Equipment needs 208 volts single phase. Supply is only 480Y/277. So a transformer is needed. How would you wire it?

First I'd see which single-phase transformer was cheaper and more readily available: 277V to 208V or 480V to 208V.
Then feed it with either a 1-pole or 2-pole C/B respectively.

You'll also need an OCP on the 208V side of the transformer.

How would you WIRE it? Where would the ground come in to prevent it from being an ungrounded system?

You only need to bring out EGC due you have line to line load that will be bonded to the transfomer.

If this is line et netural load then everything will change a bit.

Merci.
Marc

You would still have to ground it somewhere on the secondary. The easiest way is to make one leg a neutral by grounding it. I would use gray for that neutral and orange for the hot, just to wake up the next guy that there is something strange going on here. (similar to the gray/violet on the 277)
The only problem with that idea is orange is one of the BOY colors and you have 480 there.

pdh:

There are (3) ways to go on this scenario - all are based on what is acceptable for the Equipment involved.

OPTION #1

This option is the most simplest: Tapped Autotransformer.

Using a 277V wound Autotransformer, with a tap for 208V:

1. Connect the Transformer across a 277V 1 Phase 2 Wire "L-N" Circuit.
2. Place the Grounded Conductor on the "Common" side of the Autotransformer.
3. From the 208V Tap, bring out an Ungrounded Conductor.
4. From the "Common" side connection, bring out the Grounded Conductor.
5. From numbers "3" and "4" above, a 208V 1 Phase 2 Wire Circuit is created.

Insert OCPDs as necessary. At minimum, OCPD for the Primary side Feeder.
Include the EGC with the Primary Circuit, and bring it out for the 208V Secondary Circuit.
Do not bond the Grounded Conductor!!!

______________________________________________________
______________________________________________________

OPTION #2

Single Phase Isolated Transformer with 208V Secondary.

As mentioned by Ghost307, use an Isolated transformer with a Primary Voltage rating that is readily available.

Feed the Primary side from the available System:

* For a 277V Primary: One Ungrounded Conductor, plus the System Grounded Conductor.


* For a 480V Primary: Two Ungrounded Conductors.

Run the EGC to the Transformer case; bond it to the case.

On the Secondary side, bond Terminal "X2" to a local Grounding Electrode System.
Additionally, install a Jumper from "X2" to the Transformer case.

Derive an SDS from the Secondary side for the new 208V 1 Phase System as follows:

1. From Terminal "X1", bring out the 208V System's Ungrounded Conductor.
2. From Terminal "X2", bring out the 208V System's Grounded Conductor. Identify this Conductor as a System Grounded Conductor, using either White or Gray Coloring.
3. From the "Jumper" at Terminal "X2", bring out an EGC. Identify with Green Coloring.

The Secondary Feeders _MAY_ be protected by the Primary Feeder OCPD, or you may choose Over Current Protection for both the Primary Feeders and the Secondary Feeders.

Since the output Voltage is only a single Voltage (208V only), the Secondary Feeders may be protected by the Primary Feeders' OCPD as described below:

A: If the Primary rating is 9 Amps or more: Maximum OCPD on Primary side is 125% the FLA of the Primary Winding(s).

B: If the Primary rating is less than 9 Amps: Maximum OCPD on Primary side is 167% the FLA of the Primary Winding(s).

C: If the Primary rating is less than 2 Amps: Maximum OCPD on Primary side is 300% the FLA of the Primary Winding(s).

The option exist to run the Secondary side as an Ungrounded System. This does not eliminate the need of bonding the Metallic Equipment connected to the Secondary side Circuitry to a local GES.
The ONLY difference between a Grounded System -vs- an Ungrounded System is installing a Main Bonding Jumper; or in this case, a "System Bonding Jumper", between the SDS and a local GES.

For more information on Grounded -vs- Ungrounded Systems, refer to the Transformer Schematics posted at the Technical Reference section.

______________________________________________________
______________________________________________________

OPTION #3

3 Phase Isolated Transformer - 480V x 208Y/120V:

This option would be the most involved, but may be the most cost effective!

This type of Transformer is an off-the-shelf item, readily available from nearly all Wholesalers.
The costs and availability might result equal to a Single Phase setup.

This option allows for future expansion and usage.

Setup would be similar to "OPTION #2" above, with the following revisions:

A: Primary side fed with 480V 3 Phase 3 Wire Circuit + EGC.

B: Primary side Over Current Protection only covers Primary Feeders.

C: Secondary side shall be Grounded. Ground at Terminal "X0".


D: Secondary Feeders to be Protected by OCPDs. Primary OCPD may not protect Secondary Feeders.


E: Secondary side will consist of (3) Ungrounded Conductors, and (1) EGC bonded to the GES and the "X0" Star Point.


F: Secondary side Grounded Conductor _MAY_ be brought out from Terminal "X0" to a Panelboard / Disconnect if wanted.
If the Panelboard or Switch (Disconnect) will be feeding only L-L / L-L-L connected loads, the System's Grounded Conductor does not need to be included - provided the Main System Bonding Jumper exists at the Transformer, and there is a suitable EGC between the Transformer and the Panel or Switch.

................................................................................................

As far as "What Would I Do", I would do one of the three options listed above!

BTW, Abbreviations used:

* EGC = Equipment Grounding Conductor.

* GES = Grounding Electrode System.

* OCPD = Over Current Protection Device.

* SDS = Separately Derived System.

* L-N = Line To Neutral.

* L-L = Line To Line.

* L-L-L = Line to Line To Line (3 Phase 3 Wire)

Scott

Wow, Scott35, lots of detail.

I wanted to understand what people thought of a single phase 208 volt system. In the form of transformers, it is easier to discuss and understand, so that's why I asked in terms of transformers first. But really, I'm looking at this in terms of UPSes/inverters, and how a 208 volt system is formed. I have a choice between two kinds of UPSes for a bunch of computers:

1. Has one output inverter producing 208V directly (can be programmed to one of 200/208/220/230/240 volts, 50 or 60 Hz). Appears to NOT be grounded in any way. Questions to their tech support people get strange answers, but finally conclude it is ungrounded (neither terminal is a grounded conductor). Any relation to ground will be through the inverter circuitry itself. The input is 3 wire and has an L6-30P.

2. Has two output inverters producing 120V directly (can be programmed to one of 100/110/115/120/127 volts, 50 or 60 Hz). Specs also indicate they can be programmed to 120, 180, and 240 degree phase angles (I guess the 240 is there to reverse any sense of phase rotation without having to reverse the conductors). So apparently it can produce any of 173/190/200/208/220/230/240/254 volts L-L output through appropriate combinations of L-G/N voltage and phase angle. The input is 4 wire (A/B/N/G) and has no plug (why not an L14-30P).

I'm leaning toward the latter because it is grounded. The former is lower in cost and was already ordered before I was hired. Normally, when trying to evaluate equipment, vendors bring along a sample unit and often leave it for a few days trial. This isn't practical in this case.

I guess I biased my question when I replied with asking how to ground the transformer secondary. I wondered if someone might suggest to just leave it ungrounded. Would you do that for a bunch of computers with multiple ungrounded UPSes?

The problem is not going to be with the computers, they use isolated power supplies anyway that only look at line to line voltage.
Since this is greater than 150v to ground you are not bound by 250.20 so I was probably wrong about having to ground it.
This is strange territory for me. I have seen lots of 208v but it was always 3p wye or CT delta wild legs.

Of course the PSU will see the 208V L-L. But if it's floating, my concern is the phantom voltages that can disrupt interconnections between computers on different UPSes. Just a wee bit of hum on the ethernet connections can ruin their performance.

3pY would work better in the sense that it has to have the ground reference. But UPSes which output 3pY seem to no longer be made with less than 10 kVA. One rack cabinet doesn't need that much power. Not only is there a premium to the cost for 3p, but also for the excess capacity. That, plus a distribution box wired to balance the 2w1p loads across the 3p would be hard to find. 3p is just not commodity enough to be cost effective.

There is no connection between the line voltage and the output of the power supply. This is immediately converted to 20kz or so AC and pumped through a toroid transformer that produces the various DC voltages.

These power supples are designed for a global market and I am sure there are countries that do not use grounded line voltage.
Ethernet rejects 60hz hum anyway. What they do have problems with is large ground shift transients from baseplate ground in one machine to another. (tied to the the EGC)
It tends to fry the card and might even start blowing up other stuff hanging on the bus..

The NEC doesn't make you ground one leg but it doesn't prohibit it either.

Ethernet rejects induced hum from external wiring due to the twisted pair nature. Hum as a point to point potential between the interconnected devices is not so well rejected. That's the mode I always see when I do run into hum problems with it. It's a problem that can't be 100% eliminated short of running everything on DC. But the higher the voltage between equipment, the greater it can be. A pair of ungrounded 208V systems could have the effect of a 416V system, depending on the phase angles and other aspects at the time. Grounding the 208V systems on one end would prevent it being like 416V. Grounding on the center tap would be even better. There are other ways to mitigate the problem, too, but grounding the power system is one of them.

There's also the risk that the ungrounded system can exhibit higher differential voltage in a surge than an ungrounded system. This can be mitigated by additional surge protection to help maintain the voltage equality (e.g. clamping the 208V system into ground, during the transient).

Here are two different UPSes. The first appears to be a single ungrounded-output inverter. The second (more costly) appears to be grounded.

http://www.apcc.com/products/resource/include/techspec_index.cfm?base_sku=SURTD5000RMXLT3U

http://www.apcc.com/products/resource/include/techspec_index.cfm?base_sku=SURTD5000RMXLP3U

A simpler case, but much the same. I want to bring a 240V 20A circuit to my computer room at home. Output will be a dedicated 6-20R or L6-20R. As you know, this is a split-phase system, and both line wires will be 120V relative to ground and no neutral is present or needed. So I want to have a UPS provide exactly the same thing for a group of computers, without bothering with a transformer. What are the issues?

A 240V system has the 2 separate 120V waveforms at 180 degrees between each other. The only way to get 208V out of 2 120V circuits is for them to be at an angle of 120 degrees.
For that you either need 2 legs of a 3-phase circuit, or you need to play with the phase angles using a transformer.

I agree if the GROUNDS between the systems have a shift it can show up in the ethernet but in switch mode power supplies the DC voltages are very well isolated from the line voltage.
We have had to go as far as to bond the frames of remote machines together to avoid transient problems.

gfretwell ... I think that's what I'll do, since it looks like I'll probably be using the UPS that has a single 208V output only. So now I need to find a good ground strap. Fortunately all the machines will be in a room, metallically isolated from anything else (network connection via fiber), with an SPD on the subpanel that feeds them (and nothing else).

Where we have seen ground shifts were in places where there were separate grounding systems in separated buildings but the buildings were connected with LAN cables. We bonded the machine frames to prove the problem but in the end the right fix was probably bonding the ground electrode systems together.
At that time I was in the data end of the business and bonding machine frames was the fix we often used. If there was a scary amount of current on that wire we got the electricians involved. If the voltage was much more than 35 volts or so we wouldn't touch it until the grounding system was inspected but 20-30 volts was not unusual.

In the same room, I'd bond the machine/cabinet/rack frames together. At greater distances, these days, the way to go is fiber for data. Then just power each "zone" separately, with its own surge protective device. All communications in/out of any "zone" has to be restricted to fiber only unless it can be protected with the SPD for the power (not something you are likely do be doing in an I-Line, for example).

This has turned out to be quite a detailed thread!!!

I have some comments regarding the items covered so far.

Per the UPS equipment specified by the OP "pdh":

After reviewing the Products' Documentation, I am unsure if either of the Two Models will have a Grounded Output - while the Output is running Isolated via the Inverter / UPS Transformer (in Normal Mode or in Bypass Mode).
OP indicates that, after contacting the Tech. staff for verification of a Grounded system, it appears that the single output voltage system does not have a Grounded System Conductor.
This would be the UPS with an output of 208V 1 PH. 2 Wire (with EGC).

The above UPS _MAY_ have a Center Tap on the Secondary, and _MAY_ be Grounded via that Center Tap.
Results would indicate 104 VAC L-G from either of the 208 VAC lines, and would deliver a Grounded system without a Grounded System Conductor.

It could be possible that either of these UPS units do not have the Secondary Winding bonded to Ground. Only a Schematic or Field Test could verify.

When the UPS is set in Bypass Mode, the Output is directly connected to the Circuit driving the UPS Input - and the Isolated System is not connected.

As to using an Ungrounded System (setting up an Ungrounded System)...

Let us first off designate what will be considered an Ungrounded System for this scenario.

My description of an Ungrounded System is:

*** A Power System of "X" Volts AC, which does not have a DIRECT connection to a Grounding Electrode System, yet will have Metallic Enclosures Bonded to a Grounding Electrode System, along with a Measurable Voltage-To-Ground across a given Impedance ***

The Voltage-To-Ground on the above described Ungrounded System will vary, according to the Impedance between a given System Conductor and the Grounding Electrode System.

The higher an Impedance value is, the higher the Voltage between the Two points will be.

My description of a Grounded AC System, which does not use a System Grounded Conductor would be:

*** An AC System, which has a DIRECT connection to a Grounding Electrode System, yet does not utilize a System Grounded Conductor with the normal "Output" Circuitry from the Secondary Winding(s) of the Transformer, but does utilize Equipment Grounding Conductors which are Bonded to both the System and the Grounding Electrode System ***

In the case of the above System, the Center Tap of the Secondary Winding (or one Secondary Winding), will be Bonded to the Grounding Electrode System, without deriving a Grounded System Conductor from that Tap point.
All EGCs will be bonded to that Center Tap, as well as the GES.

The 60/120V 1 Phase 2 Wire System is one example.

The issues involved with high L-G Voltages experienced on an Ungrounded System may rapidly damage SPDs, as well as line Filters at the input of a SMPS.
The typical MOV would be subject to high L-G Voltages, corresponding to the Impedance of the device.

To reduce the issues connected to using an Ungrounded System - in addition to using an AC Circuit without a Grounded Conductor, the "Center Tapped" System above would be a good choice.

I agree with the Fiber Backbone idea between two or more Buildings, and/or for long distances.

Even with all Electrodes connected together, there will still be a difference in potential between the two Electrodes.
Circulating Currents are likely to flow on Conductors bonded to the GES.

Scott

Most of the places where we were having problems with remote terminals did not have enough on the far end to justify running fiber. At least not in the customer's mind.
It was usually just one point of sale terminal out in a tiki bar, restaurant or golf cart barn/starter shack.
They only got blown up when we had a thunder storm but that was every afternoon in the summer.

I protect my home computers by having the internet connection done via wireless, with a pre-configured pre-tested backup router ready to replace the one in use. Electric comes in underground (underground MV to pad transformer and underground LV to home) and goes through point-of-use protection, then UPS, then a couple IsoBar filters (not cascaded). I'm intending to add further surge protection to the panel some day.

I'm still trying to figure out a good way to switch this to 240V L-L.

I just go to laptop and power issue problems drop-out between the power supply and the internal battery.

It's the wave of the future.