A solid tip of my hat to
@Dirk,
@Mdragger88 and others on this topic, many thanks. For my part I think Dirk nailed it!! And the referenced ecmag article further nails it, in my opinion. I too have gone down some of those cleaver DIY roads noted herein. I used a cheap 12 volt AC to adjustable DC converter (5 pack for $10) mounted in a weatherproof enclosure with the enclosure and all lights more than 5 feet from the pool with all the proper cabling. All powered from a code compliant 120V to 12VAC pool/spa rated transformer with the 120V primary side also protected by GFCI. But I will soon dump all the landscape lighting portions of that in favor of either the VOLT systems or better yet fibre optics with no possible conductivity. Here’s why...
It’s basically the argument made by Dirk and others which I think amounts to the Swiss Cheese Model of Accident Causation, also known as the “cumulative cause effect” and often referenced in aviation safety. How do I know that circumstances won’t change? Someone digs and nicks a wire, a future homeowner perhaps, and/or the wire deteriorates for other reasons, rubbing on something. And later the wire touches the cage, then somebody manages to touch that source, perhaps during a rainstorm. How do I know some voltage differential won’t occur, some day, somehow, inducing a small current through a human? As noted, GFCI on the primary does not trip for currents traveling on unintended paths from the secondary. Simply stated, none of that can happen if the lighting is non-conductive fibre optics (assuming the control modules are all properly installed & protected). Downside I guess is price. Incidentally it’s probably worth thinking about whether any of your 24VAC irrigation valves are near the pool.
In related matters, I read recently that there is also some controversy about whether the NEC requires (or some locales require) that even approved transformers and their enclosures have the primary protected by GFCI. In one of those “man dies swimming due to faulty pool light” articles, the cause was traced due to an approved transformer / enclosure with proper separation of primary & secondary wiring – but the enclosure ground wire separated from the metal enclosure and a loose wire in that enclosure managed to electrify downstream secondary components. The description needed more detail, but in the end GFCI on the primary would have saved the man. After reading that I checked my 10 year old installation and indeed the primary was not on GFCI, so I added a GFCI breaker and tested its function.
This leads me to the very clever and seemingly well-thought-out post by
@setsailsoon about using a proper weatherproof DIN rail cabinet, carefully and neatly wired for sure. In such an installation, is it rated for pool/spa use? Are all secondary wirings totally and physically separated from the primary wiring as done with the metal plate in approved enclosures? Is the transformer itself rated for proper primary/secondary isolation and pool/spa use? If going down that road, at least the primary 120V circuit should be GFCI protected, methinks. But I think it’s still possible to thwart the protections in a true swiss cheese model, even if I cannot think of how, right now, today.
All that said I’m thinking of running a couple tests. In and around my pool with all tis intrinsic bonding, is it possible for me to generate even the smallest of currents between the secondary of a 12VAC or 12VDC transformer and my water or cage or light niches or railings? And if I NEVER produce such a current, is it truly safe? Or have I just failed to conjure up the right combo of swiss cheese slice alignments? Maybe stick with fibre optics!
If I want to get truly radical maybe I wonder if my pool bonding was ever done right? I mean, I’ve tested for zero resistance between all cage and niche and railing components on back to the #8 bonding wire that snakes around the equipment pad and interconnects all equipment. But how do I trust that the concrete reinforcement mesh was properly tied in? It should have been inspected but we know that too is not foolproof and I wasn’t the owner back then. So I’m just wondering if there may be some clever way of testing the bonding. Or maybe I simply need to trust, at some point, that I’ve done all I can!
Peace to all, and thank you.