[sdiy] Newbie alert :-)

[Nicolai Czempin]

> Note: a long post follows but it does contain a short answer.
> 
> 
> > What happens if I just use the default, e. g. have you received this
> > message twice? I'm pretty sure that it is a listserver setting. Was
> the
> > decision to make replies to-the-list conscious or accidental?
> 
> Yes, I got it twice. I used "Reply to All" and you can see the
> addressing on this message. You will also receive this twice.
> 
> The reply-to topic has been discussed in the past. I think the short
> answer is: "We have to live with it."

okay. I keep forgetting, and send stuff that's meant for the list privately,
but I guess I can learn :-)

> 
> 
> > >Ray Wilson's "Schematic Publisher" ...
> > >
> > Just looking at the screenshots, this seems pretty cool. Now if only
> it
> > were open-source. ...
> 
> You'd have to talk to Ray about that. He's great at answering e-mail
> questions.

I don't think I'll ask anybody to open his source if he hasn't got the idea
himself.

> 
> 
> > >Ray's website has a lot of synth circuit designs...
> >
> > Yes, looks nice, but I'm wondering whether (watch out, more
> stirring!)
> > people are really reading what I'm writing. I want something simple
> to
> > get started! All these fascinating links I will save and look into
> at a
> > stage when I'm a little more advanced. But for now I just want THE
> > SIMPLEST THING THAT COULD POSSIBLY WORK (tm).
> 
> I'm wondering if you are reading what people are writing in return.
> <G> As far as I can tell, your question has been answered. It depends
> on how you define "work."

You're right. In the "iterative and incremental" world (and "test-driven
development", you try to get something simple done that will fulfil exactly
the minimum requirements, which for an oscillator is to oscillate :-)

> People on SDIY are building synthesizers for
> musical use, so there are certain assumptions made.

And once you have something simple working, you will know better what is
missing from the minimal version (in this case, "certain assumptions" that
only people can be aware of that have seen/heard the shortcomings of the
minimal solutions.
Thanks for pointing out the hidden assumptions for me; I still have a few
comments.

> 
> Circuits that we care about are usually voltage controlled. Power
> supplies are typically +/-12 or +/-15 volts, with some exceptions
> (Ray's Sound Lab is +/-9V). Signals (audio and CV) are usually +/-5 or
> 10 volts. 

Hmm. This strikes me as a result of design (for example, using certain
opamps?) rather than an external requirement.

> VCOs tend to have exponential response (1V per octave),

This is also a result of design (or, a "second-order requirement"), although
it of course makes a lot of sense in a musical context.

> though you'll find linear VCOs, too. Temperature-controlled VCOs are
> common, since they are more useful for making music, but you'll see
> that the "tempco" stuff is optional.

Wow, how are temperature-controlled VCOs useful for making music?
Auto-compensating for temperature drift? Or have you made a typo? I am
genuinely confused.

> 
> Therefore, the SDIY group isn't very interested in a 555-based
> oscillator with frequency controlled only by a pot, except perhaps as
> a simple LFO.

Well, I'm not sure about the "demographics" of the sdiy group. Am I the only
beginner around? Because to someone like me, just getting a sound of a few
connected components is a success.
Of course, I want to advance to "the real stuff" as soon as I can, but I do
want to understand the musical reasons that make a 555-based circuit
infeasible (my guess would be that you get a very noisy signal, at least
from my experience in hearing it, and of course a potentiometer-based
oscillator isn't really of any use at all for producing the sound. For the
other stuff the VC isn't 100 % mandatory; I remember having seen simple
synthesizers where one could simply adjust filters and envelope manually,
but of course the frequency was selected by a keyboard).

My aim is to learn why all this stuff works, perhaps somehow reliving the
history of synths, to find out what the reasons were for certain choices. I
want to understand exactly how much better a square signal sounds from an
op-amp design compared to the simple 555.

"Progress" (in the sense of getting useful instruments) will be much slower,
but if I were interested only in the end result I'd just program a digital
simulation (or, indeed, just buy one).

> Besides, you can find such circuits in the "standard"
> books like the CMOS Cookbook, IC Timer Cookbook, and IC Op Amp
> Cookbook, as well as the simple Forrest Mimms books that Radio Shack
> used to sell (in the US, anyway).

Sure, I have some books like that, although not the ones you mentioned. Part
of the point of my post was possibly to get recommendations on books.
Perhaps there is an FAQ posted somewhere that I missed? Or if not, perhaps
one should be started.

> 
> So, when you ask for a VCO, people can't help but direct you to one
> that isn't a "dead end" design.

Well, I'm grateful for everybody's concern, but I'd actually like to run
into a few dead ends. If I were doing this for some commercial reason, I'd
be using a totally different approach. But since I'm doing it just for the
fun of learning, please feel free to point out things like "here's the
simplest kind of VCO imaginable, but you'll want feature A, which makes it
look like *this*, and then probably feature B, which then makes it look like
*that*.
I'll certainly try as much as I can to find all that information from
literature, and I'll document it all, so the next beginner who wants to walk
down that road has something to start with. And of course since I will
document the reasons why this or that circuit is inferior (and support the
reasoning with sound examples), he can choose to skip a few of the
"dead-end" circuits that I want to build.

I certainly won't make PCBs out of the dead-ends :-)
> 
> HERE IS THE SHORT ANSWER TO YOUR VCO QUESTION:
> Look again at the link to Rene Schmitz that was provided by Scott
> Gravenhorst:
> http://www.uni-bonn.de/~uzs159/vco4069.html
> It uses the cheap CD4069UB (must be the unbuffered UB version, as I
> recall). You can omit the "expo current sink" stuff on the left and
> you can omit the pulse/pw/pwm stuff at the bottom. This leaves you
> with a sawtooth VCO that has linear CV response. By my quick count, it
> uses half of a CD4069, 2 diodes, 2 caps, and 4 resistors. It doesn't
> get much simpler than that! :-)

Great, I'll try that!

> 
> After building the linear sawtooth osc, you can add the pulse
> converter part. Later, you can build the expo converter. Good enough?
> Simple enough? I hope so.

Sure, right up my alley.

> 
> 
> Oh, you also asked about a keyboard. From what I've seen, building a
> resistor ladder (matrix? whatever...) keyboard -- that is, an "old
> style" CV keyboard -- is a lot of work. Once again, Ray Wilson has
> done this and you can find the plans on his website. Personally, I'd
> just as soon find a cheap synth with an analog output *or* use a cheap
> MIDI keyboard with a MIDI-to-CV converter (the one from Paia.com is
> probably the cheapest and it's not too bad).
> 
> 
> One more thing: We tend not to reinvent the wheel, so a few websites
> will be mentioned time after time on this list. 

Sure, that's a perfectly sensible approach, and one that I use in my work
all the time. I try to avoid ever to write any code that's been written
before. But by now everyone will realize that for building a synth as a
hobby, I'm currently taking a different approach. I'm sure my tempo will
pick up once I feel a little more at home in this strange world where things
actually break when you drop them :-)

> That's because some
> folks have done a tremendous job of finding and creating various
> circuits. Oddly enough, we repeatedly use those which are the most
> sensible. ;-) You don't want to bother with a circuit that relies on
> hard to find parts, for example. You'll often be directed to the ASM-1
> stuff (mostly at Magnus Danielson's site, which is *full* of synth
> stuff) and the work/findings of guys like Gene Stopp (the ASM-1), Ian
> Fritz, Rene Schmitz, Ray Wilson, Jeurgen Haible, Charlie Lamm, and
> others. Harry Bissell has published a few articles, too (excellent
> guide to caps, the nifty MorphLag, and a great envelope follower).
> Apologies to those I've forgot to mention.
> --
> john
> 

Thanks very much, this was exactly the kind of answer I was looking for.

mfg Nicolai

-- 
GMX ProMail mit bestem Virenschutz http://www.gmx.net/de/go/mail
+++ Empfehlung der Redaktion +++ Internet Professionell 10/04 +++