🚀 go-pugleaf

In article <1992Apr5.205643.1@wombat.newcastle.edu.au> eerbl@wombat.newcastle.edu.au writes:
>
> I have 4 seven segment LED displays which I would like to ... multiplex
> these things ... driving them with something like 40mA per segment.
> ... 40mA times 7 adds up to quite a bit of current.
>
>	Can you suggest any good ways to do this?
>

While we're on the topic, I'd like to remind folks of a cute trick
HP used in their old LED-readout pocket calculators, presumably to
extend battery endurance:  inductors as current-limiting elements.
During each digit-display operation, to-be-lit segments are switched
ON only long enough to allow the applied voltage to ramp the segment
currents up to a desired level, and then the drivers are switched OFF.
During the rest of the digit-display cycle, the energy built up in each
inductor drains out into the forward drops of the LED and the catch diode.
Schottky catch diodes would return 75 to 80% of the stored energy to the
LED, but using the on-chip clamp diodes of the driver IC is more convenient,
and should still put 60% of the stored energy back into the LED.  Of course,
the digit rate must be low enough to allow the currents to decay completely.

Obviously, this technique depends even more heavily on the proper behavior
of the CPU; the squeamish should fuse the LED supply (which needn't be
to be regulated; 8 to 20V would be fine), and maybe even place the segment
drive enable timing under control of a nonretriggerable oneshot.

Brian Rhodefer


normally used in muxed displays isn't too much of a concern in the
original poster's