3 or 4 octave up square wave ??

[markusw]

Hi all,

whats the best/easiest/most reliable way to get a 3 or 4 ocatves up square wave signal from a guitar or bass signal?

Thanks for your comments in advance.

Regards,

Markus

[octafish]

I'll hazard a guess and say Tim Escobedo's Square Wave Shaper and a store bought pitch shifter. I guess you could build three Green Ringers for the octave ups.

[Paul Perry (Frostwave)]

Well, if you had a well-behaved and reguar square wave, you could use a pic, just count the time between transistions & divide by four (or whatever) and put out a square wave at the higher frequency, remembering to extend or cut short the last cycle to match the original signal.
The other answer is, to use a phase locked loop, but the time taken to acquire lock is the problem there.
but in practice, the problem will be, getting the square wave in the first place. If it is coming from a guitar.

[markusw]

Thanks a lot for your suggestions.

I'll hazard a guess and say Tim Escobedo's Square Wave Shaper and a store bought pitch shifter. I guess you could build three Green Ringers for the octave ups.

Do you mean first run the 3 Ringers and then square wave the signal or the other way round? Would the green ringer also work on an already squared signal?

Well, if you had a well-behaved and reguar square wave, you could use a pic, just count the time between transistions & divide by four (or whatever) and put out a square wave at the higher frequency, remembering to extend or cut short the last cycle to match the original signal.
The other answer is, to use a phase locked loop, but the time taken to acquire lock is the problem there.
but in practice, the problem will be, getting the square wave in the first place. If it is coming from a guitar.

I think I'm not ready to start with pics. Actually, my idea (not very extravagant I know  and also quite expensive) was to first square wave the guitar signal and then run it through 3-4 balanced ring mods to get the octaves. So simply running the guitar/bass signal through a comparator opamp won't do the trick??? About the phase locked loop idea I will first have to check what it actually means  Any links for a good read?

Markus

[Gladmarr]

This may help a bit with Phase Locked Loops:  http://www.uoguelph.ca/~antoon/gadgets/pll/pll.html

[markusw]

This may help a bit with Phase Locked Loops: http://www.uoguelph.ca/~antoon/gadgets/pll/pll.html

Thanks for the link  Will take me some time to understand.

Markus

[toneman]

Well, it's not a square wave, but, do a search for jonny octave in this forum.
Yes, the spelling is correct.
There was some discussion a while back about octave up...
2 Octavias in series......like the Octavia, works only on *some* parts of the guitar.
Bass?   Don't recall an "Octavia 4 bass"....
MayB someone knows if the Octavia will work w/bass.

staymultiplied
tone

[gez]

I'll second the advice about going with a PLL.  Use in conjunction with a flip-flop or two...or some other divider.

Need to feed it with a good strong square wave to get good tracking.  I just drive the hell out of simple sine to square converters and it seems to work well.  Bit of splutter at the end but you can use an envelope controlled VCA to gate it...or use a monostable to cut it off short (ouch!).

[Connoisseur of Distortion]

look at Tim Escobedo's Digital Octave Fuzz. i think that adding a few more stages like the second one could make this a ton easier. you might even fit the entire circuit using a single IC... 

EDIT: i actually think that there's a problem here... you would need to set the high threshold higher after the first octave, or else it would just flip it over. i know my language is totally messing this up, but the idea is that if it never crosses that center line, it can never be octaved. hmmmmmmm...

[Paul Perry (Frostwave)]

So simply running the guitar/bass signal through a comparator opamp won't do the trick???

No, tragically.
A ring modulator used as a frequency doubler, only works perfectly for sine waves. (you can see this from the relevant trigonometric identities, which I forgot 40 years ago).
And secondly, the idea of using a comparator to get ringmod... it's going to be pretty rrRRrrrrough!!!!!
But it all depends what kind of a tone you wnat or find acceptable at the end of the day.

EDIT: I just had a flashback.. the first FX I ever tried to build, was a bass frequency doubler! I thought I could just full-wave rectify the signal and not filter it. Well, it sounded pretty much the same.. because, the crappy speaker my mate was using, wasn't giving any response but overtones anyway!! that was the same time I was trying to fix Rex's amp, for anyone who frequents the Lounge

[markusw]

Thanks again for all you suggestions.

OK. If i got it right one possibility would be to first  square the wave and the run it to a PLL with a divider of 3 of 4.

Well, it's not a square wave, but, do a search for jonny octave in this forum.
staymultiplied
tone

I would like to have the 3-4 ocatves up as a square wave to build some add-on box for my Ring Stinger clone (I like the square wave VCO mode most with my Ring Stinger clone). Instead of using the internal VCO for carrier I would like to have the 3-4 ocatves up square wave. Should sound pretty cool (and maybe a little bit more harmonic than a fixed or oscillating freq). But thanks anyway.

I'll second the advice about going with a PLL.  Use in conjunction with a flip-flop or two...or some other divider.

Need to feed it with a good strong square wave to get good tracking.  I just drive the hell out of simple sine to square converters and it seems to work well.  Bit of splutter at the end but you can use an envelope controlled VCA to gate it...or use a monostable to cut it off short (ouch!).

So, what sine to square converter would you recommend?

look at Tim Escobedo's Digital Octave Fuzz. i think that adding a few more stages like the second one could make this a ton easier. you might even fit the entire circuit using a single IC... 

EDIT: i actually think that there's a problem here... you would need to set the high threshold higher after the first octave, or else it would just flip it over. i know my language is totally messing this up, but the idea is that if it never crosses that center line, it can never be octaved. hmmmmmmm...

Single IC..sounds cool. Will definitely check it out. Also I will try to do some sims in LTSpice to understand the treshold you mentioned.

So simply running the guitar/bass signal through a comparator opamp won't do the trick???

No, tragically.
A ring modulator used as a frequency doubler, only works perfectly for sine waves. (you can see this from the relevant trigonometric identities, which I forgot 40 years ago).
And secondly, the idea of using a comparator to get ringmod... it's going to be pretty rrRRrrrrough!!!!!
But it all depends what kind of a tone you wnat or find acceptable at the end of the day.

Did'nt know that balanced modulator chips as frequency doubles don't work with square waves. Thanks for the hint. Regarding the comparator, I wanted to use it for getting the square wave not for the ring mod. Did I miss something or could this work?

BTW, are there any schems for PLLs that might help me with understanding that divider thingy ??

Markus

[gez]

Something like this, but not this:



The second amp and its associated circuitry (IM cap and 1u cap) is the sine to square converter.  It’s just a comparator really, the +ve input being compared to the ‘average’ DC level at any one time on the -ve input.  Standard circuit straight out of the text books; primitive, but it works.  Depending on what amp you use (the one quoted has miniscule offsets) you might want to make the resistor smaller and the cap larger.

Ideally you’d want some filtering beforehand to improve tracking, but it’s not too important with octave up circuits.  The first amp in the above schematic is just a buffer, but if you increase the gain you’re going to get a more suitable signal to drive the input of a 4046.

BTW, are there any schems for PLLs that might help me with understanding that divider thingy ??

Yes, there’s a Boss feedbacker thingy out there, but I can’t remember its name (link anyone?).  It will give you an idea of connections etc.  Lancaster’s CMOS Cookbook explains the basics and plenty of books cover the oscillator in the 4046.

Actually, an interesting thing happens with the circuit I posted (not serious by the way, just a bit of fun).  With CMOS amps the outputs actually try and bias up towards half supply rather than slam into one rail or the other and stay there (the norm with most types of op- amp).  As you’d expect, bias is all over the place so its really noisy in its ‘quiescent’ state, but the end result is the sides of the output waveform slope towards the tail end of notes as the op-amp acts as a ‘normal’ amplifier.  This introduces dynamics into the mix.  Not in terms of amplitude, but in timbre.  As you’d expect its not a huge amount of dynamics (the sides don’t slope that much), but definitely noticeable.

[markusw]

Something like this, but not this


Ideally you’d want some filtering beforehand to improve tracking, but it’s not too important with octave up circuits. The first amp in the above schematic is just a buffer, but if you increase the gain you’re going to get a more suitable signal to drive the input of a 4046.

BTW, are there any schems for PLLs that might help me with understanding that divider thingy ??

Yes, there’s a Boss feedbacker thingy out there, but I can’t remember its name (link anyone?). It will give you an idea of connections etc. Lancaster’s CMOS Cookbook explains the basics and plenty of books cover the oscillator in the 4046.

Actually, an interesting thing happens with the circuit I posted (not serious by the way, just a bit of fun). With CMOS amps the outputs actually try and bias up towards half supply rather than slam into one rail or the other and stay there (the norm with most types of op- amp). As you’d expect, bias is all over the place so its really noisy in its ‘quiescent’ state, but the end result is the sides of the output waveform slope towards the tail end of notes as the op-amp acts as a ‘normal’ amplifier. This introduces dynamics into the mix. Not in terms of amplitude, but in timbre. As you’d expect its not a huge amount of dynamics (the sides don’t slope that much), but definitely noticeable.

Why something like this but NOT THIS??

Filtering to improve tracking: do you mean high or low pass filtering? 

Another Q: would there be any benefit in configuring the comparator as Schmitt trigger?

Will have a search for that Boss schem. Thanks for the tips!

About that CMOS bias stuff I will have to think/read a little more to understand it. My interpretation so far: the waveform changes with guitar signal inensity, with stronger signals giving a more or less "perfect" square wave that gets distorted (sloped) once the signal gets weaker. Sorry if this complete rubbish.

Markus

[jmusser]

Is the Square Head basically a square wave fuzz that you can use to go into flip flops for inverters for octave up or down? Does it sound pretty good on it's own? I never heard of the L7621D chip before.

[gez]

Why something like this but NOT THIS??

Because what I posted is deliberately designed to make use of the dynamics that CMOS amps provide in this circuit.  Your understanding of how it works is correct by the way.

To make it work for what you want you’d need to increase the gain of the first stage to avoid splutter (stronger square wave for longer) and use a bipolar amp so that you feed the 4046 with a nice square wave.  Actually, the input of the 4046 that normally gets used for these things isn’t too picky about what you feed it compared to some logic chips, but it does help avoid tracking problems in octave up circuits if you feed it with something it can get its teeth into, so to speak.  Otherwise you can get odd noises such as gliss as the internal oscillator tries to keep up with the input signal.

Another Q: would there be any benefit in configuring the comparator as Schmitt trigger?

You can, certainly, but the threshold for triggering is at 1/3 and 2/3 of the supply so you need to amplify your input signal a hell of a lot otherwise you’ll get a strong gating effect (you’ll get it anyway).  With the above design even small signals will cause the comparator to act, but that can be a downside too, i.e. noisy when not playing - hence the mention of an envelope controlled VCA in original post.  You can use variations on a theme to set a threshold that helps prevent this tetchiness.

Filtering to improve tracking: do you mean high or low pass filtering?

Low pass.  It’s not necessarily crucial with octave up stuff as yodelling isn’t so noticeable, but if you’re doing a few octaves up that may not be the case.  A simple 18dB per octave filter like in many octave down circuits should help, but you might not need it at all.  Breadboard and decide yourself.

[gez]

Is the Square Head basically a square wave fuzz that you can use to go into flip flops for inverters for octave up or down? Does it sound pretty good on it's own? I never heard of the L7621D chip before.

It's just a comparator fuzz (square wave) but with some dynamics.  Not really useful for driving flip flops with a CMOS chip op-amp (as shown) unless you increase the gain of the first stage of the above circuit considerably.  Even then many clock inputs of flip flops are picky.  I usually follow such a circuit with a Schmidt trigger input logic chip acting as a buffer (555 timer is good at doing this).

I've only ever used two CMOS output amps in the above circuit and both had 'timbral' dynamics.  Should think most CMOS would work, suck it and see.

I prefer it to normal comparator fuzzes.

[gez]

PS  Mouser sell the chip in the US.

[jmusser]

It's really odd I've never seen this circuit, but then again, I never new Joe Gagan had a Tarpit Meltdown Fuzz until two days ago! Thanks for the info Gez. I love the buzzy square wave sound.

[markusw]

Slowly I get it (at least I believe   ) Thanks for your patience 

Regarding the spluttering at the end of the notes and the VCA. Would it help to compress the guitar/bass signal (to increase gain at the end of the note) maybe followed by some kind of gate to turn the signal of once the signal drops below a certain treshold (to prevent noise entering the comparator)? But most likey this is similar to what you meant with

With the above design even small signals will cause the comparator to act, but that can be a downside too, i.e. noisy when not playing - hence the mention of an envelope controlled VCA in original post.  You can use variations on a theme to set a threshold that helps prevent this tetchiness.

Hm, seems that it will require fine tuning to have the octave signal as long as possible but to stop it just before it enters spluttering. On the other hand the spluttering might give an interesting sound too with the subsequent passive ring mod.

Any links for the VCA + treshold ??

The low pass filter I will give a try for sure  

I'm really curious how this all works out because the Ring Stinger gives a really cool sound (IMHO) if the square wave VCO is tuned at 3 or 4 octave up. Seems to be a bit more complicated than I initially thought though.

BTW, this http://home1.gte.net/res0658s/fatman/4046pll.html should be a good start for frequency multiplying I assume. Just hope that there are simpler ways to get the 3-4 divider part.

Markus

[gez]

Regarding the spluttering at the end of the notes and the VCA. Would it help to compress the guitar/bass signal (to increase gain at the end of the note) maybe followed by some kind of gate to turn the signal of once the signal drops below a certain threshold (to prevent noise entering the comparator)? But most likely this is similar to what you meant with

With the above design even small signals will cause the comparator to act, but that can be a downside too, i.e. noisy when not playing - hence the mention of an envelope controlled VCA in original post.  You can use variations on a theme to set a threshold that helps prevent this tetchiness.

Yup, you're right on the mark.  Compression will help.  The lo-fi way is simply to drive the hell out of an input stage so that you end up with limiting, but this makes things tetchy as far as splutter goes (slightest brush of the strings triggers the comparator).  A retriggerable monostable can be used to gate the signal once the input signal falls bellow a certain threshold.  I prefer this approach with square waves - sounds more like a synth (no dynamics) and you can get a nice clean cut off.  The alternative is to use an envelope controlled VCA to track the dynamics of the input signal.

Any links for the VCA + treshold ??

Take a look at RG's article on OTAs at Geofex.  Hook up the Iabc pin to the output of an envelope follower and it'll track the input signal.  Sounds a bit artificial with a square wave to my ears, almost like the fade out at the end of a CD track - you can hear that everyone is still partying away, so to speak.