Sv: [sdiy] Digital VCO update

[Michael Bacich]

On Apr 30, 2006, at 12:17 PM, Eric Brombaugh wrote:

> Due to limited resources an LFO will have to be a separate design.  
> It's come up often enough that I'd be interested in what folks  
> would like to see in an LFO. A few questions:
>
> * How stable does it have to be? (Is 2.5% enough?)

Are you talking about frequency jitter, or some other kind of  
stability?  I don't know how accurate an LFO really has to be in  
terms of frequency -- especially if you have a sync input so you can  
crack the whip on it with an external time reference.

> * Is 1V/Oct necessary, or is linear OK?

Most cheap and crappy LFO's have linear response, but only because  
it's cheap to do it that way, not because it's any more musically  
useful.  In fact, with 1v.octave response, it's possible to do very  
musically useful things such as using CV octave shifts to exactly  
double or halve the LFO rate.  Also, it makes the response of the LFO  
more uniform and predictable over the entire range on voltage  
control, eliminating the need for control pots with fancy log or  
reverse log curves.  With 1v/oct response, you just use a linear pot,  
and you get the same amount of apparent frequency change for the same  
amount of pot travel, at any point in the entire range of the pot.


> * I assume that hard sync would be useful. True?

Absolutely!  This will allow rhythmically synced LFO effects

> * Frequency range?

Hmmm.  Some may say that your LFO should be able to work up into the  
audio range, for cool FM effects.  While I think that would be nice,  
I think the overall range of the LFO is less important than the  
frequency control resolution that it allows inside its working  
range.  This is important especially in the slower ranges, where you  
may be trying exactly match a specific tempo.  Without enough  
resolution, you may be either too slow or too fast, and not be able  
to hit the sweet spot.  Obviously, very slow rates are always fun,  
but I worry that the waveform stepping may start to become unbearable  
at extreme settings -- only some testing will tell, right?

I think you were claiming that your current design had about a ten  
octave range, right? (20KHz to 20Hz).  Does this wide range come at  
any expense of frequency control resolution?  Or to put it another  
way, would you be able to get better control resolution if you were  
to cut the overall range, for instance, in half?

Anyway, with 1v/octave response, and the same ten octave control  
range as you have in your audio VCDO, if you set your upper frequency  
limit at 200 Hz, then your slowest speed would be roughly 0.2 Hz, or  
about 5 seconds.  That would put the typical "sweet spot" for vibrato  
(about 20-30 Hz) at a control voltage of around 7 volts (if you're  
working with a 0 to 10 volt control range -- or 0.7 of your total  
control voltage range if you're working with a different range).  If  
you raise the upper limit to 400 Hz, that would put the sweet 30 Hz  
vibrato range at a control voltage of about 6 volts -- it would also  
raise your lower frequency limit to about 0.4 Hz, or about 2.5  
seconds for one cycle.

Again, how useful this thing is going to be at the slow ranges is  
going to be totally dependent upon how much stepping is audible in  
the output waveform.  And I don't think any kind of output filtering  
is going to be very useful on this.  You couldn't even use a very  
accurate voltage-controlled slew limiter whose slew rate was dynamic  
and directly proportional to the current LFO frequency to control  
stepping, because any such slew limiter would not allow sharp  
transitions at the sharp edges of saws or squares.

Speaking of which -- how sharp and clean are those sawtooth  
transitions?  Could they reliably be used as input for a typical  
sawtooth > comparator Pulse Width Modulation circuit?  You have  
already told us that voltage controlled PWM isn't possible within the  
chip in your design.  it would be nice to be able to do it outside  
the chip by traditional means.

Getting back to your question, "Frequency range?", if you're talking  
about ten octaves of range, then 0.2 Hz to 200 Hz would probably work  
for me, and would probably be similar to the kind of range we see on  
LFO's in most modern synths (Korg Triton, etc.).  Alternatively,  
maybe you could give the chip two working ranges, selectable by switch.


> * Output voltage range (+/- 5V, or 0-10V?)

Well, this is a tricky question, isn't it?  Traditionally, in analog  
synths, we like to see different overall DC output ranges for  
different LFO waveforms, and there are musical, not technical reasons  
for this.  For example, we prefer Sine and triangle waves to have  
bipolar ranges (+/- 5V).  This gives your vibratos a more natural  
sound (it's how we vibrato when we sing, or when we vibrato on a  
violin, i.e. both above and below the center pitch).  It also ensures  
that there is no apparent overall pitch shift when we add vibrato --  
since the frequency is going both up and down by the same amount, our  
ears perceive the overall center pitch as not changing.  If the sine  
wave only went positive, you would hear a vibrato as also shifting  
the overall note pitch sharp.  (this, BTW, is what gives that  
excessive-vibrato heavy metal guitar vibrato that we hear so overused  
by guitar shredders its characteristic fingernails-on-the-chalkboard  
quality.  Since you can only bend a guitar string's pitch *up*,  
whenever you do finger vibrato on a guitar, the overall net effect is  
to raise the perceived pitch of the note.  Thus, all finger-vibrato'd  
guitar notes sound sharp.  Do we really want our synthesizers to  
sound like Yngvie Malmsteen?  No, we do not.  Therefore, Sine waves  
and Triangle waves must be bipolar.

However, it is also useful to have our square waves go only positive  
(0-10V).  Positive-going sawteeth and squares tend to be more  
musically useful (though not *always*, just most of the time).   
Particularly in the case of the square wave, since you can easily use  
a positive-only square wave to create nice musical "trill" effects  
that trill between the base pitch and some other musical interval  
(say, a fifth or an octave).  If your square went both positive and  
negative, that would make it MUCH harder to dial in the two pitches  
that you wanted in your trill, since raising the LFO's amplitude  
would change both the upper and lower frequencies of the trill.  You  
would have to not only find the correct amplitude to get the fifth or  
the octave, but you would also have to re-tune the pitch of your  
target oscillator to bring it back to the original center pitch.  i  
hope I'm explaining this clearly enough.  If not, please ask and I'll  
try to clarify.

With the case of the sawtooth, I think it usually just sounds better  
if the a positive-going sawtooth (ramp traveling upwards) goes  
positive only, and a negative-going sawtooth (ramp traveling  
downwards) goes negative only.  In your case I guess that would mean  
0 to +5V for the positive ramp, and 0 to -5V for the negative ramp.   
This is how it's done in many traditional modular synths (and on VC 
( 3 of the Minimoog, which is its LFO).  I think this leads to very  
logical and musically useful results.

If, for some technical reasons, you can't have different ranges for  
your different waveforms, and must choose just one range, then  
definitely use the +/- 5V range.  It's essential that your sine and  
triangle are bipolar.  If you have to choose just one range, but you  
still have a couple of spare output pins, you could use those pins to  
control external analog switches that could be used to do the  
necessary DC level shifting you might want for the sawteeth and  
square waves (two pins would allow up to four different selections).


> * Waveforms?

I'm a traditionalist, I guess.  Sine, Triangle, Square, Saw Up, Saw  
Down.  Those are the ones that will be used 90% of the time.  I  
wouldn't waste time including a half-ass psuedo-random "sample &  
hold" wave, nor would I include any of those fake "noise" waves.   
Hopefully, your sawtooth will be clean enough to use to make a PWM  
wave with external means, and your square could easily be used to  
clock an extrenal *real* sample and hold circuit.  Leave the weird  
stuff to the 2% of the population who thinks that a Buchla synth is a  
"meat and potatoes" instrument.  I would like to see you just build  
the best traditional LFO possible.

> * Other features?

Two things:

1.   A sync output pulse would be very handy.  One pulse per cycle.   
Or better yet, selectable numbers of pulses per cycle (1, 2, 4, 6, 8,  
12, etc.)  These could be used for rhythmic syncing of external  
events.  Alternatively, you could simply have a 24 pulses per cycle  
output pulse, then it could be used as a DIN sync signal, or could be  
easily divided down into musically-useful rhythmic units (quarters,  
eighths, triplets, etc.)

2.  MIDI sync -- sync to external MIDI clock..  Yeah, I know I'm  
pushing it here.  Whatever you do end up making will be cool, and I'm  
glad you're working on this.

MIke B.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://synth-diy.org/pipermail/synth-diy/attachments/20060430/f7500169/attachment.htm>