EDN inrush current limiter

I guess you might have to post them to edn.

There was some talk (IIRC) of using a relay to short a current limiting resistor in James A's input protection thread.

George H.

I'm sure it works good if done correctly - as a commenter mentions in this one depending on whatever time constants there are in "auxiliary power supply" etc. there's no guarantee that the relay will disengage in time to protect from a rapid disconnect-reconnect situation.

Seems like anything more complex than an NTC resistor is more non-trivial than it appears at first glance; probably why they make ICs with embedded Z80 processors and a bunch of logic and stuff to manage the switch timing in critical applications.

The "right answer" is probably "if you can accept the disadvantages of a ~10 ohm NTC resistor then accept it. If you can't then get the fancy chip"

The simple NTC resistor, because it stays hot during a short shutdown, allows full inrush current on subsequent startup. The relay bypass on that resistor is intended to correct that (among other things). Some nontrivial issues occur with the simple NTC resistor.

Is there a reference design for a version using a NTC resistor/relay that's bug-free, somewhere?

How about turning on the relay after some short delay (10's of ms?) after power up? That would take care of short time cycling of the input AC.

(Oh and you'd have to make sure the relay opens when power is lost..)

Almost, it still has hot resistance so inrush is reduced, but of course by a lot less.

IME NTCs tend to degrade over time, hot R and local theta go up as it oxidises. I wouldn't want to use them in hi-rel stuff. And as pointed out, your circuit still needs to survive a momentary disconnect.

NT

you'd need to use a SS relay otherwise a half-out mains plug would defeat the inrush limiting.

NT

I think bitrex _should_ submit his "circuits" to EDN. We're always in need of a laugh around here >:-} ...Jim Thompson

Surviving a momentary disconnect is good. The control console for a ship propulsion system I worked on a couple decades ago used a pair of $20k high shock rated plasma displays, also used in the M1 Abrams tank at the time, which dealt with the momentary disconnect (from battery) issue with a label on the front warning not to turn back on for (unremembered time interval) after turning off. Someone wondered what would happen if you did that, and found out that the display stopped working, no external accessible fuse, totally sealed requiring special tools, return to factory for $4k repair. We put in a time delay relay on the power line to insure minimum off time after that, in a box with nameplate reading "(name of idiot who fried display) failsafe device" Too bad we didn't have the name of the idiot who designed the inrush limiting circuit, which was obviously a simple NTC.

Ouch. It might have not been an idiot, but management saying, we don't have time for you to figure that out... put a warning label on it.

George H.

Jim Thompson doesn't find his own posts as comically ill-informed as many people here do. He may well be more in need of a laugh than the rest of us.

And people laugh when I tell them that reading comprehension is the skill most predictive of employment success in the modern workforce, way more than one's mathematical or technical aptitude.