That would probably depend on the impedance of the ground fault.
In an ungrounded system, you do not get a "ground fault" . If one phase goes to ground you get nothing. Because there is no ground reference at the service to trip the protection. If a second phase goes to ground, you get a phase to phase fault.
This one isn't about corner grounded, or one leg accidentally grounded.
What I see is pretty much equal voltage to ground, about 280 volts, on all three phases of a system that I know for sure originates at an ungrounded 480 delta transformer secondary.
The only additional transformers I know of are from the 480 delta to a 208Y120. Maybe there's something else I haven't found yet.
I found the anomaly while investigating a chiller that was and still is indicating phase reversal, when it still has the required clockwise rotation phase sequence.
The circuit board on the chiller is expensive enough that I am taking extra steps to find what's going on before I throw somebody else's money at it.
I'll be looking deeper into this in the next week.
Im glad we're all on the same page here.:)
No new news on your anomaly?
The board on the chiller has been replaced, and is back to normal.
I ran a Fluke 1735 on the feeder to the chiller for 18 hours, set at a 5% event threshold, and came up with nothing but a rise in voltage to 497 for about a minute, that coincided with utility switching.
I looked into the two step down transformers on the delta, and there is no cross connection I can find. The one I was most concerned with is a three wire delta on it's high side. the other is two wire 480 single phase on its high side. There are other single phase 480X 120 and 480X24 control transformers here and there.
So, it looks like until I can get a shutdown of the system to isolate individual circuits, I've gone as far as I can go. I did isolate some circuits but not the entire system.
I saw something similar , maybe 30 years ago . It was in an underground transformer vault feeding an older section of a college . 480 VAC ungrounded delta .
I did not think it was a very good idea ( I prefer 480 / 277 wye systems ) . But that was the way it was and had been for quite some time . And no one asked me for my opinion .
So , there may be nothing " wrong " with the system . And , as far as NEC , it may be old enough that the system is " grandfathered " ?
You may have a grounded delta with a stinger. In this case one phase is center tapped and this is grounded and connected to neutral.
This usually results in one leg high in relation to ground and then the other two 120 volts to ground / neutral.
Phase to phase they will measure 240 and phase to ground and higher, but I forget how much higher to the "stinger" leg.
This is fairly common in old systems. They are also often unbalanced loads if there are significant 120 volt loads. These transformers are often larger on the two 120 volt windings to correct fot this.
This might be part of the answer.
It sounds as though you are talking about a 240 delta with a center tapped ground. The high leg (usually the B leg) would be 1.5 x the value to ground of either the A or C legs.
I see this most often with old 208 delta feeds in Philadelphia.
Where a 208 three phase delta is grounded using this method, A or C to ground will measure 120 volts, and the B leg will measure about 180 volts, because the B connection is associated with 1 1/2 of the secondary winding amount with which either A or C will be associated.