If you're intending to assemble all your boards the same, I'd go for a crude current limiter using a transistor, a couple of diodes and a resistor in series with the opto diode, plus a modest series resistor to drop a certain amount of voltage and remove some heat from the current limiter.Thanks for the input. I'm looking to have inputs from 5 - 24V so of course you want a margin of error there, I would be using an industrial power supply in situations where I want 24V tolerant input. I do like optocouples I used them before for inputs and outputs that drove relays. Would it make sense to take the zener setup and use that to drive the optocouples? In the past I just used a resister network seemed to work but that was designed for 12VDC only.
If you intend to populate each board for the specific application, then a series resistor would be fine; no need for the zener. But bear in mind that at 24V input even a 2K2 resistor is dissipating 0.25W; I generally use 4K7 for a 5mA "wetting" current.
And make provision for reverse voltage connection in some way.
If your inputs will mostly be contacts, you could have a link to connect the +ve voltage to one leg of the opto circuit.
(Bear in mind that you can always link out unwanted components).
A Darlington-connected transistor always drops more voltage than a single transistor - can exceed 1 volt. Look at the Toshiba TBD62083A series as a MOSFET example - probably a few hundred mV at the sort of currents you're likely to use.The Output ULN2003 are for sinking current I wouldn't expect a voltage drop? The purpose would be to drive relays to switch any sort of load. The MOSFET version sounds interesting I was looking for sourcing versions since.
Incidentally, don't forget the ULN2803 and equivalents - 8 drivers is potentially tidier than 7.
Statistics: Posted by stevend — Thu Feb 01, 2024 1:51 pm