Reverse wired tranny inrush

[templdl]

Basically if one were to use a 3ph 45kva transformer w, 480d-208y/120 as an example the primary of HV FLA would be 54.2 and the secondary 208 LV FLA would be 125.
If you were to apply an OCPD of either applied at 125% or the next standard rating of course the LV amperes is greater. One would expect that the inrush of magnetizing current would be the same percentage if the transformer was to be energized from either the HV or LV side which is an incorrect assumption. Because the LV windings are next to the core of the transformer the inrush is a much higher percentage, of the FLA. How much higher? There are no standard values available unless and engineer does some calculations for you.

If the inrush were to be 10X at 54.2 when energizing the HV winding one would expect the inrush ti be about 540A. Then it would be logical to expect the inrush to be 1250A (125x10) if you were to use tat same transformer as a step up. This is a wrong assumption however. Because the LV windings are next to the core that 1250A could easily be 50% more or as high as 1875A.

As such defining what the word "higher" means is important. The reason for the inrush is that the transformers? core goes into saturation until counter EMF can be established by the alternating current which last about 1/10 of a second before it settles down.

If a transformer has been designed for a step up application the LV windings are either placed on the outside of the HV windings or the flux desity is reduced. On 600v transormers the LV winding is normally place on the outside where the taps would be located. Please remember that it is very difficult to locat the taps in winding that are locatre in the inside.

On medium voltage transformers place the HV coils nest to the core is not a good idea. As such it is better to be left on the outside and the flux desity of the core reduced so that the core doesn't go into saturation as easily.

Basically there are no exact figures to go by and the objective is to know if you are playing in an area which can cause you grief. Use the highest rated OCPD as you are allowed by the NEC.

 

[mivey]

...and therefore in the reverse direction the new secondary voltage would be slightly less than double the primary...

I just made a quick assumption that the voltages would be the same since the turns ratio had been tweaked to account for the internal impedance. I guess it would not have been tweaked the same for the step-up case. I suppose drawing it out would be clearer, but I don't feel like it.

 

[gar]

100216-2218 EST

mivey:

You know the following, but for clarification to others of when I was writing above and not thinking I submit:

Suppose I compensate the turns ratio for an internal 5% voltage drop in the transformer at full load, then the approximate ratio is 2/1.05 = 1.9/1 . In its step down use from 240 the no load open circuit voltage is 240/1.9 = 126 V.

When this transformer is run in reverse, then 120 is stepped up to 120 * 1.9 = 228 V. Now assume a 5% drop at full load and the secondary output is 216 instead of the desired 240 V. 10% low as should be obvious.


templdl:

Is it really true that on a large power transformer the leakage flux characteristics are so greatly different that there is a substantial difference in inrush current on a 1 to 1 turns ratio depending upon whether an inner or outer winding is the primary? I would expect more difference from wire resistance if the same diameter wire was used on both windings.

.

 

[templdl]

Power/distribution transformers are not compensated, the ratio is a ratio.
With MT transformers it is a different story. If you use a MT transformer as a step up you will be surprised that the resultant HV is significantly lower than you had anticipated because the ratios are not let's say 4:1 for a 480-120. Remember that an MT transformer has to maintain enough voltage to support the inrush of relays. Their secondary ratio is a number of turns greater than 4:1 which increased the secondary voltage. You would have a very rude awakening if one were to attempt to use a MT to drive a 120v scale voltmeter. For example, if you attempt to apply a 480-120 MT with a 120v meter scaled for 480v the meter would read way higher when 480 is applied to the primary of the MT transformer.
With a distribution transformer they would be a true 4:1 ratio. One of the required ANSI/NEMA/IEEE factory production tests is a turns ratio test.

 

[chris kennedy]

Chris, are you going to corner ground the delta?

Yes.

And with the wye primary, XO would have to be insulated.

Insulated or isolated? I'm not using XO.

 

[LarryFine]

Insulated or isolated? I'm not using XO.

The wye connection of the windings should not be connected to anything else (isolated) and the junction should be insulated.

So, both.

 

[chris kennedy]

The wye connection of the windings should not be connected to anything else (isolated) and the junction should be insulated.

So, both.

You mean the XO lug and bus? Why?

 

[chris kennedy]

My understanding is that engineers do not like corner grounded delta 480 as it stress's the insulation much more than 480Y would.

Never heard this. Any further info?

 

[LarryFine]

You mean the XO lug and bus? Why?

As I've learned here, a wye primary should be wired as a Delta load.

The POCO does not do it, but it may be different for three 1ph units.

 

[jim dungar]

The POCO does not do it, but it may be different for three 1ph units.

The nuetral point of a wye on a MV distribution system is treated differently than a LV one (i.e. MGN for an example).

 

[mivey]

100216-2218 EST

mivey:

You know the following, but for clarification to others of when I was writing above and not thinking I submit:

Suppose I compensate the turns ratio for an internal 5% voltage drop in the transformer at full load, then the approximate ratio is 2/1.05 = 1.9/1 . In its step down use from 240 the no load open circuit voltage is 240/1.9 = 126 V.

When this transformer is run in reverse, then 120 is stepped up to 120 * 1.9 = 228 V. Now assume a 5% drop at full load and the secondary output is 216 instead of the desired 240 V. 10% low as should be obvious.


templdl:

Is it really true that on a large power transformer the leakage flux characteristics are so greatly different that there is a substantial difference in inrush current on a 1 to 1 turns ratio depending upon whether an inner or outer winding is the primary? I would expect more difference from wire resistance if the same diameter wire was used on both windings.

.

Just for fun, I made some manual calcs using some typical internal impedances to see what might happen for transformer about 50 kVA in size and rated 240/120 volts.

For the calcs, I am going to assume the ratio is adjusted to compensate for losses. To get rated output at rated input, the ratio moves from 2:1 by 98.28% to a 1.966:1 ratio. This means the no-load secondary voltage is 122.0 volts. There is a 1.30 degree phase shift from primary to secondary at full load (resistive load).

For a step-up transformer: To get rated output at rated input, the ratio moves from 0.500:1 by 98.28% to a 0.491:1 ratio. This means the no-load secondary voltage is 244.0 volts. There is a 1.3 degree phase shift from primary to secondary at full load (resistive load).

But what if we kept the step-up ratio at the same as what we had for the step-down case and keep the same impedances?

In the reverse-connected step-up case, a 120 volt source now supplies the losses. The ratio was set with the 240 volt side supplying the losses so the step-up ratio is 0.509. Now the no-load voltage becomes 235.7. Using the rated resistive load from before yields a secondary voltage of 231.9 and a 1.26 degree phase shift from primary to secondary.

Even using 122.0 volt source only gets the secondary voltage up to 235.7. We would have to supply 124.21 volts to get a 240 volt output.

 

[hardworkingstiff]

It sounds like the best plan would be to purchase a transformer wired as a step up transformer instead of using a step down in reverse.

 

[mivey]

It sounds like the best plan would be to purchase a transformer wired as a step up transformer instead of using a step down in reverse.

I wouldn't say that for all cases. For a normal step down, you would have to apply more than rated voltage at the primary to get rated secondary voltage at full load. My "for fun" calculations were based on the assumption that we were tweaking the windings to get rated secondary voltage using rated primary voltage at full load.

 

[mivey]

I wouldn't say that for all cases. For a normal step down, you would have to apply more than rated voltage at the primary to get rated secondary voltage at full load. My "for fun" calculations were based on the assumption that we were tweaking the windings to get rated secondary voltage using rated primary voltage at full load.

For example, using a 50 kVA transformer but instead let it be a step up from 120 to 240 volts. At full load, the secondary voltage would be 235.9 volts.

We would have to apply 122.1 volts on the primary to get full voltage at the secondary. About the same as what we would have if we used the step-down in reverse.

 

[templdl]

With distribution and power transformers there is no winding compensation. A ratio is a ratio. The only time compensation is considered is when a customer specified a specific % regulation at full load. If a winder accidentally does not count the turns correctly, such as adding or missing one turn, it shoumd up like a sore thumb and the transformer is scrap.

 

[mivey]

With distribution and power transformers there is no winding compensation. A ratio is a ratio. The only time compensation is considered is when a customer specified a specific % regulation at full load.

That is why the secondary voltage will be below rated if we apply rated primary voltage on a full load. Our primary distribution voltages tend to be elevated a few percent above nominal anyway except maybe at the end of a long tap.

For places where we do have lower primary voltages, we have voltage regulators out on the line. For the bigger transformers, we have primary taps that we can change to compensate.

If a winder accidentally does not count the turns correctly, such as adding or missing one turn, it shoumd up like a sore thumb and the transformer is scrap.

That is odd. For a missing or added turn, I would have expected an abnormal voltage but not a catastrophic failure. Sounds like there may have been other contributing factors.

 

[chris kennedy]

Update:

The owner of the company I work for who first proposed this install is now getting the transformer that was speced. Thank you all for your help.

Please carry on with this most interesting discussion.

 

[templdl]

Did I say catastrophic failure??
The core is cut the coils are wound, Part of the core is stacked and the coils slipped on to that part of the assembled core. After that the rest of the core is assembled to captivate the coils. The primary connects are made, the secondary connections are made; the core and coil assembly is dipped in varnish and baked.
As an assembly of primary and secondary coils a turns ratio can now be done.
What happens if it fails the turns ratio test?
Back in the oven to heat and soften the varnish, while still hot the top laminates must be removed by hand after which the coils can be removed.
A new coil must be made from scratch as the defective on can not be repaired.
Can it be determined that it is a huge waist of time?
As I said, the core and coil is scrap. Nobody got hurt, the coil didn't short circuit and smoke, explode or anything else. It was simply incorrect and could not be economically rebuilt.
There was no catastrophic failure.

 

[gar]

100219-2259 EST

templdl:

Wouldn't it be possible to check the turns ratio before assembly, an air core LR test, and after assembly, but before varnish or whatever, at a much lower voltage that would not be a problem without the added insulation?

.

 

[mivey]

Did I say catastrophic failure??...

No. You said: "it shoumd up like a sore thumb". I have no idea what that means. I have heard "sticks out like a sore thumb" but "shoumd" looks nothing like "sticks" and "up" does not resemble "out".

I read your post a couple of times and thought of several phrases and settled on a combination of "swelled up" and "shroomed up" (short for mushroom).

Some clarification was definitely needed. Thanks.