Friday, August 31, 2018

Troubleshooting as an Away Game

As regular readers know, I'm a retired engineer.  Electrical engineering is a broad field so people tend to develop specialties.  Mine was radio frequency circuit design, more specifically receivers, phase locked loop frequency synthesizers, the automatic control loops commonly used in receivers and transmitters, and electronic filter design, from DC to about 14 GHz.  This is what I consider my home field; fixing broken radios, both transmitters and receivers, is my home field advantage.

For the past couple of months I've been playing an away game, troubleshooting in a specialty of electricity and electronics is so far from my specialty that virtually nothing I learned in my career is useful here, except the big picture aspects.

I'm troubleshooting an intermittent failure in my house's electrical power distribution.  Finding a ghost, if you will.

When we had the addition built onto our house for the shop back in 2014, the general contractor subcontracted to some electricians who wired the entire shop.  The AC outlets were divided into two branches largely along the west and east walls.  Each circuit is on a 20A breaker and each side had a ground fault interrupter.  Once or twice in the intervening four years, the west wall breaker would trip.  In the last few months, that started becoming more common and started to become a problem.  The east wall breaker has tripped much less often.  I can only really recall once, which was on a night a couple of months ago when both GFIs tripped. 

To be clear, nothing is completely broken, in the sense of not working.   I'm not completely sure, but I think the most often it has ever reset in a single 24 hour day has been twice.  Most of what I'm trying to troubleshoot is something that happens at random times no more than once a day.  It has gone weeks without tripping.

While I have general electronics troubleshooting tools; a few digital multimeters, and an oscilloscope, most of the rest is more specialized for radio testing.  I don't have any specialized tools that an electrician might carry.

I do have a lot of hours troubleshooting intermittent problems and know the general method is to try to make the rare failure happen more often.  Failing often would be good.  Breaking it permanently, for once and for all, would be better.  As a general rule, you try to figure out what's causing it to fail and do it more often.  If a radio or computer fails when it's cold, you spray it with freeze mist to find out which sections are sensitive to cold.  If it fails when it has been running a while and gets warm, you use a hot air gun.  If a radio fails during vibration testing, you tap it with a rubber mallet.

I've been unable to find something that causes it to fail.  The GFI has tripped most often overnight when no one is watching. It has tripped in the early evening during or after nearby thunderstorms.  It has tripped on a day with no storms.  It has tripped while I was making a part for my engine, thankfully when the mill was off and the cutting was done.  At one point, I thought it was fixed.  I went outside to check out the ground rod for the house, where the power comes into the house (we have underground utilities) and found the ground rod corroded away - if you look in front of the yellow tag, you can see the point on the ground rod. I had the ground rod replaced and the GFI stopped tripping for two whole weeks.

What I've ended up doing is spending time looking for stories about things that cause GFIs to act strangely and do things like this.  Someplace I read said that refrigerators or freezers shouldn't be plugged in a GFI circuit because ... some reason that sounded plausible.  So last Sunday I removed the east side GFI outlet and replaced it with a conventional outlet.  Then I took the east side GFI and replaced the west GFI with it.  Comparatively, the east side GFI was much less likely to trip.  It took two days for it to trip on the west side. 

One difference between the west side and east side of the shop is that the west side has three surge protected AC outlet strips.  A search for reports that combining surge protection and GFI might cause the GFI to trip more often said that the surge protected strips might be a problem.  They theorized that when the surge protector dumps current from one side the AC line or the other over to ground, it looks like an imbalance between hot and neutral, which is what the GFI is detecting.  In other words, when the surge protector does what it's there for.  I took one of the surge protection strips out of place and just plugged in only the things I was using.  Today is the second day and it hasn't tripped. 

I have basic knowledge of how these things work, and nowhere near enough experience to look at something that's bad and know where to start looking.   "Away game" is probably too lame a comparison.  AC house power follows the laws of current and voltage I'm familiar with, but everything else is its own language.

All suggestions will be appreciated.


  1. Something to check. Electrical contractors work for speed. They will use the direct insertion holes to wire up the duplex outlets instead of the screw terminals. Not that great a contact. Check each outlet, pull the wires out of the slip in contacts and hook them to the screw terminals.

    1. concur. This is a common problem, to the point that many specifications for hospital construction forbid use of the insertion holes. My bet is that rewiring to use the terminals will cure the problem.

    2. When I took the GFI off the east wall and moved it to the west, I found that they had been plugged into holes on the back (what I assume you're referring to). On the new outlet and on the GFI when I moved it to the other wall, I put the wires under the screw heads and tightened down onto them.

      You're saying to do that to all the outlets, not just the GFI itself?

  2. You don't put fridge or freezer on GFI because it might trip and you don't notice before you have a box of rotten food.
    GFI can also get weak over time?

    1. GFI can also get weak over time? They can, but most sources I could find said to expect a 10 year life. They're 4 years old.

      That's one of the reasons I swapped the east one onto the west wall. It seems really unlikely both of them could be bad or both would wear out early.

    2. The cheap stuff from China and south of the border doesn't last as long as the cheap stuff that used to be made here. "Quality control? We don't need no stinking quality control!"

  3. And it may be an intermittent fault in the wire itself. Especially if the contractor used carp wiring or recycled stuff.

    And, as Unknown said above, check all the connections. Especially anywhere a junction or splice is in. (Sometimes having exposed wiring or exterior conduit runs (not inside the wall) is a good thing.)

  4. Terminology is a bit mixed up - "Once or twice in the intervening four years, the west wall breaker would trip". Is it the 20A breaker or the GFCI as you continue further talking about GFCI tripping.

    Check the neutral screw on your load panel? A number of surge protectors have 3 mov devices with 2 to ground. A floating neutral does cause problems.

    You could borrow a Megger that can measure leakage - required instrument to test the electrical in mobile homes after moving. But would be surprised.

    How clean is your service power - look on a scope? You might be running hot on the supply due to bad power factor upstream.

    1. Terminology is a bit mixed up - "Once or twice in the intervening four years, the west wall breaker would trip". Is it the 20A breaker or the GFCI as you continue further talking about GFCI tripping.

      D'oh! I try not to do that, but say it all the time. It was the GFCI, not the breaker. The breaker has never tripped.

      I have NOT done anything inside either the breaker panel in the shop or the main one in the house. Until your question about checking the neutral screw in the panel, I wouldn't have known anything to look for.

      As far as we've gotten is that we've thought that perhaps the ground wire off that side of the shop isn't connected well to the rest of the grounds but haven't looked into anything.

      These are the sorts of recommendations I was hoping to get.

    2. I should add I have one of those three light outlet checkers that will indicate if an outlet is wired properly, had an open neutral, open ground, hot and neutral or ground swapped and so on.

      Every outlet in the shop look perfectly normal. That means if I have a problem like an open neutral, it's an intermittent, once-every-two-days-to-two-weeks thing.

    3. Talking with another engineering buddy - do you have any moisture problems with the new construction. A moist downstream receptacle (or hidden junction box in the wall) has been known to trip gfci's randomly - but it's probably a Vancouver thing.

      and the obvious - the downstream receptacles from the gfci are properly wired (line and load)

    4. That was (is) one of my suspicions. When the wires had to go through the concrete block exterior walls for anything, the electricians filled the area with some sort of foam in a can, and I was wondering if there could be enough openings for moisture to be oozing into the junction boxes.

      We don't have quite as much moisture as the Vancouver area, but we get constantly high humidity. My only two week vacation in the Seattle and Olympic NP area made me think the rain there was sorta like our days of "chunky humidity".

      Arguing against moisture oozing into the outlets is that there are NO other signs of humidity issues. No drywall staining, nothing on the baseboards.

      Right now, the only thing recommended here that I can definitely say might be an issue is that the electricians used the "wire stuck in a hole" mounts and not "wires wrapped around the screw". So far, the GFI hasn't tripped after removing one of the surge protected strips back a few days ago - but that's less than a week and I won't feel comfortable thinking it's fixed until I get to a month without it tripping.

  5. Surge protectors using GEMOV's tend to get very leaky after the device has been clobbered a few times. When the GEMOV fails hard, it usually destructs with a nice pop and flash. But in the meantime, they get leakier and leakier, and start to unbalance the line, and can occasionally go 'low impedance', and then clear themselves.

    It causes GFI's to get all upset.

    The reason I know this is because the electricians on the ships I worked on monitor the power very closely. Several of our equipment areas were causing problems, and the ship's electricians found several bad surge protector power strips in each room that were leaky enough to trigger their equipment.

    Try removing all surge protectors and see if it helps.

    1. What I've been doing is trying one power strip at a time, rather than shot gunning all of them at once. More of "troubleshoot to the faulty component" rather than just get the damned thing fixed. I should really do that.

      That's from my training and doesn't make sense here!

    2. HAH! I'm the same way about a lot of things.

      Troubleshooting a $20 surge strip to the component level is a fine example.

  6. Never use those plugin wire holes in the back of an outlet. If you crack one open, you will find it makes a connection with the EDGE of a piece of spring metal. The spring gets weak over time with movement and heat cycling, and you get a poor connection. Then they draw enough current to get warm with use. I hear that fires can result.

    Cockroaches love outlets and especially GFCI's for some reason. Their droppings seem to be conductive (at a guess), since they can kill a GFCI in a year or two.

    The makers instruct you to use the test button at least every month. Not a good idea, as doing this test seems to wear out the unit. Kids kill the unit doing this.

    Refrigerator motors are not compatible with them. Put them on a separate dedicated circuit. Microwave ovens SEEM to be compatible, as they don't appear to trip them, but they do seem to be over-representative in kitchen GFCI's that die early.

    Circuit breakers, and GFCI's, get weaker each time they trip, although they don't mention this in the documentation.

    In one old building I maintain, with a non-grounded system (2 wire outlets), the environment is such that I consider GFCI's to be a consumable component. Other buildings with what appear to be original ~30 yo GFCI's are just now beginning to fail.

    1. Cockroaches love outlets and especially GFCI's for some reason. Their droppings seem to be conductive (at a guess), since they can kill a GFCI in a year or two.
      Considering Florida is cockroach mecca, I've seen remarkably few droppings in the boxes, but I'll keep an eye out.

      We have a GFI breaker in our house's main panel that an old refrigerator has been on since the 90s and it has never popped. I've also never tested it. We used to have one to a lamp post in the front yard that popped all the time, until we pulled that lamp out and terminated the wires.

    2. I have heard that newer GFCI's and AFCI's can be RF sensitive. This can also vary from brand to brand. Check your instances of tripping against your transmitting on the various Ham bands.

    3. My station transmitting is one of the few things I can rule out.

  7. Sig, nuke it from orbit, the only way to be sure....

  8. IIRC the GFCI works when there is an imbalance between neutral and hot, and has nothing to do with the ground connection. Good grounding is essential for other reasons, so good thing you fixed that.

    Inductive loads can cause GFCI activation. An inspector once required us to put our dimmers on GFCI circuits for an outdoor event. That was a cluster....

    I'd look for anything that turns on and off- blower motors, fridge motors, led motion lights, sump pump motor, condensate drain pumps, etc.

    Something is causing the issue, and it's more likely a physical process/change than random heating or cooling wiggling a connection.

    Try putting a mechanical electric clock on the circuit overnight. You'll be able to see exactly when it activated... If you don't have one, try Goodwill.

    And finally, the power company has an interest in this and may install a power quality monitor for free. Tell them your lights occasionally get BRIGHTER, and they'll be over in a hurry :-)


    1. And finally, the power company has an interest in this and may install a power quality monitor for free. Tell them your lights occasionally get BRIGHTER, and they'll be over in a hurry :-)

      I was working on my Sherline CNC one time that the GFI tripped and the funny thing is that I noticed that the lights did change brightness and flicker a few times before it tripped.

  9. Timings of trip/no trip don't seem correlated enough with the changes in storms or ground rod or surge protectors for me to believe they are casually connected.

    Do you have any scope-type things whose triggering could be set up to watch both hot and neutral and capture the previous second before a trip?

    Insects or frogs in any of the electrical boxes?

    Whole-house MOVs are a 2" cube which only costs $50 at Lowe's.

    I prefer the "screw clamp back wire"/"screw-actuated clamp" outlets where a straight wire inserted goes under a washer under a screw.

    Tighten every screw clamp in the breaker box and you may find lots which are not as tight as they should be.

  10. Worked with a pool contractor where a motor would snap the ground fault interrupter whenever the power was applied. The factory tech. said nothing else could be on the circuit but the motor. I asked 'not even an incandescent lamp'. "Yep" he said but 'its not even on'. "Doesn't matter", he said it is still in the circuit with the ground and neutral. Who Knew?

    1. That's a complication - I always have things plugged in and off. Wait, was the factory tech from the motor company or the GFI company?

      If there can never be anything even plugged in on the GFI but turned off, that just ain't happening in my world. I'm running too many things for that to be possible.

    2. The tech of the manufacturer of the equipment that had the motor on it. He termed it that the lamp was introducing 'noise' into the circuit.

  11. I haven't done much mains troubleshooting myself, but maybe both of your GFI outlets are bad.

    How about replacing them with a GFI breaker?

  12. Just guessing here, but: is there any heavy industry hooked up to your local substation? Snapping big loads on-or off can unbalance things.

    Is there a good connection between the branches of the neutral line?

    Apparently, current is sneaking away somewhere. Could there be leakage from neutral to ground? Theoretically, they are supposed to be at the same potential, but if there is a load on one branch, neutral can be pulled around.

  13. Check all the outlets on the circuit that has a popping GFI. Especially outdoor facing or in moist areas. This is one of the primary causes of random triggering.

    Someone also mentioned heavy inductive loads going on and off, that is possible but should not be an issue unless there are other issues such as mechanically loose wiring or wet wiring.

    Good luck.

  14. You mentioned the GFI's don't last. Also surge suppression strips don't last. If they have been dumping surges, they are probably not lasting faster than normal. (Eventually they won't act like surge suppressors and the surges can start damaging equipment.)

    I would consider replacing the surge suppressors. Even though they cost more than they should.

  15. This comment has been removed by a blog administrator.

  16. Hello SiGb:

    Did you ever solve your problem? If you did, then I would love to read about it as a follow-up post.

    1. Hi Joe,

      I'm sort of in the philosophical quandary of "did I fix the intermittent problem or make it more intermittent?"

      The GFI hasn't tripped in 7 or 8 weeks. What I did was replace most of the surge suppressor strips. 3 out of 4 of them. Yes, one of the suspect strips is still in place and I'm plugging another one in just when I need it.

      So I'm not completely sure it's fixed, but it's much better. Did the improvement come from replacing the strips or from the generally better weather or something else? Don't know.

      Hard to get much of a conclusion other than it's easier to live with right now.