1 Nov
2010
1 Nov
'10
1:02 p.m.
On 2010 Nov 1, at 10:23 AM, Dan Roganti wrote:
On Mon, Nov 1, 2010 at 3:10 PM, Brent Hilpert
<hilpert at cs.ubc.ca>
wrote:
We must have different specs at hand (different manufacturers perhaps):
Fairchild'76 says 270 ohms typ, 1000 max at 5V/25C, Harris'92 says
470-1050. Even if you account for the R in the LED drop R, part of the
problem is the variability between devices, even though intra-device
variability is low. Based on the specs I have power dissipation may
also be an issue once you're pushing current through that internal R
(100mW max per switch).
Neat suggestion to modify the existing A-F displays, but once you
have the
'138 and a bunch of diodes you could just revert to the technique of
generating 8-F, and get by with 2 chips (247,138), 1 inverter (to
disable
the 247 for >=8), and diodes.
The use of the 4066 does present a problem for direct drive of the
LEDs: ON
resistance of the transmission gates is on the order of hundreds of
ohms
(with considerable variability over devices and temp).
oh yea, that would make it only 3chips - we have a winner !
Although the utliz.of the inverter chip is very low.
You could always just use a transistor.
I just didn't want to leave the '247 abandoned ;)
The on resistance of the 4066 (~80ohm) is way less than the
4016(~200ohm) -
which is why I chose that one - but marginally higher when operating a
the
lower 5volts. But you could still compensate by lowering the LED
resister to
keep it just as bright. I never rule out CMOS :)