My son works in IT at a local college. I got this from him a few minutes ago:


Thanks for any ideas you might have.

How about those "smart" power strips? Plug a device into the master that draws enough of a load for the strip to sense. Then the switch into the slave.

When the power goes out, both lose power. But when it comes back, the slave shall remain out longer than the master. As it needs to sense the master having enough of a load.

Since he has over a hundred locations. He may need to purchase 10 different manufacturers/models for testing. Keep a log to see which one suits their needs.

Remember, these power strips needs to be programmed for the master to determine the load tolerances.

KOT

How about replacing power packs with something other than the cheap ones most devices come with. The older in line style size of a 2 smart phones seem to hold enough juice to get past a 1 second power outage. Just an idea.

Oh, or design a capacitor inline with the wall wart and device. Wired in parallel of course. Have a large enough capacity for the amperage draw of the devices to maintain the 1 to 2 second power outage.

KOT

Ya a power supply with a decent sized cap or an add in cap is the way to go. But I'm having a hard time finding a simple plug in inline one. I did come across capacitor trip devices that can trip breakers on an outage but that isn't an ideal solution either.

Wattbox.
Have it ping the device that's hanging up.
When device fails to ping, automated script can reboot that (and other) outlet(s) with whatever delay you specify.

Wattbox like Mario suggests above or a UPS. Do it right. If he works in IT, then they should provide/purchase what is necessary to keep IT working.

The cost of the wages paid for rebooting the devices outweighs the cost of the equipment needed to perform this function. That should be the argument for the requisition request.

Car audio systems often use a large 'stiffening' capacitor when the power amps draw current that's in excess of what the charging system can provide. However, it's only for short-term fluctuations in the voltage, so a battery is better. These power supplies don't use a lot of current, so a large cap might help, but someone needs to find out what size of cap is needed, it needs to be in some kind of enclosure, needs an input and output connection and it needs to be able to justify its existence by doing its job reliably and for a long time.

They need to pick their poison- there's no easy solution to this and it WILL require some hardware. Doesn't sound like they're using POE for everything- that would simplify the solution because they wouldn't need as many devices for delaying turn-on.

OTOH, it IS California.

An IT guy we've just started working with responded with this:


I love it when the need for something delicate and high-tech is met by a hammer made to smash a crude problem in another field!

What I see happening now is disbelief that anything at this price can do what's needed, and of course a few of these need to be tested.

Interesting device, it says overvoltage 3 minute disconnect, I'm not sure if under voltage you get a timed disconnect. It definitely would need to be tested to see if it handles the problem.

Hmmm. Also there's the question of whether the power company's power has overvoltages which are now not triggering reboots. It would be most unfortunate if the interruption problem were solved by this, only to introduce more interruptions due to overvoltage. I mean, it's likely that the power supplies on campus are 100 - 240 VAC type supplies.

My Fluke DMM will log minimum and maximum AC voltage. A little testing before applying a fix might be advisable.

Well isn't that slick.

Thanks for the suggestions about adding capacitors, etc. I might do that in my home system but that's completely taboo in any commercial anything. Warranties are voided, performance can be affected.

For instance, when you thought about adding a capacitor to the DC so the DC will stay high longer after a power loss, what did you engineer into the system so that the power supply is safe when power comes on and the large capacitor must be charged? Capacitors look like shorts when discharged, and the better and beefier the capacitor, the longer it would stress the supply when powering on. That is, the better this would work to hold the DC up after a power loss, the more dangerous it would be to the power supply. Unless you add a transistor, some resistors, etc...

And at that point, is the supply operating within the parameters specified by its UL rating?

In a system such as this, you don't crack open ANYTHING. You don't add a capacitor by making a male to female cable with a capacitor across it. You use the equipment out of the box, with any adjustments being only those described in the manual.