Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!cs.utexas.edu!uunet!snorkelwacker!ai-lab!mikec
From: mikec@wheaties.ai.mit.edu (Mike E. Ciholas)
Newsgroups: comp.sys.mac.hardware
Subject: SID Board: UPDATE!!!
Message-ID: <8223@rice-chex.ai.mit.edu>
Date: 3 May 90 00:28:21 GMT
Reply-To: mikec@rice-chex.ai.mit.edu (Mike E. Ciholas)
Organization: MIT Artificial Intelligence Laboratory
Lines: 64

(SID is a sound digitizer for any Mac.  Download info-mac/sound/program/sid.hqx
on sumex to see what its about).

UPDATE:
	I've gotten about 70 responses.  Getting close...  If you want to
	make sure it goes, better respond and tell me so.

DESIGN:

How close is close enough?

Due to the very limited availability of 1.558Mhz resonators, I have been
searching exhaustively for a replacement circuit.  Criteria are: cheap, 
avaiable parts, and close frequency.  But how close is close enough?

Here are the following choices:

1. 800KHz resonator with a divide by 2 followed by an 18 state conversion.
	- conversion frequency is 0.145% low
	- virtually no parts changes (just the resonator)
	- parts: $1.00

2. 1.000MHz crystal with a divide by 3 followed by a 15 state conversion.
	- conversion frequency is 0.145% low
	- more stable than resonator
	- minor parts changes
	- parts: $4.00

3. 6.144MHz crystal with a divide by 23 followed by a 12 state conversion.
	- conversion frequency is 0.028% high
	- two extra chips
	- longer assembly
	- parts: $6.00

4. 22.1184MHz crystal with a divide by 71 followed by a 14 state conversion.
	- conversion frequency is 0.012% low
	- two extra chips
	- longer assembly
	- more power
	- parts: $9.00

Some additional data:

Resonators vary in tolerance from 0.05% to 0.5%.  I don't know how stable
they are.  Somewhere I got the idea that the 1.558MHz resonator from muRata
Erie was 0.5% tolerance, which means any of the above circuits are better.
Even if it is 0.05%, 3 and 4 are better initial tolerance.

The difference in frequency between 'C' and 'C#' (sharp) is about 5.5%.  If
I went with the first circuit, could you hear the 1/38th interval?

I may be possible to "pull" the resonator frequency slightly.  I'll see if I
can find a way to do this without an adjustment (it is usually done with
some sort of trimmer capacitor).

My personnal favorite is #1.  People who have already bought parts can still
use it, the freq difference is minimal, and it is cheap.

Mike Ciholas

email:  mikec@ai.mit.edu
snail:  289 Highland Ave, #108/Somerville, MA 02144
phone:  (617) 623-3563 
air:    N1909C, 1954 Cessna 170B