Question regarding a Switched PDU internals ?

[Black Ninja]

I am considering getting a switch PDU for the obvious reasons like remote reboots etc. For the moment I use a basic PDUs. Some of the equipment I use like Dell RPS720 makes a spark on insert, I understand that is normal and with some devices it barely happens, with other is much more severe depending on the electrical flow trying to pass before the connections is made.
I am trying to avoid disconnecting this device unless I really have to, because the spark it creates is eating a little bit of the plug (making a black spot too) every-time when I plug it in.

And now my question : How a switched PDU will deal with that ? Will the same thing happens on inside of the PDU when I remotely turn on and off ?

 

[Bidule0hm]

The same will happen but it'll be between some relay contacts and they're designed to handle it, so no problem. Maybe some PDUs don't even use relays but triacs instead so there's no moving part, no sparks, zero crossing switching capability, ...; so this means less EMI and an infinite number of on/off cycles without any contacts wear.

 

[Black Ninja]

That's what I was afraid of. If relay is used and they got eaten from the sparks internally that will be worst than manual plug in-out by hand, cause at least I can see and replace the power cord if it start looking too bad at the end.

Relays I know.
"Triacs" I don't , but I like what you are saying.

How APC PDU's were made then ? Is this a real problem ? That suggest to me that I rather not use the switched PDU because suck a devices that create bigger spark can easily damage internal parts - relays and make them inoperative ?!

 

[Black Ninja]

I guess I was right to be worried. I finally found related info on the subject which confirms all my fears. It's seems switched PDU's has very good possibility to weld their relays if used too often. That was my main reason from kept away from them in a first place.

P.S. I was kind of hoping that somebody will prove me wrong so I can start using them, but if even @Bidule0hm can't prove me wrong I guess is hopeless.

 

[joeschmuck]

You could always buy some PDU that you can remotely control and then add your own electronic switch. It's pretty easy to do. You could also just take it super simplistic and just add heavy duty relays. Basically you would use the PDU as a low current switch and the relay or triac + diac would power the heavy load.

Also, how much of a load are you talking about per circuit?

 

[Black Ninja]

Also, how much of a load are you talking about per circuit?

The loads will be pretty low. For example this RPS720 is not more that 50-60W, but it likes to make good sparks despite the low load. Some devices are like that because of parts they are build with(like capacitors). I have a old dell laptop's charger that likes to spark a lot. While other devices that suck more power 100-200W don't make sparks.

P.S. I kind of start feeling embarrassed. Is it me the only one who don't know what "triacs" are ?

 

[Bidule0hm]

I guess I was right to be worried. I finally found related info on the subject which confirms all my fears. It's seems switched PDU's has very good possibility to weld their relays if used too often. That was my main reason from kept away from them in a first place.

P.S. I was kind of hoping that somebody will prove me wrong so I can start using them, but if even @Bidule0hm can't prove me wrong I guess is hopeless.:)

Ah... welcome to the world of "almost everything is made in china now, deal with it". Well, then I think I'd open the PDU and change the relays for better ones (Finder, Omron, ...). And I'd add a snubber network (a resistor and a capacitor) to suppress 99 % of the sparks ;) Maybe I'd start with just the snubber, it should be good enough here.

The loads will be pretty low. For example this RPS720 is not more that 50-60W, but it likes to make good sparks despite the low load. Some devices are like that because of parts they are build with(like capacitors). I have a old dell laptop's charger that likes to spark a lot. While other devices that suck more power 100-200W don't make sparks.

Because the lighter loads typically have a crappy PSU with (almost) no filtering you have the inductance of the primary of the transformer (almost) directly connected to the plug. Caps aren't inducing sparks, it's inductors (usually transformers' windings) here.

 

[jgreco]

The loads will be pretty low. For example this RPS720 is not more that 50-60W, but it likes to make good sparks despite the low load. Some devices are like that because of parts they are build with(like capacitors). I have a old dell laptop's charger that likes to spark a lot. While other devices that suck more power 100-200W don't make sparks.

And just how many times in its lifecycle do you plan to need to reboot your switches from remote?

We used to see problems with the control logic on the Sentry Power Tower XL's but the APC AP7930 and its relatives have been very good. I'm having trouble recalling any issues at all with them over the last ten years, or whenever it was that they were "new".

P.S. I kind of start feeling embarrassed. Is it me the only one who don't know what "triacs" are ?:)

Probably. Why would the average person know?

 

[jgreco]

Ah... welcome to the world of "almost everything is made in china now, deal with it". Well, then I think I'd open the PDU and change the relays for better ones (Finder, Omron, ...). And I'd add a snubber network (a resistor and a capacitor) to suppress 99 % of the sparks ;) Maybe I'd start with just the snubber, it should be good enough here.

Re-engineering the innards of a UL listed device because of some fearmongering by a rack PDU competitor is actually probably a lot more dangerous than the risk that maybe someday after a few hundred cycles something bad might happen to the relay and it might fail.

 

[Bidule0hm]

When I see switches with a PSU's PCB that's less than a millimeter apart from the metal case with no insulation at all I guess I don't trust standards that much...

 

[joeschmuck]

P.S. I kind of start feeling embarrassed. Is it me the only one who don't know what "triacs" are ?:)

Probably. They are basic components in the electronic field. A TRIAC is two SCRs tied together but facing the opposite direction and the triggers connected. Do you know what an SCR is? I'll tell you, and SCR is a switch controlled diode so it works in a DC circuit. You can apply a momentary trigger voltage to the SCR Gate and then power will flow continuously until the power going through the diode is interrupted. Now you can't really use an SCR in an AC circuit because the voltage reverses and it doesn't allow current flow (because it's a diode) and turns off. If you put two SCRs together and trigger it in an AC circuit then current will flow in each direction, BUT each SCR will shut down unless there is a constant trigger source. A light dimmer has a TRIAC in it. The trigger is provided by a DIAC, Capacitor, and Resistor in a very basic controller. Fancy controls will replace the DIAC and Resistor with a digital control to provide the trigger to the TRIAC. Here is a schematic diagram of a light dimmer. D1 is the DIAC, A1/A2 is the TRIAC and the "G" is the gate, and P1 is the potentiometer (variable resistor). Hey, do you know what a Transistor is?;) You know I'm messing with you but do you know how a transistor works or how it is created? Silicon is easy to understand (maybe I accept it easier), try to understand vacuum tubes, that shit is crazy.

 

[Black Ninja]

Probably. They are basic components in the electronic field. A TRIAC is two SCRs tied together but facing the opposite direction and the triggers connected. Do you know what an SCR is? I'll tell you, and SCR is a switch controlled diode so it works in a DC circuit. You can apply a momentary trigger voltage to the SCR Gate and then power will flow continuously until the power going through the diode is interrupted. Now you can't really use an SCR in an AC circuit because the voltage reverses and it doesn't allow current flow (because it's a diode) and turns off. If you put two SCRs together and trigger it in an AC circuit then current will flow in each direction, BUT each SCR will shut down unless there is a constant trigger source. A light dimmer has a TRIAC in it. The trigger is provided by a DIAC, Capacitor, and Resistor in a very basic controller. Fancy controls will replace the DIAC and Resistor with a digital control to provide the trigger to the TRIAC. Here is a schematic diagram of a light dimmer. D1 is the DIAC, A1/A2 is the TRIAC and the "G" is the gate, and P1 is the potentiometer (variable resistor). Hey, do you know what a Transistor is?;) You know I'm messing with you but do you know how a transistor works or how it is created? Silicon is easy to understand (maybe I accept it easier), try to understand vacuum tubes, that shit is crazy.

Thanks for explaining. The idea about the opposite diodes to get the AC through make sense. I know how transistor works and I like it for that reason How is made it's different story, perhaps I'll have more luck with vacuum tubes. I've seen videos how silicon die is made , but can't quiet connect it with real world components. It's very interesting stuff, but daily life takes heavy toll on one's time and I just there is no time for everything. Still, I appreciate the time you took to explain it. Now I know what TRIAC is wand what is for.

But PDU's use electro-mechanical relays (EMR) for switching and that means they can weld their contacts ?

 

[Black Ninja]

When I see switches with a PSU's PCB that's less than a millimeter apart from the metal case with no insulation at all I guess I don't trust standards that much...

True. I have this habit to open the device first , check the components first , make sure nothing is loose , nothing to close to something it shouldn't be or anything I don't like cause I think could potentially become a problem later fix it and then close the device and use it.

 

[Black Ninja]

Ah... welcome to the world of "almost everything is made in china now, deal with it"

Don't blame the people who made the soup , blame those who order the ingredients.

Because the lighter loads typically have a crappy PSU with (almost) no filtering you have the inductance of the primary of the transformer (almost) directly connected to the plug. Caps aren't inducing sparks, it's inductors (usually transformers' windings) here.

I didn't know that. I had the impression that the caps are pulling the spark for the reason that they are trying to charge themselves instantaneous on contact and therefore making the arc, no ?

 

[Black Ninja]

Why would the average person know?

I've been called worse.:)

 

[Black Ninja]

I have a feeling the there is no way around this problem. I am just trying to understand how big of a problem is it. Perhaps the relays switching faster (compared by hand) will minimize the spark effect.



P.S.I am glad I didn't use my UPS's ability to switch outlets on regular bases for reboot modems or something like that. I rather replace burned PDU than the whole UPS when it happened.

 

[Black Ninja]

And just how many times in its lifecycle do you plan to need to reboot your switches from remote?

Hopefully .....never. I got the RPS720 mainly so I don't have to reboot when switching power sources.

We used to see problems with the control logic on the Sentry Power Tower XL's but the APC AP7930 and its relatives have been very good. I'm having trouble recalling any issues at all with them over the last ten years, or whenever it was that they were "new".

Perhaps there are other factors I don't take in consideration that are minimizing the spark problem. I wish I had disposable RPS720 and switched PDU so I can turn on and off 100 times, and then open the PDU's relay , and then everything will be revealed:)

P.S. I want to post this again, cause the video at the bottom is very interesting to see.

 

[jgreco]

That's a problem, you know a "little-p" problem. As in it doesn't make the top 10 list of issues I have when considering power designs, or even the top 20.

 

[joeschmuck]

But PDU's use electro-mechanical relays (EMR) for switching and that means they can weld their contacts ?

It's all possible but if they are rated properly then it shouldn't be an issue. Think of an HVAC Condenser unit (Whole home A/C Heat Pump outside unit with the big fan) and it has a dual pole relay called a "contactor" in that world. The sparks fly when the circuit closes and sparks fly when the circuit opens. My HVAC unit goes on an off thousands of times reliably. But these are parts which do fail but they do last a long time. I have two of these units for my house and in almost 8 years I have not had a single issue. So if someone puts in the correct parts for the PDU design, then it will last a long time. The real trick with relays which have a good arcing to them is a strong spring to force the contacts apart. Also in this design, it's four contact coming apart vice one or two. I'd tell you to get some contactors and use those but they switch on a 24VAC source so it's not a plug and play situation.

 

[Black Ninja]

That's a problem, you know a "little-p" problem. As in it doesn't make the top 10 list of issues I have when considering power designs, or even the top 20.

I am hoping you are right for everybody's sake.