NE 566 + 8038

[gstopp at fibermux.com]

     Regarding the 566 and 8038 - 
     
     I think that the issue should be clarified regarding the idea that 
     "these chips shouldn't be used for musical purposes". I know what the 
     intention behind this statement is, and I agree with it, but it 
     doesn't say it all...
     
     See you can use anything for musical purposes. You can use toy pianos 
     and Speak 'N Spells. You can make a keyboard synthesizer with linear 
     VCOs. It may be "cheesy" compared to something like an ARP 2600 or a 
     Minimoog, but hey - that may be an advantage! If it's used for a solo 
     in a song and it makes listeners go "what the heck is that - it's 
     cool!" then you've made a musical statement.
     
     If you want to build a VCO yourself, and you want it to oscillate, 
     these chips are probably a fun thing to try. In fact I'd like to get 
     an 8038 after all these years and mess with it. But - if you want to 
     make a wide-range, high-precision waveform, exponential response, 
     accurate-keyboard-tracking synthesizer VCO, ya just can't do it with 
     these chips. They weren't designed with synthesizers in mind. Although 
     the "VCO-on-a-single-chip" concept is attractive, there are only two 
     one-chip options that meet these strict and highly specific functional 
     requirements - an SEM VCO chip, or a Curtis VCO chip. In fact the SEM 
     chip needs an external op-amp or two anyway. If you can't get either 
     one of these, you *must* build a discrete VCO. I know that these are 
     good circuits that are simple:
     
     1. Rhodes Chroma-style VCO using a CA3046 and and EXAR 4151
     2. Electronotes VCO using an NPN pair, CA3080, and LM311
     3. Electronotes VCO using a PNP pair and a couple CA3080s
     
     Each one of these uses three to four chips plus a few misc. 
     components, and have great specs for musical intervals, rivalling the 
     best modular systems (including the fabled Serge). The first two 
     generate sawtooth waveforms, and the third generates a triangle and a 
     square. Adding other waveforms is fairly easy.
     
     So - it depends on what your expectations are. The 566 and 8038 may be 
     great for non-tracking uses like drum note synthesis, so you can have 
     a whole bunch in a single project. Note that the PAIA Stringz 'N 
     Thingz chorus module uses four 566s - two for the high-frequency delay 
     line clocks, and two for the LFOs to modulate the delays. That's a 
     good "musical purpose" I think.
     
     I do have a schematic for an exponential 8038-based VCO, from 
     Electronotes, but according to the author it gets really hot and needs 
     a while to drift into proper operation, as I recall. They're not too 
     keen on it, they kind of go "oh that's nice" and then go on talking 
     about their wonderful discrete stuff. I can dig it up if you want.
     
     - Gene
     gstopp at fibermux.com


______________________________ Reply Separator _________________________________
Subject: Re: NE 566 + have you checked out the 8038
Author:  MACHINE MEDIA <cdmaster at netcom.com> at ccrelayout
Date:    5/16/96 2:56 PM

> One chip I am currently planning with is the 8038 Function Generator chip.
> This is a lot more versatile, with Sine, Square and Triangle outs. There is a 
> Sine wave adjust pin, and a Duty cycle adjust (which works for all waveforms) 
> enabling Ramp both ways.