Arcomatic Rosinator - bowed string effect

[anotherjim]

I'm a little confused by the schematic so here is a noobish question. What is Vcc as it seems to just be +9v and Agnd and Vee seem to be the same am I missing something?

Ah well, I'm not used to Eagle. I used it for this schematic as it seems to be the most common tool to generate a pcb layout - although I don't run to making pcb's myself. Anyway, Eagle's LM324 schematic doesn't seem to have supply pins. Vcc and Vee happen to be the supply designations of a bipolar chip. So Vcc = +9v (pin 4) and Vee = 0v (pin 11). Agnd (analog ground =  0v) is my convention because if I have something with BBD or any digital stuff, I'd have a Dgnd too.

I left the footswitch wiring off the scheme. This is entirely conventional TBP on my prototype. I haven't bothered to fit an indicator LED.

A few things that have occurred to me since...

The Zener is probably actually 3v6, and as I suggested, 2 led's in series do give equal performance (tested).
2x  LM358 dual are same as 1x LM324 quad if you prefer them.
I did try a cheap  CMOS rail-rail amp - the LMC662 dual (LMC660 is quad). It works well. This type isn't usually found in audio projects, but because it's not in the output signal path or driving a screened cable, it's good enough here as a low power alternative (also good for LFO work).

Zool, I'll have a go at checking your layout.

Again, thanks for all the interest & comments from all.

Jim

[Ben N]

Cool stuff! To some extent, this is the territory guitar players have tried to cover with their volume pedals, but the sawtooth+envelope just ups the ante considerably.

[anotherjim]

Better post gutshots of my build - I know you like 'em warts & all.

http://s1374.photobucket.com/user/ashdalestudio/slideshow/Build%20photos

Jim

[Luke51411]

Ah well, I'm not used to Eagle. I used it for this schematic as it seems to be the most common tool to generate a pcb layout - although I don't run to making pcb's myself. Anyway, Eagle's LM324 schematic doesn't seem to have supply pins. Vcc and Vee happen to be the supply designations of a bipolar chip. So Vcc = +9v (pin 4) and Vee = 0v (pin 11). Agnd (analog ground =  0v) is my convention because if I have something with BBD or any digital stuff, I'd have a Dgnd too.

Ok, I figured as much just wanted to make sure before I start breadboarding. Thanks for the clarification.

[alparent]

I'm at work so I can't do much now, but I will update the files you sent me and I will send you tips on Eagle (like how to invoke the power pins of an IC.)

[anotherjim]

I've tried to make a veroboard layout for this, but I'm not too experienced in drawing layouts from schematics. So if someone could verify it, that would be great 

I've checked against my schematic.
Very good. You've swapped the amps which had me puzzled until I realized.

I found the following...

IC1 link between pins 9 & 10 should be 8 & 9
R16 not numbered (it's the only stand up one).
D3 polarity not showing.

...and that's it! Well done

Mind you, there was method in how I've picked IC pin outs in the schematic. It's my habit to put the highest impedance or highest gain circuit on the grounded side of the chip so as to reduce the chance of noise pickup from the + supply. Since there's no LFO or high current stuff in this circuit it's probably an unnecessary precaution.

[duck_arse]

ahh, geeze, you KNOW how much I like looking at warts. I caught a brief glimpse of some, on vero, then photobucket took over and started helping me look, screen went blank, waiting, page closed. oh, well.

find the "invoke" button in eagle, and it will show you what's your IC still has hidden.

[Zool]

I've tried to make a veroboard layout for this, but I'm not too experienced in drawing layouts from schematics. So if someone could verify it, that would be great 

I've checked against my schematic.
Very good. You've swapped the amps which had me puzzled until I realized.

I found the following...

IC1 link between pins 9 & 10 should be 8 & 9
R16 not numbered (it's the only stand up one).
D3 polarity not showing.

...and that's it! Well done

Mind you, there was method in how I've picked IC pin outs in the schematic. It's my habit to put the highest impedance or highest gain circuit on the grounded side of the chip so as to reduce the chance of noise pickup from the + supply. Since there's no LFO or high current stuff in this circuit it's probably an unnecessary precaution.

Thank you for checking the layout! Yeah I swapped the amps around, I didn't know about these things you mention about the noise and grounded side of chip, but I will keep that in mind  

I've corrected the things you mentioned, and here is the updated layout (D3 polarity dissapeared because I turned the opacity down to let the trace cut seen under it, now I made the color of the stripe to green, so it's visible now). I've also changed the Zener to 3V6 as you mentioned in another post.

[Zool]

Btw I'm looking forward to build this, because I really like slow attack effects Unfortunately I don't have some parts around, so I had to order them, and now it is waiting time for them to arrive...

[anotherjim]

ahh, geeze, you KNOW how much I like looking at warts. I caught a brief glimpse of some, on vero, then photobucket took over and started helping me look, screen went blank, waiting, page closed. oh, well.

find the "invoke" button in eagle, and it will show you what's your IC still has hidden.

Photobucket slideshow has a sense of humour - if you pause while it's doing its transition you get all black.

Invoke? Ok, so long as it doesn't involve calling up a Demon.

[midwayfair]

Great work! It sounds amazing.

[deadastronaut]

right click on opamp...'invoke' then you will have summoned the demon + - 

[anotherjim]

Ok many thanks, I get the Invoke thing. I did get a knock on my door. It was Mr Stopholese or some such name offering me the deal of a lifetime. I told him to go and Faust himself.

Working on the breadboard now is the next development.
An octave down from a 4013 and one opamp reused as an output filter/buffer. Output filter is necessary if it isn't used with a guitar or bass cab - too fizzy played straight into a mixer. Might split the amps into two dual ones so the high gain input & output filter have a better quality chip. Trying a TL072 at the moment.

[anotherjim]

A question I have...

Would the octave down be better "invoked" with a footswitch or on a blend/mix knob?
I'm about ready to build a V2 (I won't aim it at London) which will be in a larger Eddystone box (95mm wide).

[bluesdevil]

Looking forward to "version 2"! I would probably just go with a blend control for the octave down, myself.

[anotherjim]

The bad news is that blending octaves either sounds like an octaver effect or a dual osc synth tuned to 3rds. Nice, but not appropriate.
So I think I'll just go with a stomp switch to change between 1f and f/2. The good news is that it has confirmed some ideas I've been having for a VCO-less synth effect.

I'll have to get on and order some PT2399 delay chips.

Now, a Solina and other String Synths had a BBD based "ensemble" chorus. There were 3x BBD's in parallel. A sine LFO was phase shifted twice by 120° so for each of the delays the chorus sweep was spread around. The main "Chorale" LFO speed was fixed at about 0.6hz. A second "Vibrato" LFO at about 6hz is mixed with the Choral LFO. The result is lush and spooky - even in mono. With the chorus switched off, the instrument sounded like either a Harmonium or an Accordion depending on the Envelope controls.

I'm hoping I might get near enough just using 2 delays. One modulated at Chorale rate and the other at Vibrato. In series or parallel - I'm not sure which.

[Mark Hammer]

1) Sounds excellent.  How much of the grind we hear in the samples is from the circuit alone, and how much stems from some form of pre-distortion feeding the circuit?

2) Slow attack with quickly-played notes is kind of an oxymoron.  I have yet to play anything non-digital that adapts the attack time to inter-pluck interval.  It is whatitis, and the user has to always adapt their picking style to the device.  Tis like that for any analog "swell" device.  That said, I hear no shortcomings whatsoever in this circuit so far.

3) The samples are essentially continuous sound,so we don't getto hear how notes end.  So, what's the decay/release like?

4) One of the things people tend to overlook is that the most harmonic content from a plucked string will occur during the initial attack - the very part that is "suppressed" by such circuits.  That's not a complaint at all.  Rather, it means if you want notes to "cut", you need to do something, like use pre-distortion, for there to be some presence to the note/tone after the initial swell is over.

[anotherjim]

Thanks Mark,
There is no pre-distortion, although until the note is well into decay there is obviously monster distortion in the first amp - but that isn't heard.

The envelope is on the "safe" side. That is, it decays as fast as it can without too having much ripple. The squaring schmitt trigger is still detecting the pitch after the envelope has shut off. I do get a slight pump up from the attack control with moderately fast playing, but it isn't as obvious as it is if you mute long enough for the envelope to fall all the way to zero.

The grind is due to wave shape. The narrower the "pip" after the square wave has been differentiated and half-wave rectified, the scratchier it gets. There isn't really much of a fundamental wave present at all initially. The fundamental pitch is only "described" by the period of the pips. Between the pips, it want to be a straight line, but AC coupling, low pass filtering, and the whole chain including our hearing, tends to supply an underlying sine wave derived from the cyclic disturbance caused by each pip. At least, this is what I think is going on.
That said, you have to keep applying some finger vibrato to maintain the illusion of a string player, which is an advantage over the string synths which didn't allow any.

At heart it's an extreme distortion and doesn't like chords. A staple of string sections - arpeggio, is very difficult without some fancy finger work to mute the last note. You can trill and hammer on though, so arpeggios along one string  are do-able.

[Mark Hammer]

Thanks for the followup info, Jim.

Your samples are played monophonically - one presumes, in order for the finger vibrato to work its magic - but I gather the circuit itself is perfectly capable of handling a chord.

Your comments about maintaining the illusion of a string player got me thinking.  A phase-shift vibrato, with foot-controlled modulation intensity, would be handy for mimicking the "vibrato creep" of bowed strings, more than one note at a time.  I mean, I suppose one could attempt to do the same thing via the vibrato arm of the guitar, but some of them are difficult enough to work in "nuanced" fashion, as it is.  Having the intensity/width of the modulation fade in smoothly would be hard work.  Easier to just let your foot do the work.

[anotherjim]

Oh no it's rubbish at chords, like most circuits that slam your signal into a square wave. More than one note at a time and the strongest one wins after some chaotic coming and going until it sorts itself out.
I'd thought of just high pass filtering a clean signal, but that wouldn't create the narrow pip that you get from starting with a square wave, and rectifying isn't kind to chords either.

I definitely think of it as monophonic. If you want chords you need a string section - build several!

I have just experimented with adding a smooth bass tone. The f/2 coming from a 4013 swings the whole 9volts so it only needs passive R-C low pass to get a decent bass voice which doesn't change the apparent timbre of the bowed voice. It doesn't need to go via the VCA. Obvious really, but I'll only add it if there's space left after working out the chorus part.