What is the output circuit of a switched PDU?

[Magnox]

APC make some great network-connected switched Power Distribution Units for racks. What is a suitable output circuit for such a thing, given that it's controlling a SMPS in very expensive server/network gear?

I'm asking because I want to make one for my man cave, to save my old legs having to carry my fat, lazy body upstairs all the time to turn something on in my lab. Mostly Cisco switches and routers and some small servers. Rated power consumption on them ranges from 100W to maybe 500W.

I have a handful of big Crydom SSRs, but the leakage current (through the snubber?) when they are off is enough to make a laptop SMPS blink its power light all the time. That can't be good for it.

I guess a simple triac output might 'misfire' due to the way an SMPS draws current. Not sure if that's actually a problem though, provided it can turn on when the SMPS needs current and off fully once the trigger is removed. Google isn't really helping much, seems to be confusing me more than helping.*

Of course, a relay would make those points moot, but do APC use relays? Besides, I'm trying to make this without spending any money (because I don't have any) and I have plenty of SSRs and triacs (not sure I have isolated drivers for those though). I don't have any suitable relays; I think the ones I have (nice little latching ones, 250VAC,8A but only 2000VA/240W) might die soon from switching SMPSs.

So... what do the big boys do?

*I don't really use Google, I use Ixquick.

[Gallymimu]

You could use the SSRs and bleed the leakage current through a resistor, though that is a terrible terrible idea

[Magnox]

Yeah, that's one of the solutions I vaguely came up...

However, at 10mA rms leakage through each SSR, having a resistor value low enough to be useful when off would dissipate far too much power when on. Multiply by 16 if I have everything on at once and by the time I've bought enough heatsinking and forced-air cooling, I might as well buy the PDU. Never mind the environmental impact...

Terrible idea 

[Sushilogic]

I'm not sure what the big boys do, but a relatively low cost solution could be to use an optocoupler triac driver (with zero crossing detection to keep the switching clean) and drive one of your triacs from there. Something like this : https://www.fairchildsemi.com/products/optoelectronics/triac-driver-optocouplers/zero-crossing-triac-driver/MOC3163M.html

They are fairly cheap.

Datasheet : https://www.fairchildsemi.com/datasheets/MO/MOC3163M.pdf)
App Note : https://www.fairchildsemi.com/application-notes/AN/AN-3004.pdf

More here : https://www.fairchildsemi.com/products/optoelectronics/triac-driver-optocouplers/

Drive the input from a microcontroller and if you want it network connected then you could prototype something using the GPIO on a raspberry pi or an MCU dev kit with ethernet or perhaps even better,  wifi. Maybe something like Electric Imp could be used to get the whole thing controlled from a web browser?

[Magnox]

Well lucky me, I just found a tray of MOC3162M in a drawer, so I'm sorted for drivers and triacs.

For the network side of things I have a few options. I'll probably end up using a Lantronix X-Port Pro; I have a couple loitering around looking for a purpose in life. Clever little things they are. Not much bigger than a normal network socket, with a mini linux core and a couple of output/UART pins.

I would still like to take apart someone's APC PDU though, to see what they do.

[Sushilogic]

The X-Port Pro sounds pretty cool. Yeah - I'm curious now myself as to how APC does it. I'll let you know if I find anything on that.

Sounds like a fun project though that could be used for general purpose AC switching - add some analog inputs or at least one-wire digital and you could use it to control all sorts of things.

[Magnox]

I got my answer...

APC PDUs use relays, and bigger ones that the babies I have a drawer full of! I guess there's a good reason for that.

[zac]

I got my answer...

APC PDUs use relays, and bigger ones that the babies I have a drawer full of! I guess there's a good reason for that.

You might consider latching relays which will avoid the continuous coil current. 

http://www.mouser.com/ds/2/357/105A_755-7090.pdf

[Magnox]

Yeah, I have a box full (about 50 I think) of latching relays that I would love to use, I mentioned them in my OP, but they are probably too small to handle the potential surge current.

I would also want to implement a dying-gasp circuit to reset all the relays on power loss.

Then again, it's not really the holding current on relays that's an issue, just that I'm trying not to spend any money on this project. Given the price of decent relays (especially latching ones), I could almost buy a second hand PDU for the same cost if I'm patient on the auction site.

I might be best to sell the box of relays and use the funds to buy a second hand PDU...