Jump to content
Light-O-Rama Forums

GFCI circuit TESTER


Recommended Posts

Posted

I am thinking about making a tester for testing circuits attached to a GFCI so I know which one is the problem.

When you have 50 channels or more tied to a GFCI and it trips.. then what? Which is the one causing the problem? Do you manually figure it out? What a pain that is. I am tired of the guess work. I have some trees that have 1800 lights on them. With a small amount of leakage current from each string do they add up and cause the trip, or is it one string in particular??? I need a way to tell me what I want to know.

Anyone interested? Not sure of the cost, the size or even if I can do it like I want.. but if you are interested let me know your thoughts.

Craig

Posted

In general, nuance trips like that are caused by the accumulated loss.

Posted

In general, nuance trips like that are caused by the accumulated loss.

Yes they are.. and there is no way to be sure that I am aware of. When you put 18-20 strings of lights on a big tree each string can have a very minor amount of leakage current when the whole thing is wet.. or it could be just one string.. so how do you troubleshoot it??

For that matter, with LEDs there are so many controled circuits you can put on a single 20 amp power circuit that it is very time comsuming to track down. ie... I have 8 16 channel controllers on a single outlet. All channels driven.. And who wants to do that in the driving rain?

First you have to narrow it down to which controller, then which channel.. then zero in on the offender.. this is only for small circuits.. for the scenario with the large trees that is not so easy.. Some of the strings are way up on the air...

I want to make some sort of low cost plug in device with a threshold that has non volitile memory and will show you when a specific amount of leakage current has been exceeded. Something about 2" square at max. At least that is the first thought.. Looking for others to input what they might want to see..

Posted

My point is that it's usually an accumulation. No one channel is to blame, they all are.

If you REALLY want to build a device, you are going to have to build a super-sensitive meter that can measure to micro-amp resolution the difference in current going out the hot vs back in on the neutral for a 20A total load.

(If I'm reading the specs correctly) this Benning CM9 does what you want, on 1 lead at a time. But it will set you back about $460 + shipping.

http://www.test-equipment.com.au/current-leakage-clamp-meters/benning-cm9-digital-leakage-current-clamp-with-1ua-resolution

Posted

I actually have testers that will do it.. but one circuit at a time.. like that one you referenced..

So my dilema is that if the accumulation is the issue but none are to blame individually.. then what? Code doesn't allow for non GFI circuits outside.. unless it is all low voltage of course..

If I were to put detectors on at all the channels at one time... something small and pluggable.. I could track down whether or not one particular channel is the offender or whether they all accumulate together.. If the problem is accumulative I would have to break them up to separate GFIs.. but if there is a single offender or two.. I could deal with it or them by relocating those sensors to the strings involved...

Perhaps I am just over thinking it.. a gang of GFIs would be cheaper and solve the problem easier perhaps... just take up a lot of space... perhaps that is the way to deal with it.. Any one else? Ideas are a good thing!

Posted

I'll jump in here...

Bleed currents can and do accumulate to amounts above the threshold of the GFCI trip level. For those that are curious, a GFCI device compares the current on the hot and neutral, and in the event that there is fault current above the trip threshold on either wire, the device opens the circuit (trips). This presents a challenge for keeping a display lit in wet conditions.

The only way of reducing and hopefully eliminating GFCI trips is keeping the circuits isolated from ground and other circuits (fed from other GFCI devices). In order to achieve this, the display elements would have to be non-conductive to ground (i.e. non-metallic) or the light string supports isolated from any conductive framework. Additionally, extension cords (read: plug ends) laying on the ground also need to be isolated for trouble-free service from a GFCI. Cords that are ran in areas that are prone to puddling with water can be troublesome considering that a small pin hole in the cord's insulation is enough to trip a GFCI.

Some ways that I have sought to isolate my strings from a pathway to ground have been to use rubber hose where a string loops over a metallic hanger (mega tree) and allow the light string to hang on the rubber rather than the metal. The mega tree is grounded with a driven ground rod (thunderstorms can visit us from time to time). Extension cords have been lifted from the ground with non-conductive materials (PVC pipe & tie wraps/zip ties). Display elements have been made from PVC where able. And there aren't any power or comm cables routed through areas prone to collecting water or subject to being drowned during rainy weather.

In addition, design the power system for quick identification and isolation of a faulty circuit. Your bank of GFCI devices is the ticket here. Each of my controllers is served by 2 GFCI receptacles (each pair of GFCIs fed by a single 20A circuit from the panel), one GFCI for each bank of channels. My entire display runs on 2 20A circuits. In the event of a trip, I would know immediately which eight channels to put under scrutiny. This will reduce tracking time considerably.

In the event a GFCI trips, the eight channels being served and quickly be cut in half to either identify or eliminate the problem channel. For example, the GFCI is tripped, so isolate/unplug channels 1-4 and see if 5-8 trip the GFCI. If they don't, then add 1-4 one at a time until you find the offender.Should channels 5-8 trip the GFCI, then isolate those and see if it holds 1-4. Once you narrow it down to the channel, then you have to take a look at the load served by the problem channel. Hopefully it is a simple correction from there, but if not leave that unplugged and let the show go until the appropriate corrections can be made.

As for incorporating the ability to monitor all of the channels, my humble opinion is that would be mondo expensive, overly complex and another headache that isn't needed during the season. While it would be very cool, I am thinking it isn't practical for 99% of the folks that are decorating. A GFCI receptacle runs around $12 at the home improvement centers, so for $24 the controller could be served by two separate GFCI devices (one per bank) and I would almost wager that a system to monitor all 16 channels (or even the two feeds) would cost more that the GFCI devices. YMMV

If you do look into building up one of the systems, do keep me/us informed. That would be a fun project to watch unfold, practical or not.

Posted

I had my first experience with driving rain - and mostly won... I had each controller plugged into a seperate GFI, so when #3 tripped out, the problem was isolated to at least those channels... What I did find ultimately was that when it is driving rain - don't fight it - The next day in examining the icicle led's, I found that a couple of the plugs/socket joints had been left over the top of the gutter - and showed signs that the roof water was falling directly into them during the rain... I pulled those out and left them hang over the front - away from any water path.

I also checked all the other channels on #3, blowing out water in the plug and making sure they are protected from direct water. I used foam split pipe insulution and left the split open on the bottom as sort of a rain coat.

Best to do is isolate the controllers to seperate GFI breakers, and when you have a fault you can know which controller it is. I think you could also double up and put in two GFI breakers per controller isolating it to only 8 ckts.

When I was troubleshooting I found my controller channel assignment tables were very helpful. (Excel) It allowed me to know the use and loads by controller and channel on a page I carried with me.

Posted

I kind of answered my own moment of weakness by saying the bank of GFIs would be much cheaper and perhaps the way to go... My problem is someone different I think.. I have 9 trees that are about 30-35' tall. Each tree takes about 2k lights. I break those up coming down the tree to about 3-6 different string feeds.. Two or three can be on one extension cord dropped to the bottom. Then all are combined and we have a single channel for a single tree. My problem is that the lights will run for a bit, then trip the GFI... troubleshooting is a pain when it won't do it continuously.. The only path to ground for me is litterally thru the tree itself or the wetness of the bark... not so much the latter. Everything is two wire until it hits the controller.. Even if I have a single tree narrowed down.. I have 20 stings of lights to search thru to find the issue..

Anyway, it only happens when the rain is really bad.. or with high winds.. so I am thinking I am ok for now...

As for the project, we put a pencil to it and decided it wasn't worth the time and expense.. but we did noodle some cool ideas.... I have some really nice little Hall Effect sensors laying around.. ha ha!

Posted

Is this what you're looking for? http://www.idealindu...er_brochure.pdf

We call those meggers in the control business.. and I have a few of them... I never really thought about it before, but that might just be the ticket once I narrowed it down to which tree... it is exactly what we do with them for motors and such.. The only issue is that I would have to take every string apart to megger them, at least sections of strings... but still... Great Idea Ken... I might just give that a shot.

See, that is why I like to think "out loud" so to speak... someone will ring the right bell and my pea brain will shake loose!

Thanks All

Guest
This topic is now closed to further replies.
×
×
  • Create New...