Floating neutral and UPS

[pc1966]

Here is my dilemma. Picture this arrangement: a system with a 100A three-phase supply will have a RCD incomer (and possible some generator transfer switch in the future). As well as the SWA supply cable's earth there will also be some lightning protection so a few earth rods as well. From the distribution board there will be around 4-6 single phase UPS, of 3-5kVA each, and they will be distributed over all 3 phases vaguely tiring to keep the load balanced.

What worries me is when the RCD trips (testing, if nothing else...) or the transfer switch is going between mains and generator, the neutral to the board will be disconnected and floating. But the UPS will still be powering everything (after all, that is their job) and it bothers me that we might see the neutral getting to 50V or 100V and so one phase going over 300V with respect to earth depending on the leakage currents and how they might balance (or not). In the past those UPS show 10-20mA each.

My first thought was to put something like a indicator lamp between neutral and earth to:

• (a) act as a warning that something odd is up (i.e. significant N-E voltage), and;
• (b) keep the neutral below 10-20V w.r.t. earth if only tens of mA unbalanced leaking when the neutral is isolated.

Now I know the regs say not to join N & E at more than one place for various reasons (like high earth loop currents, etc) but is this sort of proposal, a medium impedance from N-E for leakage voltage control, either prohibited or just foolish for some other reason I have not yet spotted?

TL;DR: Is putting a small lamp between N and E prohibited?

 

[Lucien Nunes]

You don't mention what type of UPS is in use here, which is important as some pass the neutral through continuously and others don't. If the UPS inputs are DP isolated during supply loss, then there is no problem as far as the DB neutral is concerned, it will be dead. Downstream at the UPS outputs, you might needs to ensure that the neutrals (which will all be isolated from one another and supply neutral during power loss) are correctly referenced to earth if the loads and protection scheme requires. (it might not matter, if the loads are OK with an IT supply and DP switchgear etc is used). Nothing here differs from a normal installation with various small UPS loads located around it.

If as you imply the input neutrals are not isolated from the loads during supply loss, and the UPS outputs are not synchronised, then all sorts of voltages including pulsating DC and 460V AC can be present between the lines as the phase relationship that normally exists between lines on the DNO or genny supply will be lost. I would seek the advice of the UPS manufacturer.

 

[pc1966]

I think all the UPS are Dell models but not all the same. Of course Dell do not make UPS so they could be re-branded Eaton or APC units but have to check the part numbers. All are of the "continuous conversion" type so they are converting via DC at all times so no break on supply loss, etc.

That is a good point about them losing synchronisation when power goes off so any assumed balance in leakage would be invlaid so it could be as much as 100mA if 5 at 20mA all in-sync after several minutes of isolation.

 

[Pete999]

Probably best if you went for 1 one UPs to cover your needs rather than tryint to manage with multiple 1 phase units, IMO

 

[pc1966]

Probably best if you went for 1 one UPs to cover your needs rather than tryint to manage with multiple 1 phase units, IMO

As the joke about the village idiot being asked for directions goes: "If I were you, I wouldn't be starting from here".

Thing is they already have the UPS, they fit various racks for different jobs, and replacing with something better (i.e. 20kVA 3-phase unit) is going to be something around £10k extra. Also you have the issue that a fault on the central UPS takes down everything. In the past they have seen at least 1 (if not 2) single phase UPS go bang during self-test and drop the load. Hey UPS, you had just one job to do!

Just now they are being moved from separate buildings, each on single phase supplies, to this one location with a centrally controlled 3-phase board, hence the rather odd problem here.

 

[Pete999]

As the joke about the village idiot being asked for directions goes: "If I were you, I wouldn't be starting from here".

Thing is they already have the UPS, they fit various racks for different jobs, and replacing with something better (i.e. 20kVA 3-phase unit) is going to be something around £10k extra. Also you have the issue that a fault on the central UPS takes down everything. In the past they have seen at least 1 (if not 2) single phase UPS go bang during self-test and drop the load. Hey UPS, you had just one job to do!

Just now they are being moved from separate buildings, each on single phase supplies, to this one location with a centrally controlled 3-phase board, hence the rather odd problem here.

not if you have an Automatic By pass configured in the UPs and a no break set up with the standby set

 

[pc1966]

not if you have an Automatic By pass configured in the UPs and a no break set up with the standby set

Do you have any suggestions / part numbers to take a look at this option?

When the single phase units went bang they just were dead, fault code on display and no possibility to bypass internally them to get power out (other than going externally). And yes, that really sucked!

 

[Pete999]

Do you have any suggestions / part numbers to take a look at this option?

When the single phase units went bang they just were dead, fault code on display and no possibility to bypass internally them to get power out (other than going externally). And yes, that really sucked!

Been a while Mate, but I used Emerson UPS 10.15 or 2oKVA were my briefs UPs FED FROM mains and ST.BY set output SP, suggest you google Emerson UPs and go from there. they were always will to provide a solution, easy to connect up. simple to use, it's been a few years since I used them so no new information sorry. I think the units I used were AP20 for 2o KVa unit, should be a simple google search.

 

[Simon47]

Do you have any suggestions / part numbers to take a look at this option?

Can't help with nubers or anything, but most manufacturers will offer you a pre-built bypass switch. You can make your own - it really is just a changeover switch with make-before-break action. As you can imagine, with a switch that momentarily connects UPS input and output together, operating it requires the operator to know what's they are doing ?
I've attached the schematic for a commercially available unit - which I modified to add a power inlet for alternative power. As supplied, the AC Input switch is just on-off. The changeover switch is badly drawn, in reality it's Off-UPS-Bypass with make-before-break between UPS and Bypass.
When switching between UPS and Bypass, you MUST put the UPS into local bypass mode which I'm fairly sure every half-decent UPS should have - and anything but double-conversion types will always be in what is effectively bypass while there is mains present.

When the single phase units went bang they just were dead, fault code on display and no possibility to bypass internally them to get power out (other than going externally).

Seen enough of those. It's one of those hard things to count, because unless you check logs (or have automated monitoring to do it for you) then you don't know how many mains problems the UPS the UPS has saved you from. On the other hand, without careful configuration, when the UPS fails, then your load fails too. It's possible that I've seen as many failures due to the UPS as due to mains issues - but it varies according to quality of mains.
Where I used to work, the big UPS failed and we ended up running direct - I had servers with up times of well over a year. We were lucky in having all underground mains, and we were fairly close to the 132kV substation. Some of our customers "really needed" their UPSs

And yes, that really sucked!

There is one thing worse than finding everything quiet with a failed UPS ...
That's when you are doing something, flick a switch, and instead of "nothing happening" there's a clunk and just the sound of all the equipment fans and disks winding down. That sound is usually followed a few seconds later by colleagues asking "what happened ?" (or words to that general effect )

 

[pc1966]

I once had an external UPS bypass switch, was some fancy spring-loaded rotary switch that was quick enough you could change over "live" and the PCs, etc, did not seem to notice any break in the supply. Only 2kVA or so though unfortunately.

One other suggestion is to use a transformer so the UPS feed could be configured to have N & E bonded at the transformer secondary (star-wound), so a floating neutral on the incoming supply (i.e. transformer primary, could be star or delta) would not cause the UPS to see anything other than a loss of power.

But I doubt that a 20kVA transformer will be much less than £2-3k and also I am not sure it would be that quiet and/or cool to stick in the equipment room along side the rest of the stuff. Any further thoughts/suggestions?