k at lowest.   Worry less about the extremes end of the trimpot range functioning exactly correctly...it's all about compromise.    At that point, maybe a pot is even merrited on your build...try and see.   Just some thought.

In fact I am using a real pot, not a trim pot. I like to be able to play with that pot because it lets me choose different sound textures and adjust tracking depending on the pickups I am using. I have tried at first with a 1M pot, then with a 500K pot with resistors, then with a 100k pot with different combinations of resistors. To my ears, the best sounds were obtained with the values I mentioned in the previous post. It may be lowering the input impedance, I don't know but it just tracked better like that. I was able to get very long sustain on the low E when adjusting this pot to the sweet spot 

Along with that, I tried it with higher voltages myself.  I didn't notice any significant difference - in fact I noticed it was a bit worse, because the higher voltage makes the BJT stage drive the rest harder....  You need a fair amount of gain to drive the F-V and oscillator, however, too much and it goes wacky.  Note that you may have liked the higher voltage either because of a weakened battery by comparison, or because of lack of designed amount of gain, in one or multiple places...as mentioned above.

When using different voltages, the bias pot really comes in handy. I had to adjust the pot depending on the input voltage. For me, that's another reason to keep it as a pot instead of a fixed v divider.

So the need for the trimpot and prefering higher voltage  leads me to believe that your issues may come surrounding the gain stage, if not more...  Be sure the emitter bypass cap (47uF) is connected and properly sized (at least 22uF or biffer).

I had this set up correctly

That being said,if you're up for it, I've found following the circuit with an op amp buffer feeding a Baxandall like this (http://sound.westhost.com/dwopa2.htm#baxandall), which would give more interactive control over boosting/cutting both sub bass and treble than passive treble cut controls and a 2 position bass switch.   Give it a shot....

All that said, again, thanks for your input!  I'll tinker with your ideas in short time I'm sure.     And more great tweaks coming! Arnoud (~arph) is helping me out....and I found another synthy TRUMPET sound in here...to be continued!  

I plan on keeping the synthbox on my breadboard for a little while. I will try to post some sound samples so we can compare with real results. It will give you an idea of the sound differences when adjusting the bias pot (which I now call the tracking pot)

I can't wait to hear your trumpet sound!

I will keep in touch!

cheers

[gigimarga]

I'm waiting for the trumpet...COOOOOOOOOOOOL!!!

[~arph]

it sounds like a sax allready.

[isildur100]

LOL

Sound samples please!!!  

I was able to get a somewhat trumpet-like sound by putting a 4.7n cap across the base and the emitter of the third transistor (where the diode is (leaving the diode there)). Again it could be adjusted just right with the bias pot I talked about.

I also tried it before my superfuzz and man, massively heavy sounds!  

[thedefog]

This thing is quite a different animal in comparison to all of the other octave-type effects I've heard/owned/built. And it's pretty simple too! Definitely going to build this one... Possibly tonight.

[liquids]

Okay, here's the schematic for the op amp version.  http://sites.google.com/site/liquidselectronics/synthbox-2

It adds an active EQ and some tweaks for filtering, in addition to some new sounds via the "attack" pot.    The attack pot offers the more 'trumpet like' sounds, but don't take this too literally...with the sound clips of the synthbox as its base sound, the above schematic can get more 'trumpet like' in terms of the attack of the note.   It does offer a cool variation on the tonality of the original sound in my opinion, with subtle shades of blending between the two in between with the added attack pot...the 100nF is crucial, for whatever reason this will not happen with the stock 'shocktave' schematics 1uF cap in there.

I'm hopefully done tweaking it!  Other people may tweak it to their liking.  I tried biasing the base of the first BJT per isildur's concept, and twisted the pot to hear the range it offered, then set it to the place I liked it best by ear.  Then I measured it with a DMM for voltage at the base and the resistance values, only to find it exactly where the 330k/680k resistors bias it.     I also didn't find it that useful, so I'm leaving the biasing as is since I find it optimal.

Note that if you breadboard this up, you can actually get even more interesting sounds, like a 'pure' trumpet-type sound...though bare in mind, it's more like a cheap guitar synth trying to 'imitate' a trumpet sound...but very cool nonetheless.  I found it a bit too much of a novelty and too much trouble to merrit all the controls necessary to add it to the design, and some of tim escobedo's circuit snippets could probably be tweaked similarly....    For those who like that sound may find it merits a completely separate build with filtering designed around that sound.  If you breadboard it and are interested in hearing it, try this: starting with the above version (you'd have to tweak values a bit if you start with the original version), break the connection between 9V and the collector of the transistor that is fed by the new 'attack' pot/10k.   This would leave the collector of that transistor with only the 100nF cap connected to ground.  This will now make the tone 'sweet and smooth,' and remove all agressiveness from the sound.  But it will be very dark and muffled.  You will then want to remove, or greatly reduce the 68n capacitor that comes after 'B' and the 10k resistor to something like 2.2n.    

Tracking is a non issue when you break that 9v connection as mentioned above, so the 68n is not needed, though it's pretty key to good tracking with the stock synthbox sounds along the range of the attack pot.  Likewise, some of the other filtering stages may beg tweaking for the best sounds when you break that connection, even after removing the 68n cap.  

Eventually I'll get someone to do a PCB of the op amp version for me and build it--I don't know that I have the patience or room to work out a vero for this one!  And if you want sound clips of the new sounds available...well, breadboard it.    Enjoy!  

[isildur100]

Hi liquids,

Congrats for the new design! 

It would be nice if you could put some sound samples just so we see how different it sounds from the first version.

As for the biasing question, the thing is that your fixed setup may be good for the voltage and guitar pickups you have but someone else with another guitar, different setup etc it will not necessarily have it tracked perfectly like you. It's only my opinion but I think that having the possibility to adjust that is a plus And I don't know if you experienced the same thing but I find there are more than one sweet spots offering different sounds and better tracking for playing higher or lower notes.

I will definitely try your "attack" pot with the version I have on the breadboard.

cheers

[liquids]

Finally 'finished' the vero layout for the 'first' All BJT version.  I simplified it via removing the tracking pot so that it would all fit on 15x30 piece of vero.  I also updated some of the component values.

If you search for synthbox in the layout gallery, you'll find both the vero and the schematic that more closely matches it.  I needs some other eyes to look at it before anyone should build it, since it may have errors....I can only look at it so much.  

Since it is unverified, and rather than load it as a huge image, and since there are so many cuts and so many components, I just loaded it as a working file to be downloaded and opened in DIYlc.  This way, you can save it yourself, and also 'remove' all the components just to see the board and cuts, invert the image how you like, and drill properly, since there are so many cuts under components.

I like to cut almost everywhere that traces aren't needed.  It takes more time, but I find it easier in the long run...since there are more places to drill than not, I just need to know the where NOT to drill rather than count off the trace breaks.  Then, when I'm  inserting components or soldering, I'm less likely to make an error with most component's placement or bridge a trace....however, you can remove all but the essential trace breaks if you like, after you download the file.  Since I imagine there will be many, PM me if you have questions about the layout, corrects, etc, and I'll fix anything that needs clarification or correction.  Good luck!

[liquids]

Anyone else try the circuit, successfully or unsuccessfully with single coils?  My experience, testing, etc for this is mostly (if not completely) limited to humbuckers.  I've been working with isildur, and I'm now wondering if some of the different results and complications we've had are due to pickup differences.  Any experience and feedback about that (via PM, ideally) would be appreciated.

if so, and it's simply a matter of a way to boost the signal, a simple gain control (as a mod to the input buffer, or in addition to it) could probably be implemented.  I may experiment with that one myself, as a matter of fact...

[~arph]

Hi,

Yes I did put it on breadboard, also without the tone section, but with an opamp buffer at the end.
I had far worse tracking on the opamp version over the original transistor version. I am using the same guitar, high output single coils (lindy fralin vintage hot's). Tracking seems to be the best at a certain input level, not too loud ,and not too soft. This leads me to the conclusion that to make this thing really good we have to ensure a constant input for the f-v section. I am currently experimenting with a NE571 in front. Still looking where to place this, before or after filtering the signal.

Also I found that the filtering after the 5089 right before the F-V converter has no noticable effect on the tracking (I both versions of the synthbox)

Still playing around with this.

I'll report back.

[soffa]

I've got the circuit working but I've only tried it with single coils. I have a guitar with a humbucker as well so as soon as I get a chance I'll try that as well.

So far the circuit sounds pretty close to "correct" but I don't always get the octave down. I am going to try some of the suggestions in this thread WRT that problem but I haven't had time to get back to it yet. Based on what others are saying maybe I'll try to get it working  with the humbucker-equipped guitar first and then move back to the single coils...

[thedefog]

Sorry to continue an older thread, but I'm having some issues with my build. I followed the perfboard layout by isildur, and I am getting ticking from the oscillator. The gate doesn't seem to be working properly. I used 2n5088s where the 2n5089s were listed in the schematic, and MPSa18s where they were supposed to be. Any thoughts on this one?

[liquids]

No problem reviving this...it's still a 'work in progress.'  I may even have a 3rd version which is mostly op amp, and more user friendly in the near future.

Ticking? There is no LFO so I'm not sure I understand you....do you mean, when you aren't playing it is still 'making a note?'   If so, that should be fairly easy to resolve.

Following the layout, Check your traces for any stray solder connections especially around Q3.  Get your voltages on Q3. Either the base or the emitter - I can't recall which right now, but one of them should mostly be variable as you play a note.  The 'no note playing' voltage should be below .6 volts, which is a key part of the 'gating' and 'decay.'   Examine this area carefully, I suspect that is where the problem is.  Audio probe carefully here as well...it may give you clues. 

Also, the electrolytic cap that feeds that stage, right before the diode and Q3...you may want to reverse it.....and if you get it working, try replacing the 1uF and 47k pair with .1uF and 10k, you may like it better (maybe not!)

Good luck, let us know what you find.

[gigimarga]

...
I used 2n5088s where the 2n5089s were listed in the schematic, and MPSa18s where they were supposed to be.

I've built it using BC549C and it works like charm, so I don't think that transistors are the cause of the problem.
Anyway, if I'm remebering right, 2N5089 is the same as 2N5088, but with lower noise.