Replies

1. 
   khaag | Oct 02, 2012 05:18pm | #1

   Thanks DanH,

   I've got the floor soaking now.  Will be interesting to see what happens.  Think I'm going to put a millimeter on the ground return when I do my wet check later in the week.  When I measured the 60K with the ohmeter open circuit the value climbed like it was measuring capacitance, for whatever that's worth.  There was obviously enough leakage to trip the GFCI so it's not simply measuring the capacitance between the ground and the conductors, there's more to it than that.

   BTW, the GFCI is built into the SunTouch controller, so is a special order (from Lowes, anyway), and not as "inexpensive" as a standard GFCI.  That's why I haven't just tried a new one.

   Kurt

2. 
   DanH | Oct 02, 2012 07:52pm | #2

   Actually, disconnecting the ground wire doesn't make things that much less safe.  The GFCI should still work as intended for people protection.  The ground is just a "belt and suspenders" thing.

1. 
   khaag | Oct 02, 2012 09:14pm | #3

   Yeah, I've been walking on the floor with wet bare feet and am still vertical.  I am somewhat worried about how a new homeowner might react (not that I'm planning to sell anytime soon) if they found that I disconnected the "safety" ground.  And it's interesting that the GFCI doesn't see an imbalance in current when the ground is disconnected.

1. 
   DanH | Oct 02, 2012 09:41pm | #4

   Likely the ground wire slightly shortens the "circuit" to ground, especially when the floor is dry.  The thing is clearly on the "hairy edge" anyway, so it's the straw that made the cookie crumble.

3. 
   Scott R. | Oct 03, 2012 12:34pm | #5

   Ground Fault troubleshooting

   Hello,

   You have a ground fault and it will need to be addressed.  The system should not be operated without the grounds attached.  A licensed electrician can use a variac to send a very small amount of AC down the ground and the wire that has continuity with the ground.  What will happen is that he may be able to heat this "shortened circuit" and see exactly where the fault is located in the floor.  To do this, you will need to use a thermal camera.  With the camera you will be able to see the section that is heating.  Where the heat ends is where the fault is.  You can then carefully remove the tiles above the area and repair the wire.  

   The only person doing this testing should be a licensed electrician.

   Just so you know, I work for WarmlyYours.

1. 
   DanH | Oct 03, 2012 08:52pm | #6

   With all due respect, you're wrong on several counts.  With 60K resistance you can't send enough current down the line to heat anything (you get 0.7 watts at 200v).  And even if you could, you could not dig up the cable and make a reliable repair.

   The thermal imaging can be used to find where a heating cable is broken/cut, but in this case presumably the entire cable is heating properly, it just has some "leakage".

1. 
   khaag | Oct 04, 2012 07:55am | #7

   That's part of what makes this tough to troubleshoot.  I agree, the standard diagnostics aren't likely to reveal much.  I kind of like the idea of the variac, but between hot and ground to see if I can burn out the 60k leakage path.   Don't have one, but maybe a good electrician would?

1. 
   DanH | Oct 04, 2012 08:49am | #8

   Note that the minimum trip current threshold for a Class A GFCI is supposed to be 4ma.  120V through 60K ohms would produce a peak current of about 3ma (taking into account that the peak voltage of 120VAC is about 170V).  So you're right on the hairy edge

2. 
   Scott R. | Oct 04, 2012 12:50pm | #9

   Service techs

   Hi there,

   Where are you located?  There may be a tehnician located near you that has worked on your type of wire.  If you would like to contact me, you can email me at:  srosenbaum AT warmlyyours DOT com.  I will be glad to help if I can!

2. 
   Scott R. | Oct 04, 2012 12:52pm | #10

   RE: All due respect

   Thanks for the reply!  Just an FYI, I have repaired dozens of floors this way.  And an FYI, an experienced technician can and does make a reliable repair by carefully removing the tile/s above the fault.  Once the tile is up, the thermal camera will display exactly where the bad spot is and the tech carefully removes the thinset around the area like an archeological dig.  The cable is repaired with a splice or jumper sealed in a waterproof, heatshrink tube.

   If the ground fault connection is not sufficient (many times it is sufficient to set up a circuit), a technician can use a Thumper or hi-pot to arc the gap and fuse the ground and line together.  That's when you get a really good connection and you can start getting that repair done.

   I have personally repaired broken heating wires in asphalt, in outdoor stairs under limestone caps, etc.  We need to make sure people know that a wire damaged during the installation does NOT mean a floor has to be replaced.  It just needs to be diagnosed and repaired by a professional.  These repairs are not inexpensive.  The equipment needed to do this is very expensive and quite specialized.  Even with this cost, the cost of the repair is usually a 10th of replacing the floor and incredibly faster.

   By the way, I repaired 2 floors using this technique just last week, that's why I got a kick out of your post.  I have also personally watched 2 other professionals repair other brands of heating cable doing this exact method.  These types of repairs are done hundreds of times a year, by many experienced technicians.

1. 
   DanH | Oct 04, 2012 08:02pm | #11

   Let's just say I'm skeptical, on several counts.

1. 
   khaag | Oct 07, 2012 08:35pm | #12

   Update

   Ok, I wet the grout lines for a few days, then checked the mAmps running through the safety ground.  My cheap Craftsman multimeter said 0.2 mA - way below the GFCI threshold.  It's autoranging, and read DC mA, which obviously indicates something amis, but I couldn't find a way to force it to AC.  Figured I'd go back to reading resistance between neutral and ground.  Still 60 K as before.  Reading was climbing so I let it go for awhile.  Reached 200 K and still climbing.  But it had been tripping the GFCI so something more must be going on??

   Maybe capacitance between the ground and neutral with the long parallel run of wires is causing the meter reading to climb, and allows a surge current via the ground when the mat cycles on?  Tried curling the wires into a few loops along the shaft of a screwdriver to make a bit of an inductor.  Put everything back together and no trips for almost two days - weird but good!  Maybe my little inductor helped?  Maybe the dampness reduced the capacitance?  Maybe it'll trip again now that I've jinxed it by writing this??  Don't know, but I do appreciate your interest and input, and I'll let you know how it's working in a few days...

4. 
   khaag | Oct 14, 2012 08:17pm | #13

   Still No Trips

   Going on a week now and no trips.  Could my coiling the wire have really made the difference?  Or just random rearrangement of the wires in the utility box reduced coupling??  Thanks agin for all the comments.

5. 
   khaag | Oct 22, 2012 08:48am | #14

   Still good

   Stay away ghosts!

1. 
   DanH | Oct 22, 2012 08:05pm | #15

   I predict it'll trip on the 31st.