Was just discussing this section of code with a coworker and we noticed it says "for circuits over 250v to ground"
Since the PV source circuits and output circuits are ungrounded their voltage to ground is 0.
We are still installing bonding bushings, but this got us thinking about it and I wanted to see what others thought about this and are doing.
Anyone omit a bonding bushing because of this verbiage and end up having a discussion with AHJ about it?
In general, a ungrounded transformerless system will "float" with each polarity at approximately half of the operating voltage on nearly equal and opposite sides of ground. Bi-polar systems work on a similar principle, except the difference is that they are center-tapped with a grounded conductor rigidly constraining the system to be half above and half below ground. Ungrounded/transformerless/non-isolated systems will not be as rigidly linked to this trend.
Because there isn't a firm guarantee of where ground will be, you need to treat it just as you would treat any ungrounded AC system regarding the "voltage to ground" figure, whether it is for work space or for bonding practices. So suppose one polarity is unintentionally grounded somehow, what will the voltage be on the other? Also, you need to use either the cold VOC, or maximum system voltage, rather than VMP or operating voltage.
I see this as a situation in excess of 250V to ground, where you probably should be installing more substantial bonding means when required by 250.97. Remember, if you use up all the ring knockouts, or if you don't have them in the first place, you can still use standard locknuts for electrical continuity when this rule applies.
Mike Holt does a good job teaching the logic of 250.97. He gives the simple example of a fault through a 1 ohm of resistance, and shows how a 277V fault gives over 5 times as much heat generation as a 120V fault, straight from P=V^2/R. So to keep excessive heat from building up in a faulted ring knockout that isn't meant for a 277V fault, you simply provide an alternative path through the bonding bushing.
Then he gave an example specific to PV systems on the DC side, and realized that this is uncharted territory. This isn't a conventional situation where the source can provide kiloamps of fault current, and where voltage directly drives the current. This is a situation where the irradiance & cell temperature define an intimate curve relating current and voltage. With a short circuit, your voltage diminishes to zero. And in any case, short circuit current is limited by the configuration of modules itself, usually single digit amps per string. He admitted that applying this requirement to PV systems is stupid, but it is what we have to do anyway.