> Most versions ideally place them 90 degrees out of
phase with one
> another. This has two advantages: [...]
> (1) is admittedly not very relevant if you're
using a real mouse and
> even vaguely modern electronics, since on even a high-res mouse
> edges are very far apart on digital logic timescales.
Did you catch the ID in EDN about 6 months ago
No. What's EDN? It doesn't look like an obvious acronym to me
offhand. But, in any case, it's unlikely to be anything I follow.
for a circuit using a single 4538 on a rotary encoder
that provides
pulses on either of two lines depending on direction?
Well, when I built my Tempest setup, I used two 74123s (or similar,
such as 74LS123 or some such - I don't recall). You can find a GIF of
the schematic for what I built in
ftp.rodents-montreal.org:/mouse/misc/quadrature-counter.gif. Chip
numbers are 74xx-series numbers, choice driven largely by what I
happened to have on hand at the time. The pin numbers at the bottom
edge are a DB25 interface to connect it to a dumb peecee parallel port.
I thought it was very clever--much better than the
usual flip-flop
decoder.
However, it does depend on the count edges not occurring too close
together (faster than the one-shots can generate pulses), which is
where my (1) is relevant.
I spent a while thinking about other alternatives and was unable to
come up with anything better that would count every edge (which was
important in my application, since the rotary shaft encoder I used had
128 cycles per rotation and I needed to use all 512 countable edges or
it wasn't as sensitive as I wanted).
It should be possible to instead use each signal as clock with the
other as enable, with various combinations of inverters, driving
edge-triggered counters. That avoids the one-shots, but requires four
different counters, which pushed it into impracticality for my purposes
("buildable with the chips at hand" was pretty close to a sine qua non
at the time, and I was lacking either '161s or '83s, I forget which).
Using dual-clock up/down counters instead would have fixed this, but
I'm pretty sure I didn't have (enough of) those on hand either. It
might be possible to use direction-and-clock up/down counters; I'd have
to think about it. I don't recall whether I didn't have them, decided
I couldn't do it, or just didn't think of it then.
/~\ The ASCII Mouse
\ / Ribbon Campaign
X Against HTML mouse at
rodents-montreal.org
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B