30 Oct
2010
30 Oct
'10
1:23 p.m.
On 29 Oct 2010 at 22:10, Tony Duell wrote:
Yes, I should have remembered that...
Anyother thought I had is to use a 7447 for the
lower 8 or 10
patterns (as it's designed to do!) and add logic for the higher ones.
I don't know if that saves chips.
The 7447 has a problem in that the "6" doesn't have the top crossbar,
so that it's indistinguishable from a "b". The '247 does include the
top segment when displaying '6', which is why I mentioned the 247 and
not 47.
This brings to mind an ancient "fix" for the 47 "6" display--a
pulldown diode connected between segment "e" and segment "a"--in the
Been there, done that :-)
display of 0-9 there is no time when segment
"e" is active that
segment "a" isn't also active. The converse, however isn't true--and
this "fix" will mess up your display of "b" if you use the method
described previously to display 0-F.
Of course. For 'b' you mmed the 'e' segment without the 'a'
segment.
That's what distinguishes it from '6'
But if you allow diodes, then there's no reason not to use a 4-to-16
demux and a mess of diodes to do the decoding for 0-F, is there?
while my original question didn't preclude the use of discrete
components, and while I happly ageee that the odd diodes, pull-up
resisotrs, etc can lead to interesting solutions, I do feel that such a
diode matrix is outisde the spirit of the problem. After all, you could
say it can be solved with no ICs at all, just lots of discrete
transistors, etc. :-)
-tony