GFCI rearing it's ugly head - with a twist. Help!

[jtomason]

So, the show this year is once again tripping the GFCI outlet after 3-3.5 hours of running.   So, I break out the handy current leakage meter and start looking.   I've isolated it to the LOR 50 watt flood that is peaking at right around 3ma of current leakage during peak strobe events and .5ma of leakage current at idle.   I should note at this juncture that there hasn't been rain since everything has been set up and it's all bone dry.   

So I check the power cable.  I lift it up, no change.  Replace it.   No change.   Unplug the CAT5 from one leg - no change.   Unplug the CAT5 from the other leg - drops to zero.   Huh?!

So now I trace that back.  It goes through 3 LOR AC controllers before going inside to the laptop.   Nothing really changes until I unplug the CAT5 from the RS-485 - and then it drops to .2ma.   

None of this makes sense.   I'm hoping that someone can shed some light on this.  

 

Setup:  

Laptop - USB - RS-485 - CAT5 12' under deck - LOR AC - CAT5 8' on deck - LOR AC - CAT5 40' on ground - LOR AC - CAT5 40' on ground - Floodlight ------- More CAT5, other LOR boxes, etc. 

I've made sure that CAT5 is not running close to power lines, and where they all meet it's as short and perpendicular as possible.    But hey - that's noise troubleshooting - shouldn't have any bearing on leakage current...  I even tried elevating the CAT5 off the ground on the theory that it was some wacky capacitive coupling issue - but no change.  


So basically my brain is fried and unless a revelation comes from you fine folks, I'm gonna have to pull the floodlight from the show for now.    

Anyone have words of wisdom?

 

Thanks, Joe

 

[k6ccc]

It's stuff like this that makes me glad that almost everything in my display is RGB.  The only AC props I have year round are all on permanent house wiring, and for Christmas, two 11 channel singing faces that are hanging from the roof (on rope).  Not much of a GFCI trip there..

Now Joe to your issue.  That's a strange one.  See what your meter shows with the USB to RS-485 adapter adapter unplugged from the computer  As a test (since you can do this with a laptop), see what your meter shows with the laptop unplugged.  If the power source for the laptop can be polarity reversed (some can, some can't), try that. 

I have only had a GFCI trip twice (never during Christmas).  Both times it was totally legit.  I have a 12V 29A power supply in one of my brick columns in the front yard that powers most of the year round landscape lighting.  Brick is not waterproof by itself and twice when I had not done the annual water sealant on the brick, water leaked in and of course poured right onto the power supply.  Fortunately the water resulted in the GFCI tripping before any real damage was done.

 

[TheDucks]

1 hour ago, jtomason said:

So, the show this year is once again tripping the GFCI outlet after 3-3.5 hours of running.   So, I break out the handy current leakage meter and start looking.   I've isolated it to the LOR 50 watt flood that is peaking at right around 3ma of current leakage during peak strobe events and .5ma of leakage current at idle.   I should note at this juncture that there hasn't been rain since everything has been set up and it's all bone dry.   

So I check the power cable.  I lift it up, no change.  Replace it.   No change.   Unplug the CAT5 from one leg - no change.   Unplug the CAT5 from the other leg - drops to zero.   Huh?!

So now I trace that back.  It goes through 3 LOR AC controllers before going inside to the laptop.   Nothing really changes until I unplug the CAT5 from the RS-485 - and then it drops to .2ma.   

None of this makes sense.   I'm hoping that someone can shed some light on this.  

 

Setup:  

Laptop - USB - RS-485 - CAT5 12' under deck - LOR AC - CAT5 8' on deck - LOR AC - CAT5 40' on ground - LOR AC - CAT5 40' on ground - Floodlight ------- More CAT5, other LOR boxes, etc. 

I've made sure that CAT5 is not running close to power lines, and where they all meet it's as short and perpendicular as possible.    But hey - that's noise troubleshooting - shouldn't have any bearing on leakage current...  I even tried elevating the CAT5 off the ground on the theory that it was some wacky capacitive coupling issue - but no change.  


So basically my brain is fried and unless a revelation comes from you fine folks, I'm gonna have to pull the floodlight from the show for now.    

Anyone have words of wisdom?

 

Thanks, Joe

 

Pin 6 in the CAT5 cable is Ground (to the +9 on pin 3).  You have the hated ground loop. That means the there is a device or outlet with less than stellar grounding (green).

💡 Leave the network connected and one-by-one unplug the power cords feeding the AC controllers, monitoring the leakage

[jtomason]

45 minutes ago, TheDucks said:

Pin 6 in the CAT5 cable is Ground (to the +9 on pin 3).  You have the hated ground loop. That means the there is a device or outlet with less than stellar grounding (green).

💡 Leave the network connected and one-by-one unplug the power cords feeding the AC controllers, monitoring the leakage

So you are saying that one of the controllers has a bad connection on Pin 6 on one of the cables?   When I disconnect the CAT5 from the flood, it drops to nothing. 

With the flood disconnected (just coupled the CAT5 runs on either side together) and monitoring the show tonight, so far only 1.78mA max leakage. 

 

[LaughsBrightly]

Just learned something. I need one of these meters. Joe, what meter are you using?

[TheDucks]

19 minutes ago, jtomason said:

So you are saying that one of the controllers has a bad connection on Pin 6 on one of the cables?   When I disconnect the CAT5 from the flood, it drops to nothing. 

With the flood disconnected (just coupled the CAT5 runs on either side together) and monitoring the show tonight, so far only 1.78mA max leakage. 

 

No Pin6  is how logic grounds are being connected (which is normal). Loops happen when the 3rd pins are not all really grounded. Sometimes called a HRG (High resistance Ground)

You have 1 or more bad players attached to that network. It could be bad extensions. Outlets that have no bare/Green wire back to the Panel. These are SOB's to find

[TheDucks]

14 minutes ago, LaughsBrightly said:

Just learned something. I need one of these meters. Joe, what meter are you using?

ONE 🙃 

A simple Autorange DVM  ($20) for basic voltage, continuity and DC current (<10A),

A Killawatt power meter for AC consumption and Power Factor  (Cheap supplies do not have PFC circuits. it can add up. all you can do is break up their loads across different GFCI)

Clamp on DVM most do AC. A few do DC as well.  While it works, I found the controls on the $90 Harbor Freight fussy. Gimme separate buttons rather than punch until the Mode display says what you want)

Flukes are top of the line, but bring a fat wallet (most lighter do not need this good).

FWIW I learned on a Simpson 260 (the original) Analog and used that (for 30 years) until forgotten Batteries leaked and corroded stuff

[dgrant]

16 minutes ago, TheDucks said:

 

FWIW I learned on a Simpson 260 (the original) Analog and used that (for 30 years) until forgotten Batteries leaked and corroded stuff

For me, it was a PSM-6 then a Radio Shack, analog something, which I still have!

[TheDucks]

24 minutes ago, dgrant said:

For me, it was a PSM-6 then a Radio Shack, analog something, which I still have!

PSM6? They were only up to PSM-4 when I was in.

The 260 was my dads (I was 8), they were up to 260 5 series when I was in. (I ran a meter qualification/repair bench for a while)

[jtomason]

1 hour ago, LaughsBrightly said:

Just learned something. I need one of these meters. Joe, what meter are you using?

I'm using this one: 

https://www.aemc.com/store/products/clampon-meters/clampon-meter-565

 

[jtomason]

57 minutes ago, TheDucks said:

You have 1 or more bad players attached to that network. It could be bad extensions. Outlets that have no bare/Green wire back to the Panel. These are SOB's to find

...As opposed to the RGB flood that I disconnected?  Or are you saying that someone is likely wrong with the flood?

[TheDucks]

1 minute ago, jtomason said:

...As opposed to the RGB flood that I disconnected?  Or are you saying that someone is likely wrong with the flood?

the flood has 2 thing in this picture. A power cord and a LOR network cable. Open the back up. You might have WATER. The next choice would be Line filter components (between the cord and the PSU section)

While in there look for smoked areas, it might have just flashed over and left residue

[jtomason]

So last night the flood was totally disconnected (CAT5 and power) and the show ran fine.    

Today I plugged the CAT5 in but did NOT plug in the power.   The GFCI started tripping about the same 3.5 hours into the show. 

Tomorrow I am going to just sanity check the SPT-1 extension cables since there are a lot of them and make sure that none have any damage.   I have a 16 channel LED light strand mega tree, so I may consider elevating its extension cords off the ground if I can't find anything else to correct. 

Just spitballing here, but how many AC devices do you folks tend to run off one outlet/circuit?   Maybe I need to share the wealth a bit.   I am already on 2 outlets on different circuits.   Starting to think that I will need an electrician and a Christmas sub-panel and dedicated circuits/outlets.    For those of you with larger displays, what electrical work have you had to undertake as your show got bigger?

Also, does anyone have a good source for GFCI / current leakage troubleshooting for Christmas lights?  Just about everything I have seen just deals with moisture, which is not a factor here. 

 

Thanks all...   Joe

 

[TheDucks]

5 hours ago, jtomason said:

Just spitballing here, but how many AC devices do you folks tend to run off one outlet/circuit?   Maybe I need to share the wealth a bit.   I am already on 2 outlets on different circuits.   Starting to think that I will need an electrician and a Christmas sub-panel and dedicated circuits/outlets.    For those of you with larger displays, what electrical work have you had to undertake as your show got bigger?

Also, does anyone have a good source for GFCI / current leakage troubleshooting for Christmas lights?  Just about everything I have seen just deals with moisture, which is not a factor here. 

 

Thanks all...   Joe

 

If you are not branch current limited.: just convert a Single GFCI to 2 GFCI.  For indoors, the make a 2 gang box with a single gang opening in the back. That screws in place of the original outlet (additional wood screws recommended on the Stud side). I did this in my garage so the Freezer has its own GFCI (I put a child cap in the other hole) as the 2nd one runs my 'corner' controller

I use timers because Idle (off show hours) power per location was 16-26W (no lights, just the smart/dumb PSU's). When the timers turned things off, the GFCI would trip (reactive collapse current) if there were 3 units. So I spread the leakages over 5 GFCI (but only 3 branch circuits).

As to testers?  You have to be careful. You don't want to use a Megger (the tool, not the brand) as it hits 600V and will kill surge devices.  A DVM with a ohms that reads in the 1-2 MegOhms rang will  show resistive leakage.

If you have AC (capacitive) leakage, I have no real idea.

BTW that flood is probably full of moisture or corroded. The metal stake creates part of a galvanic (soil acts as a battery) loop to the power panel ground

[jtomason]

I opened the flood up - looks perfect.   This is the 50 watt flood - no stake, it has a bracket.  I have it sitting on a plastic riser on the lawn, and the power supply is zip-tied underneath that plastic riser.  Both look pristine.    I am starting to believe that this is more about having too many controllers on one circuit and internal component leakages are adding up and tripping the outlet rather than there being ground leakage. 

What I still don't get is why this happens 3 or 3.5 hours into the show.  Maybe the capacitive components have heated up and increased the leakage by then?

Sunday I had the flood disconnected entirely and the show ran OK.   Last night I left the CAT5 runs connected but the power DISconnected and the outlet tripped and wouldn't stay reset for more than a minute. 

Tonight I'm going to remove it entirely again and see what happens.  

 

What I may have to do is split the load on that circuit - just too many controllers, even if the amperage is low enough.   

 

[hasslerk]

In my case, I run 14,766 LED lights off one duplex plug on the front of the house, which is a combination of AC controllers, all my pixel controllers, some DMX, dumb RGB, and 2 LOR floods.  Another 4000 LED mega tree off another duplex by the garage, but both these sources are behind the same GFCI in the garage.  I have another 1600 lights down the driveway but they run off a circuit from the Carriage House (using an ELL and AC controller).  Under normal conditions I have no issues.  But when it snows, the GFCI will trip, which means I have water entry somewhere (probably from an exposed female pigtail, but this is a different problem than yours, but it shows my GFCI actually works).  I run my entire show at 30%, except for the floods.

...and I keep meaning to buy a meter.

[Wayne K]

21 hours ago, jtomason said:

Just spitballing here, but how many AC devices do you folks tend to run off one outlet/circuit?   Maybe I need to share the wealth a bit.   I am already on 2 outlets on different circuits.

I have 24 a/c controllers running off of 8 gfi plugs around the house. Since I am all led's, it is not so much the load as it is the gfi problem once in a while. I also try to group things together on 1 gfi. Example my 2 controllers for the singing trees are on 1 gfi, My 2 mega tree controllers are another gfi and so on.

Wayne

Edited December 8, 2021 by Wayne K