Alt-BEAM Archive
Message #08642
To: "'beam@sgiblab.sgi.com'" beam@sgiblab.sgi.com
From: Wilf Rigter Wilf.Rigter@powertech.bc.ca
Date: Wed, 29 Dec 1999 10:48:56 -0800
Subject: [alt-beam] SONIC NOTES
The phone rang three times in the dark and the sonic circuit chirped (wished
it would turn the light on instead)! Wrong number? Back to sleep and sweet
silence ...... ...... ...... Umm, three times? Shouldn't that be six?
Well, earlier in the night there was the cat knocking something over and the
clock rang 1 o'clock, that's two: So what happened to the missing pulse....
Arggh! I was obviously suffering delusions of accuracy!
Morning and still curious to see how stable the accumulated charge on
capacitor CA is but you can't measure it with a digital voltmeter cause that
drains the capacitor charge through the 10M input resistor of the meter. So
hook up a CMOS opamp as a voltage follower/buffer and measure the output
without affecting the capacitor charge. The circuit only drifted one bit out
of six over a period of about 1 hour so this analog counter may have a
useful "short term" memory, at least with the tantalum cap and at room
temperature, but forget about long term memory. For that you would need a
analog circuit that can ramp up but has stable "notches" which would make it
multilevel analog logic right? (mental note to design this later). In the
mean time, it appears that the analog counter can accumulate errors over
longer time spans especially with higher divider ratios. One solution is to
use this analog counter only for application which require shorter term
memory (ie <1HR) and add a reset function, either a feedback resistor to
deliberately "forget" or to "auto-increment" counts over long times or some
logic that senses a "long" beep as a reset (or synchronizing) signal.
Hey, adding that opamp made this analog counter into a kind D/A converter.
Every sound pulse increments the analog output voltage by a small step. Now
what could that be used for? Change a servo position?
cont'd
wilf
Home