Transformer secondary conductors are un protected to the first over current device. Is there any section requiring the use of a ground bushing at the transformer and panel or disconnect? Should they be treated similar to service equipment as far as bonding is conserned? I learned that this was a requirement years ago as a cub, and have always done it, but I can’t find anything to support it now.

Nick,

Ref. 2002 NEC

Transformers are considered separately derived systems per art. 100 definitions and should be grounded as such. See 250.30 for the particular requirements of the bonding jumper, grounding electrode conductor, grounding electrode conductor taps, grounding electrode, equipment bonding jumper size, and the grounded conductor.

Hi Nick,
The flex between the XFMR and panel is not an equipment grounding conductor (250.118) and the flex connectors aren't listed as a grounding means either (unless they are marked "GRND"--per UL Electrical Construction Equipment Directory) so you need bond bushings to bond the flex. If given an either/or choice, I bond the bushing that is at the panel end because that is where the fault curent has to go to clear a ground fault (assuming the neutral is bonded to the equipment ground at the panel). The link doesn't have to be bonded at BOTH ends, like we do for any metal raceway enclosing a grounding electrode conductor, because any fault condition has current traveling both ways, thus there is no "choke effect".
You also might be semi-remembering from your 'cub' days that when the circuit voltage is over 250 volts you have to use bond bushings too(250.97),this would probably apply to your primary side.

Thanks for the replies. A couple of things though. There is nothing that says you have to use flex to connect a transformer. In fact, is not allowed to enclose the GEC.

Fault current is not going to stop at the panel. It is going to flow back to the transformer regardless of where the bonding jumper is. Under a fault condition most of the current is going to flow on the metal raceway. For this reason I think it is a good idea to put ground bushings at the transformer and panel but I don’t think it required.

Also, for circuits over 250V to ground the requirement to use ground bushings only applies when concentric or eccentric KO’s are encountered. Transformers usually do not have these.

I looked through some more literature and still don’t see that they are required do to the nature of the conductors involved. Maybe it’s one of those things you learn when starting out and eventually find out it’s not true. I should write a book of all the things people have said to me 'are in the code' that simply are not. Might be interesting.

Nick, you might take another look at 250.97 to see that you need a bond bushing for the primary side of the XFMR.
As for the 'transmission of vibration or noise' issue that requires one to use flex, that info is usually part of the transformer installation instructions.
Assuming you use an equipment grounding conductor on the primary side, which also connects through the bond bushing lug, fault current on the primary side will conduct through the low resistance grounding conductor and eventually back to the service neutral to close the primary side hot-to- neutral loop and trip the breaker feeding the transformer (hopefully).
On the secondary side, the equipment bonding conductor going from the transformer case to the equipment ground bus in the load side panel, if connected to the bond bushing at the panel end, will conduct fault current to the equipment ground bus, through the main bonding jumper of the separately derived system, on to the grounded conductor (neutral), back through the flex to Xo and complete the hot-to-neutral loop. This enables the fault current on the secondary to create an overcurrent situation on the primary and thus trip the supply breaker of the transformer. I put the bond bushing at the secondary panel end of the flex to shorten the fault current path and also to avoid having the choke effect. Don't let the info you undoubtedly have about 98% of fault current traveling on the raceway enclosing a grounding electrode conductor get mixed up in this scenario. If you connect the equipment bonding conductor at a bond bushing at the load panel end you won't have any net magnetic fields to create that choke effect.

[This message has been edited by Elzappr (edited 01-12-2002).]