Conference napalm::commusic_v1

    
    Well, having a natural bent away from compression in any form (yeah,
    that's the classical-music influence in me), I haven't _built_ any
    such thing... but I can imagine how one _might_ be built.
    
    [now, I'm thinking while I'm typing, so don't expect this to work right
    off the bat...]
    
    
    We need some sort of device whose resistance we can control (no, I
    don't want to attempt to design a current mirror or transconductance 
    system; I know they can be done but I'm not the one to do it).
    We can either go with a CMOS bipolar switch or we can go with 
    an analog [augh] device like a CdS photoresistive cell (Radio shack
    has 'em, 10 for $2.00. )  
    
    Since the CMOS is tricky to handle (and has some interesting
    signal-sensitive nonlinearities when you don't saturate it either 
    ON or OFF) let's go with the photoresistive cell.
    
    Now, if it's a photocell, we need to give it an amount of light 
    proportional to the *conductance* we want the cell to have (conductance
    is high when resistance is low, conductance is low when resistance is
    high.)  Once we have that, we can put the CdS photocell into a 
    circuit just as though it were a resistor, and all should work.
    
    First, let's make some light.  We can do that with the old op-amp
    amplifier chip, one resistor, one potentiometer, and a low-voltage 
    light bulb (try a SMALL three-volt flashlight bulb first).  Wire it so:
    
    	
    		|\          optional      v--- this is a light bulb :-)
    	 IN ----|+\        resistor
    		|  \----.------/\/\/\---(aa)-----
    		|  /    |                       |
    	    ----|-/     |                       |
    	    |	|/      \                      ---
            |           /                      ///
            |__________>\ variable pot,
                        / 100K ohms
                        \
	                |
                        |
                       ---
    		       ///
    
    I'd say use a diode and LED except that CdS photocells aren't very 
    red-sensitive; you'd need a green LED but those aren't very bright.
    Try it if you are in the mood someday, but realize that it's only
    an experiment.
    
    How this works: the IN signal is amplified by op-amp, and the gain is
    set by the variable pot.  Try to size the light bulb so that when
    the op-amp is saturated (i.e. is putting out max current- read the 
    back of the op-amp data card) then the light bulb is at max current
    too.  If the op-amp is much too powerful for your light bulb, then
    put in the "optional resistor".  I'll tell you how to size this
    resistor later, if you need it (you probably don't need it unless
    you start blowing out the flashlight bulbs)
    
    Now, take that light bulb, and wrap electrical tape over all of the
    electrical connections.  Don't wrap over the glass part.  Take a CdS
    cell, insulate the leads and back part, (so that only the front plastic
    surface is exposed) and place the light bulb right on top of it.  Tape
    the two together with electrical tape (or, if you feel very confident
    that you insulated everything  perfectly, wrap first with aluminum foil
    to get more of the lightbulb light onto the CdS cell).  Leave a 
    little hole in the wrap so you can see if the light bulb is lighting
    up.
              
    Your CdS cell is now a dynamically varying resistor.  You can plug it
    anywhere into your CQM you want to have a decreased resistance with
    incoming signal.  For instance, you can use it like this:
    
          
    
    	  10K ohms
    			|\                                                 
     IN 2 -\/\/--.------|+\                    
     		 |	|  \----.---- OUT                
     		 |	|  /    |                        
      		 |    --|-/     |                        
      		 |    |	|/      |                         
      		 |    |         |                         
          	 |    |_________| 
                 |     
                 &  CdS cell,
    		 &  with light bulb
                 |
    		---
    		///
                                         
    
    Note that I labeled "in" on the second diagram with IN 2, not just
    IN.  This is to remind you that the two IN's are different- the first
    one is the "control" input (is the signal that controls how much 
    compression) and the second is the "data" signal (the signal that
    actually gets compressed).  So, if you keep them separate (or maybe
    have two jacks and a switch that bridges the two jacks together) you
    can do other useful things than compression, like "ducking" a signal 
    under another signal.  [this is a great trick for making vocals 
    clear even over a powerful metal lead guitar- put the vocals into the 
    "control" input and the cranked-up guitar into the "data" input.  
    Then, whenever the vocals talk, the guitar is automatically (and
    nearly instantly) potted down.  Likewise, when the vocals stop, 
    the guitar is back up again in milliseconds. ]
    
	-----
    
    Enjoy- and remember - I have NOT built this; I just think it ought to
    work.  You might have to do a bit of debugging.  
    
    	-Bill

Seems to me when reading an index of EM, they mentioned a do it yourself
compressor project.  Anybody keepold back issues?

Chad

I have a PAIA dual compressor that I built from kit. If anyone wants a cheap
compressor and a "hot springs" reverb too (power supply included) get ahold 
of me. FIrst $15 takes it...


dbii

    light bulbs are too slow.  There are I think compressor limter chips on
    the market, probably based on transconductance amplifiers.
    Anyway PAIA might still have a kit <$100.  But you'd probably have to
    build a power supply or always have batteries at the ready.
    Tom

    
    Actually, it's arguable about the speed of light bulbs; you don't want
    to take of the leading edge (the attack) of most acoustic instruments.
    If you take it off, the instrument sounds "dead" or "boxed in".  A
    light bulb gives you this dynamic transient capability as a nice
    side effect.
    
    If you want *speed*, then go ahead and use green LEDs, or try a red
    LED and see if the CdS cells you got are red-sensitive.  LEDs are
    good for many MHz, so you should use a diode and a capacitor to 
    keep the LED from flashing at audio rates. 
    
    Build it, and see if it does what you want.
    
    	-Bill

    If there's an op-amp on the compressor input, then it acts sort-of as a
    pre-amp, right?  So technically, it's a compressor/sustainor of sorts
    because as the signal gets weak, the resistance goes down, and the op
    amp amplifies it more [than if the instrument was dry].  Right?
    
    I suppose this would work, too:  to allow the attack to come through,
    have a pot on the light, because most lights take a while to reach
    their brightest, especially when they're dim.  This would allow the
    attack to come through.  Or, have the light set very bright (comes on
    quickly) to muffle it.  I think I like this idea!
    
    Why do you have to resort to a physical relay?  Doesn't a transistor do
    something like that (limit a source voltage with a control voltage)?
    
    I don't know.  That's why I asked!  Thanks for your help so far.
    I like that light bulb idea, though.
    
    MikE

    
    Compressor-sustainer:  Yeah, you could make it a sustainor by
    setting up the second op-amp circuit to have lots of gain when
    the light bulb is off.  But watch out- the noise level builds
    up very fast when you start using megohm-sized resistances in
    the op-amp feedback path (and a CdS cell with the lights off
    will go up to about a THOUSAND MEGOHMS!).  You may want to set 
    a potentiometer in parallel with the CdS cell to limit the
    amount of gain that the 'sustainer' can put on.
    
    	[Transistors and voltage]
    
    Actually, NO, an "ideal transisitor" does NOT vary it's
    emitter-collector current when you change the emitter-collector
    voltage (this voltage is called Vce).  An ideal transistor passes 
    current proportional to the base current, and it doesn't matter
    if Vce is one volt or a hundred. 
    
    	Electronics lesson #2: An NPN bipolar transistor will pass
    	collector-emitter current proportional to some multiplier
    	of base-emitter current.  The multiplier factor is around 25 for
    	big power transistors, 100 for medium-sized transistors, and
    	250 for little teeny transistors.
    
    	Electronics lesson #3: An NPN bipolar transistor passes the
    	amount of current from #2, and doesn't give a rats a** what
    	the voltage is as long as it's at least 1.4 volts in the forward
    	direction, and the voltage is less than what will fry the transistor.   
    
    Real-world transistors _do_ have a slight Vce/Ice effect; that's
    usually small enough to ignore.
    
    Field-effect transistors (FETs) _do_ vary their resistance depending on
    the control voltage- but they have this other nasty effect.  They
    "switch on" to "switch off" over a fairly wide range- something
    like a couple of volts- and that voltage is relative to the "data"
    signal, not to ground!!!  So, if you built a
    compressor/limiter/sustainer using them and don't play very carefully
    (say, by using a current mirror at constant voltage) you end up with
    a compressor that compresses the + peaks and expands the - peaks !
    This doesn't do much for signal fidelity! 
    
    -----
    
    Another neat reason to use light bulbs is they Look Sooo Cool!
    
    [hey, this *is* art, isn't it? :-) ]
    
    	-Bill
    

       
       
       
       
       Craig Anderton did one of these in a projects book for guitar
       players along time ago.  His was mono, but you could build a pair
       of them and sync them by summing the error signals.  I proto'ed
       one in college so I could make 8-track tapes for the car that
       were consistently loud.  You should be able to make a pretty
       clean one, with the opamps you can get now.
       
       
       
       

I've got the Craig Anderton book mentioned in .8.  It uses an opto-isolator
instead of the light bulb/photoresistor used in .1.  The opto-isolator he uses
is a Clairex CLM-????? that is difficult to find.  He also hints that 
substituting another opto-isolator won't work.  Unfortunately, I'm a software
person and don't fully understand this electronics stuff, so I don't know if
that makes sense or not.  Perhaps one of you hardware types could enlighten
me on how "generic" opto-isolators are, and whether they would be suitable for
this application?

-Hal

P.S.  Bill, your CQM plans inspired me to get into this hardware stuff.  I'm
      going to build a mixer soon.  I want some sort of EQ in it, so I'm 
      going to have to do some experimenting first to figure out what kind
      of EQ will meet my needs (simple high/low shelving, parametric, etc.).

CLM-6000 I think it's been superseded by a newer part.

dbii

    All Electronics has it (CA): CLM-6000   $2.50
    
    Specs:  500 ohms off, 500K ohms on.
    	    2000 V isolation, forward voltage 2 Vdc    <--  meaning?
    
    This looks a lot like the light bulb/opto-resistor combo, Bill.  Could
    I use one of these?  Probably a little expensive, though.  Not easy to
    troubleshoot (can't see light).
    
    Mike

    
    Except that I think you've got it switched around (the ON resistance
    should be around 300 to 500 ohms, and the OFF resistance around 500K).
    But besides that, it looks like it might work.
    
    The forward voltage thing makes me think it may have an LED inside
    instead of a light bulb.  LEDs fry out if they get whacked with too
    much current.  Also, you can't easily TELL if you've fried the 
    LED.
    
    I'd suggest sticking to what you can see, at least at first.  Go
    visit Radio shack, get some flashlight bulbs and CdS cells.
    
    	-Bill

 >   All Electronics has it (CA): CLM-6000   $2.50
 
I can't find the darn things anywhere around here (Baltimore, MD)!!!
Could you post the phone number of All Electronics?

    I'll dig the number up and post it tonite.  Should I put it in Dealer
    Names?  I don't think so, since they do parts, not machines.  I'll put
    it here.
    
    	Haven't had time to try the feedback-compressor yet, Bill.  I leave
    for school in a couple of days.  I'll get back to you when I do from my
    Internet account.
    
    Mike

    These are the numbers, I think.  I got home at 2:30 last night and I
    didn't bring the catalog in.  So these are from 411.
    
    	Order line (this should work, ask them for the info line if the one
    I give is wrong):
    
    			800-826-5432
    
    	Info line	818-997-1806
    
    Mike

    Rep .13
    
    Hal,
    
    Digi-Key also carries the CLM-6000. I got one from them about 4 months ago
    (it's still listed in the latest catolog).                                  
    
    Digi-Key part # = clm6000
    price = $3.23
    phone # = 1-800-344-4539
    
    
    Chris (Landover MD.)
    

PAIA sells CEM (Curtis Electromusic Specialties) VCA (Voltage Controlled 
Amplifier) chips.

I've noticed that lots of devices (like automated mixing machines) use dbx
VCA chips.  No idea where to get them, though.


Assuming you have a VCA chip, what should the rest of the circuit look like ?