160504-1501 EDT
Belknap24:
If you measured 6 A on each input line with the secondary unloaded, then I do not believe you have a transformer problem.
Some measurements from a small, 175 VA, transformer. Nominal input rating is 120 V. Calculated full load input current is 175/120 = 1.46 A.
No load measurements:
123 V input
0.19 A
4.0 input power
24.2 VA
0.16 PF measured, calculated 4/23.4 = 0.17
For your transformer measured unloaded VA per leg is about 277*6 = 1662. Calculated PF is 1662/15000 = 0.11 . Lower than my smaller transformer and to be expected. I used a VA ratio here instead of power but it gives us an estimate.
You should really use a power input measurement to the unloaded transformer. If your transformer efficiency was 98%, then you would expect an unloaded power loss per leg of about 15,000*0.01 = 150 W by assuming transformer core loss is 1/2 of full load transformer loss.
With a power factor of 0.1, then 150 W becomes 1500 VA. 1500 VA at 277 V equals 6.4 A.
Make sure all three legs read about 6 A with no load. Then your problem is in the load or how energy is being measured.
I have not proofread or checked my calculations.
..