Home Server Room Power Upgrade and Multi-Room UPS
blog.networkprofile.org
blog.networkprofile.org
This is very insightful. I've seen outlets installed upside-down to designate them as switched, but never heard any other justification.
This actually makes sense and the charred measurement tape further demonstrates what can and has happened.
Dug up the Reddit link for those also curious: https://www.reddit.com/r/OSHA/comments/ilb7wq/please_unplug_... and maybe original post was to Twitter: https://twitter.com/meyerweb/status/1289302216826925057
Personally, I like "normal" outlets to be ground pin down and switched (as in, connected to a lightswitch) ground pin up.
1. Why isn't your plug all the way in the socket? 2. Why haven't you replaced that outlet that doesn't hold a plug properly?
It's not a great reason, but here we are.
The cable exits the bottom, not the front, so it doesn't matter how thick or heavy it is, and pulling it out of the socket by the cable is very difficult.
Further, our earth pin is at the top by default, and our line and neutral pins are insulated to the extent that if the plug is far enough out of the socket to expose the uninsulated portion of the pins, it is no longer far enough into the socket to be receiving any voltage.
How is this even a concern?
Do people in America go around treading on cables all the time or something?
I'm not so annoyed now that I'm aware that the recessed nature is actually really helpful to avoid that failure mode.
Another reason is a lot of manufactures of grounded/polarized wall-warts and GFI plugs tended to orient the damn things for ground pin down making ground up installs prone to falling out. So you wind up with electricians following ground down to accommodate this design flaw (or habit).
I dislike the nema 5-15 blade design as it doesn't hold up well to wear. I'd much prefer a pin and sleeve design with the plug recessed in the socket like CEE 7/Schuko. But that's another argument.
The ground prong being the last to lose connection means it will not lose connection while the others a plugged in. In the reverse case, where it is the first to loss connection, it will lose connection while there is something to ground.
I've taken to using these kinds of 30cm / "1 foot extension cords" just to get wall worts a bit of distance away:
* https://www.amazon.com/dp/B00F1CLF7O/
* https://www.startech.com/en-us/cables/computer-power/externa...
There could be stored charge in capacitors as in older TV sets or switching power supplies. Theoretically I suppose a short to one of these could energize other metallic components and would be immediately discharged if the ground is still connected.
Also useful if you are using metal cover plates, as the author did, because if the plate falls off, it won't as easily short hot to something else.
It’s not uncommon in other countries for the plug to be designed in such a way that by the time conductive surfaces are exposed, then the plug is far enough out that it’s no longer live. There are a few ways to accomplish this, either from recessed plugs to requiring that the exposed conductors be half coated in an insulator so that the first part exposed as the plug is removed is non-conductive.
I don't think the extension cords are a problem because they're not in the wall (and I see actual Romex used in the wall).
However, I would take the time to carefully research exactly what pieces of this are, and are not, code compliant. Not because you need to redo it, necessarily, but because it is a very impactful way to learn these things and learn deeper concepts of safety and failure behind the rules. Nothing stays as well learned, or remembered, as details relating to work you have actually done.
A lot of lessons learned, hard lessons learned (like fire, expensive equipment damage, injuries and blindness, and big sparks) have happened to get electrical code to where it is.
[1] https://www.leviton.com/en/products/3325-c
Then you don't have to worry about a "tight fit" or something dislodging (or half-dislodging) the plug.
I’ve only used it for specific Amp/voltage required situations (UPSs actually).
https://en.wikipedia.org/wiki/NEMA_connector#/media/File:NEM...
Surprisingly, there is no physical shape (that I'm aware of) to distinguish between 120V and 240V, it's up to the user to be aware of possibility of alternate voltages. I have seen color schemes used to distinguish voltages.
Another alternative would be to use Lutron's Claro AC outlets designed for dimmable control. They've got a smaller prong for the hot leg.
I have some sockets for outdoor holiday lighting that I've colored to avoid mis-using them for anything else.
The whole point is I can easily expand in the future to charge from a solar panel array; solar panels are cheaper and cheaper, the biggest cost is the batteries and installation. Until then, the batteries can be charged from the utility company and the efficiency is close to an APC Smart UPS.
Seeing the limitation in how many of the UPS sockets can be used, I'd say it's a little bit of a shame not to have used high amperage IEC connectors - e.g. C19/C20 ones.
Most UPSes have options to work with these natively rather than having country-specific and often painfully shaped connectors (e.g. UK sockets are huge!)
A subpanel would make future expansion easier.
1- https://download.schneider-electric.com/files?p_File_Name=AS...
If you wanted to do this up to code, how would you do it?
If they sell those power inlets which are NEC approved, how can you use them without breaking code? So it MUST be code. But then you ask 10 people and get 10 answers
There are larger UPS's designed for hard-wiring, but they are almost always VERY large. I think that might be the only way to avoid an inlet cord like this
At least in my jurisdiction (America) the only way to achieve what you are trying to achieve is by installing a dedicated subpanel fed by the UPS with appropriate breakers. I believe that such an installation would only be code compliant with a UPS that was designed for such an application (with a hard wired output connection, and appropriately designed disconnects), but a lenient inspector might let you get away with feeding a subpanel from an extension cord as long as as of the breakers and capacity factors work out correctly and the grounding meets code.
The biggest issue I see is that the system is not labeled correctly (at all?). That is a pretty serious issue because if, for example, an emergency worker were to come into the house under the expectation that the power to the house was cut (at the service disconnect) but the house would have live outlets that are otherwise unmarked. See https://www.mikeholt.com/download.php?file=PDF/11_Generators... for more information (the relevant section for your system starts at page 16)
As other commenters have pointed out, just because you can buy something that says NEC approved on the sticker does not mean that it is approved for your application. A good example mentioned elsewhere is using a 16awg SJTW extension cord to power a 15a load--all of the connectors will plug together but the combined solution is obviously not code compliant and even if it doesn't melt/catch fire it will not supply the load with clean power.
So the first question to answer is what NEC allows, but you will then need to double-check for your specific context.
Further, the power inlets are NEC approved but that in no way means that they are approved for this use. I'm far from an expert on the NEC but I can pretty confidently assert that it allows such inlets only for specific uses, and you would need to check carefully that what you propose is one of those uses. A very rough rule of thumb is that you are not normally allowed to have more than one "user-removable plug" on the path to a device. There are a number of exceptions to this, but any time you're plugging into something that plugs into something else, you need to check carefully if what you're doing is allowed.
The situation that is dangerous here is if a power source were to be connected to one of these inlets and it did not have a sufficiently-sized breaker and the other side has a dead short, you are looking at a fire inside your walls.
I just installed a 50 amp generator inlet on my main panel, and even though my generator has a matching 50 amp circuit breaker, I installed a separate matching 2-pole 50a breaker on the inlet circuit. This protects the electrical system from even the most screwed up of generators being connected to the panel.
Here there is a breaker, and if the UPS is disconnected, there is no power. Of course someone could go and plug this into a 30a outlet, however that's something that I wouldn't do, so its not really my issue
That would be against code.
There's nothing stopping you from running bare wire through your walls.
There's nothing stopping you from not bothering to ground any outlets.
There's nothing stopping you from using 14AWG romex on a 50 amp breaker.
Etc., Etc.
All of that is against code, and I've personally seen all of that in the wild at people's homes, offices, etc.
If you wanted to do things the #rightway, you'd most likely need a subpanel for the UPS to feed.
> however that's something that I wouldn't do, so its not really my issue
Electrical code is designed so that even an incompetent operator cannot produce a dangerous situation inside a building's electrical system. OSHA, USCSB, et. al. would be the first to tell you that relying on a specific individual to interlock a dangerous situation can result in extremely adverse outcomes.
Flexible cords and cables shall be used only for the following...(11) Between an existing receptacle outlet and an inlet, where the inlet provides power to an additional single receptacle outlet. The wiring interconnecting [...] shall be a Chapter 3 wiring method. [...]
I don't know what Chapter 3 is though, and in the below study guide link it seems to show a double outlet in the graphic.
https://www.electricallicenserenewal.com/Electrical-Continui...
the code you referenced states 'Between an existing receptacle outlet and an inlet'
the circuit is also unlikely to be arc fault protected, which would mostly likely be required.
also how the romex is run, is not to code.
One of the things I think would be most likely to throw up a red flag and immediate fail would be the "suicide plugs".
Even though they're not energized (but rather, used to deliver power to the remote end), most electricians and inspectors wouldn't sit around waiting for you to explain your brilliant power delivery scheme.
The only real way to do this stuff code-compliant would be to purchase and install a real hard-wired UPS system, use properly-installed Romex (or other suitable code-compliant wiring), and get things inspected by the local authorities.
Usually if it's not installed by a licensed electrician, the inspectors will be even more stringent.
Suicide plugs refers to extension cords which have been built (or modified) so they are male/male. One in which the far end will be exposed while energized. In this system, the male plugs will never be energized while exposed.
Mostly because 'sparkies' often wire things up like this and use random extension cords in walls and create fun fire hazards.
I agree that the installation is in 'the spirit of the code', but there are enough little variances that I would take a bet that a random inspector would not pass it on first glance.
Not sure where all this worry of code when selling a house comes from. Code changes. No 20 year old house is up to current code, doesn’t mean you can’t sell it.
After the author began the article by whining about the previous owner's outlet wiring, I expected better attention to detail.
And as mentioned elsewhere, I doubt the JHA or the manufacturer of the UPS would approve of this kind of installation. There are UPS specifically manufactured for a hardwire installation and that's what ought to be used here. That said, I am sympathetic, because UPS can become very expensive very fast. I'm balking at the cost of 2U rack models from Eaton, he will have had to shell out a good chunk more based on his VA requirements but at least he avoided the significant premium for hardwired.
It's called 2x Tesla Powerwalls and a 7.5Kw solar system! :-)
I wonder how much money utilities could save if everyone had a big in-home UPS/Solar and 99% or 95% grid reliability were acceptable.
They do exist, however, and hopefull they will get better over time. This will allow more people to add a moderately sized battery system to a grid-tied system, and then we end up with the best of all possible worlds (or something much closer to it).
There are also grid-tied inverters that will provide limited online power even with the grid down. These also seem to be not as reliable as the "grid down? i'm down" design.
The really tricky part is that for the inverter's output to be instantly switchable with the grid, it has to be phase aligned with it, and so the inverter has to see the grid phase and match it.
They’ve gone 403.
https://web.archive.org/web/20140325203319/http://www.mdisc....
Also, are they hot if nothing is plugged in?
They are reversed from an outlet because don't ever want the hot part to be exposed.