Foolproofing open-collector outputs

Mar 03, 2021 Last reply: 5 years ago 25 Replies

Is there a relatively foolproof way to protect open collector outputs against over-voltage/over-current, when they're on terminal blocks that are user accessible? Designed to connect LEDs with limiting resistor, small incandescent lamps or relays up to 500mA rather like the output of the ULx2004.



But they may connect anything and would prefer in the case of most reasonable errors like shorting a collector to the supply outputs not be destroyed.


Ixys has some self-protecting SSRs that are handy for field outputs. Isolation is a bonus.

John Larkin Highland Technology, Inc The best designs are necessarily accidental.

You could include an emitter resistor and a potential divider on the base so that the o/c transistor behaves as a crude constant current sink. That would limit the power dissipation if shorted to supply.

John

Zener diode to ground and current limiting resistor to the output?

ULx2004 in a DIL socket?

Adrian C

onsdag den 3. marts 2021 kl. 18.31.46 UTC+1 skrev bitrex:

formatting link

or similar

I tested a few self-protecting mosfets that didn't. Some oscillated too. I don't think any includes reverse protection.

That plus an overvoltage clamp might be all I need, here. Might be able to protect several outputs with one chip like this:

and PFET in series + zener in parallel with the load, at the relatively low currents the on resistance of the fet shouldn't matter too much and might be able to scan all the collector voltages on a bus with an analog multiplexer, if one goes out just clamp them all is OK.

Zener in parallel with the switch transistor, rather

50 cent per channel in quantity is kind of budget-hurting

Particularly if they don't work!!!!!

I like it but not as much the price

sure, but there are lots of manufacturers of such protected switches can probably find much cheaper

if you can use SPI how about

formatting link

Perhaps these devices:

Littelfuse PGB1010603MR SESD0402X1BN-0010-098

The PGB1010603MR is something like 0.04 pF--not bad for RF ports. The SESD0402X1BN-0010-098 might be more to your liking tho.

Also, consider a shunt (to GND) 10M to 100M resistor if there will be some pull-up in the target: These days we need to protect ourselves against Corona.

There is no protection from Mexican beer!

Excellent suggestion; a lot of transistors are fine with a milliamp or ten and safe-to-handle voltages. A diode to clamp negative voltages is a minor embellishment. There are mass-produced 'biased transistors' with the potential divider onboard, if necessary.

There are 2-cent beta-binned transistors. One could perform the blasphemous act of beta limiting.

John Larkin Highland Technology, Inc The best designs are necessarily accidental.

There's no reason to avoid Modelo draft.

John Larkin Highland Technology, Inc The best designs are necessarily accidental.

Especially since poor JT isn't here to pour poorly though-out scorn and derision on the idea. ;)

I designed a little diode laser driver for a start-up that used a beta-graded PNP for short-circuit protection :

Q1

0----------* *-------* +3.3V V / | ------- V Laser | T R | Q2 R 301k \| R |--- pulse train R V| | | GND R R R | GND

The pulses were short enough that the stored charge didn't get depleted, so the PNP didn't come out of saturation in normal operation, but limited the laser current to 60% or so of the peak current in fault conditions.

Cheers

Phil Hobbs

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required