Monday, December 30, 2019

Here Be Gremlins

After the old nautical charts that would have the legend "Here Be Dragons" over unexplored waters.

Since we put in our shop in '14, I've been plagued by random triggering of my ground fault interrupter outlets.  Long time readers with extra sharp memory will know that I've talked about this before, and did one long post about it.  The bottom line about this is that it has never gone away.  Our electric utility replaced the last transformer in the distribution we're on - the one for us and our three closest neighbors - in the last year and the problem became more intermittent.  When they replaced it we were having our lights flicker or pulsate randomly, and when the flickering was bad, the GFI would pop.  We'd get the outlet popping once a day or a couple of times in a day, and then not again for a month or six weeks.  We've thought that it correlated best to rain and then seen driving rain not matter.  This fall/winter it has seemed to correlate to strong winds.  We have underground utilities, so there are no wires moving in the breeze in the last mile to us.  We've thought it was caused by the surge protected AC strips I was using, but those have all been replaced with only one being surge protected.

Last week, I finally decided to replace the GFI outlet itself.  It was originally the less likely of our two GFI outlets to trip, but I pulled that outlet put a non-protected outlet in its place, and moved it from our east wall to our west wall.  Briefly, the west wall has all of my machine tools plugged into it except one - my air compressor - and the place where AC-powered machinery and water meet is the cooling system on my milling machine.  The east wall has far less plugged in.

Saturday, I'm making some sanding passes on the cylinder in the big lathe and cranked the RPMs up from its minimum (100 RPM) up toward 200.  The GFI pops!  Coincidence?  Random event?  So I reset the GFI and try again.  Slower this time and by 170 RPM it pops again.  Now I've run that lathe at higher RPM many times before and it never popped while using the outlet I replaced because I thought it might be bad.  I tried the experiment a few more times (Goldfinger's law) and it was completely repeatable.

I open the junction box and pull the outlet out.  It's a 120V 15 A GFI just like the circuit, I tell Mrs. Graybeard.  See that breaker in the box right there?  It's clearly marked ... 20A.  Oops.  Bought the wrong size GFI.  OK, put back the old one, try it again, up to 250 RPM with no hassles.   Never pops the GFI.  Then I look up the spec on the lathe: that motor is never supposed to need more than 12A so unless there was a short transient over 15A, it should have been OK.

Regardless, I tell myself, I should replace the one I'm suspicious of with another 120V 20A GFI and resume the experiment of seeing if it pops randomly (not related to the lathe).  We waited for today because it was the weekend and picked up the third GFI outlet for that spot (True Value hardware is starting to love me).  Put in the new outlet today, now it's time to test it with the lathe.

Sure as Schiff it pops when I get above 150 RPM!

I realize this gets confusing - it gets confusing for me.  I find myself with two different, brand new GFIs, different brands, a 15 A and 20 A that do the same thing - the lathe makes them pop open when I go above about 150 RPM.  I have one, old GFI that's also a 20 A that the lathe never pops.  It's not really easy to make sense of that. It seems to me that the lathe speed controller is probably doing something wrong during the transients while speeding up. The "good" two see that while the possibly "bad" one doesn't.  I'm not even sure the lathe is doing something wrong, although it looks that way.

I'm in that awkward position of not knowing quite what to do.  No schematics, no manual besides the lathe's operator's manual.  There are relatively few of these compared to the 7x12 mini lathes


  1. Does this lathe have a DC motor? Some DC motors and controls have a tendency to couple a lot of noise onto the ground wire of the unit, and it might be disturbing your GFI.

    Quick test....Take a grounded outlet adapter (the little grey "plugs" with the green pigtail) and use it as a "ground bypass" adapter by not connecting the green wire to anything. It defeats the safety ground, by disconnecting the machine ground from the AC outlet, and lets the machine "float". I'm sure you've done it with test equipment to make measurements.

    If letting the machine ground float makes the problem go away, you've pretty much confirmed it was ground noise upsetting the GFI.

    1. The manual just says "Brushless motor", so I'm assuming brushless DC. Which implies the kind of cheap PWM controller bhudson mentions below.

      This putting in a ground isolating adapter is a simple experiment and I can measure the ground current while doing it. Even hang the oscilloscope on it.

    2. How things have changed! "Brushless DC" motors and controllers used to be far too expensive for a "Consumer" item like your lathe, so all you got was a brushed DC motor and simple SCR speed control.

  2. Is your GFI intended for resistance type loads? I think it is looking for a 6 ma

    between the black and white leads. The GFI might be seeing the inductive load

    from the motor as a fault. Also, the GFI should not be relied on for over load

    protection. GFIs can be confusing. Wait till you have trouble with AFIC breakers.

  3. 1. Replace all your GFIs on the circuits that are blowing as once the GFI starts tripping, it makes it more likely to trip the next time. Sounds like you've done this, but make sure you've changed all outlets on the circuit. (I had a bad outlet on a circuit that was all over the place. 1 outlet killed a lot of stuff. Got to the point of calling an electrician, he came out, took out his multi-tool-electric analyzer thing and tracked it down in 10 minutes. That rotten outlet was causing problems all the way to my main panel, which he replaced both the affected circuit breaker and the main breaker)

    2. Sounds strange, but call the utility to check on the drop to your house. They're responsible for the line from the distribution line to your weatherhead, including weatherhead. You may have some damage in that line from the lightning strike. If your utilities are underground, still have them check from the distribution loop to your house. Lightning can and does destroy underground lines in new and unusual ways.

    3. Swap out the breakers on the circuit that are affected. Like the GFIs, they tend to pop more readily, and throw garbage down the lines, as they get old or damaged.

    And finally...

    Check the main breaker for any damage. Then check the breaker box for damage. Lightning and line spikes can do some awful funny things to your circuit breakers.

    If you are still stymied, and don't have all the high-faluting test equipment yet, then the next step is to... call in an electrician, who has all the high-faluting test equipment.

    Sucks, but that lightning damage probably fragged a lot of stuff in your house that won't show up for years.

  4. It is difficult to troubleshoot anything remotely. My background as industrial electrician may provide some assistance. Bear in mind that anything I say is based on an old plant, over 100 years, updated as and when required. We did have GFCI circuits on newer installations. But never had them on any circuits that carried motors.

    Your speaking of controlling speed indicates a D-C motor. It could be a single phase variable frequency controller but that's more than a hobby level machine justifys. So, assuming D-C, the most likely is a PWM drive. There are several ways of controlling DC motors, but again, being a hobby level machine, PWM is the most likely. Chinese PWM controllers have little to no filtering, at least at the low end. Higher cost controllers will likely have some, but (again) hobby level machines would not.

    When I rewired my home, near 40 years back, I used my experience in industrial as a guide. FWIW, we have no air conditioning and the heat is an old school gas furnace. As in a floor furnace with no blower. The big issue is that during power outages we still have heat. It does get cold in Alabama, especially as I age.

    The end result is that the only motor loads here are the refrigerator and fans in the summer. Now in the shop is a different matter but that is a seperate building that feeds just below the main entrance. The only GFCIs are in the bathroom feeds and the kitchen.

    My industrial experience is that GFCIs are NEVER run in motor circuits. The same in my house. On the rare occasion that (current) wife tries to run a fan in the kitchen, the GFCI trips. If a scope were to be run on the power line, there are "imbalanced" currents for motor loads. These are written off as "transients" although I don't know the technical reasoning.

    My suggestion is to remove the GFCI protection for the shop circuits. If you are troubled with getting zapped by a machine, look to the individual machine for a fault. Use rubber mats if you have them on a concrete floor.

    GFCIs look for a difference in line current vs neutral current. Motors can cause an instantaneous difference between line and neutral. It only lasts a part cycle but is enough to trip most GFCIs. Even standard circuit breakers have a time delay to compensate for motor loads at startup.

    I don't cover GFCIs in my meanderings;
    but looking at this may yield some insight to an electronics engineer. Keeping in mind that electricial and electronics are two distinctly different disciplines, there are repeated attempts by Mother Government to make them the same. 'Taint so, sez I. I don't like GFCIs, only using them when specifically called for around moisture by the (NEC) code authorities.

    Always remember, codes are a minimum acceptible practice. They do not define all approaches to construction. Note that my house was built in 1887, long before electricity was in common use. Yet it is still standing and has weathered storms (tornados included) that have collapsed "code approved" buildings. When I bought the place in 1975, it had 120 volt 30 amp electrical service and knob&tube wiring. Everything here has been my own work with a 240 volt 220 amp service. I'm still alive at 70(ish), though only God knows how. Wife has been here for near 30 years. The only problems we have seen was in 1991 or so, lightning struck nearby and she lost the electronics on a fancy new stove. The stove worked fine for 20+ years, just the oven timer went south with the lightning. I DON'T LIKE GFCIs, period.

    Bill Hudson, Artificer


  5. GFCIs and motors don't play well together, especially if they run intermittently, like refrigerators.

    Normally, 'fridges are run on a dedicated line, or a bypass loop with a single outlet. The single is to keep someone from plugging into that non-protected circuit in the kitchen. (it's normally hidden behind the 'fridge, and you can get away with using a duplex outlet there, but best not to take a chance. Most houses don't have enough outlets today, so if you leave one available, someone will find it and fill it.)

    If you run those motors on a GFCI, eventually it will trip while you are away, and present you with an unpleasant surprise on your return.

    BTW, cockroaches love outlets, especially the GFCI type. They will dramatically reduce it's life. I suspect their waste may be conductive, as the electronics will fail in as little as a year in a heavy infestation.

  6. These three comments - Beans at 2348 on 12/30, bhudson at 0132 and Will at 0542 this morning are simply great. I can't thank you enough!!

    Let me describe something possibly useful to know. This is an addition to our house we has built back in '14 while planning for retirement. We have two 20 A breakers in the wall breaker box with two rows of outlets that go down the east and west walls. The west wall has all of my shop equipment on it: Grizzly G0704, Sherline mill, two Sherline lathes, and a drill press. Everything that's always plugged in, except the computer and three CNC control boxes has motors. I've used a bunch of handheld power tools: pad sanders, jigsaws, belt sanders, wall powered handheld drills and more on that west wall. No piece of equipment has ever caused the GFI to pop until now. The Sherlines are all DC motors. I think the Grizzly mill is a DC motor as well. After resetting the GFI yesterday, I tested the mill's variable speed motor and the GFI didn't pop.

    When we first started getting into the problem of the GFI randomly opening, I learned about freezers/refrigerators and GFIs and moved that outlet (which rarely popped) from the east wall to the west. That was the first time I swapped out a GFI outlet in there. I replaced the GFI on the east wall with a plain outlet.

    I added above, in a reply to Drjim, that I just found one reference to the lathe's motor being "brushless", so I assume it's exactly the kind of cheap PWM controller, you mention, Bill.

    I have a Fogbuster, water-based coolant on the mill but never on the lathe. Because of the combination of water and electricity, I'd like to keep the GFI for that. The simple expedient is that the lathe is plugged into the west wall via extension cord. I can move that cord to put the lathe on the east side with just about zero effort. The lathe's manual says just to use a grounded outlet, nothing about using a GFI.

    It doesn't seem like this could have anything to do with the random popping of the GFIs because that happens when nothing is running. I'd say the majority of the time it pops nothing is running. I assume that since everything plugged into that wall is 120V single phase, they're only switching the hot line, which could mean that motors (as just big inductors) are on the neutral side of the power line at all times.

    I think what I'm going to do today is find a ground isolating adapter around here and plug the lathe into that while on the GFI, then see if it pops. I can put my oscilloscope onto ground and see what it looks like. Maybe there's a way I could clip the scope onto the power line itself.

  7. from:

    According to a study conducted by the American Society of Home Inspectors (published in IAEI News, November/December 1999), 21% of GFCI circuit breakers and 19% of GFCI receptacles tested did not provide GFCI protection. Yet, the circuit remained energized! In the examined cases, failures of the GFCI sensing circuits were mostly due to damage to the internal transient voltage surge protection (metal-oxide varistors) that protect the GFCI sensing circuit. This damage resulted from voltage surges from lightning and other transients. In areas of high-lightning activity, such as Southwest Florida, the failure rate for GFCI circuit breakers was more than 57%.

    Also: most GFCI receptacles are rated as 15/20 amp. This means they have 15 amp receptacles on them (two parallel vertical slots Hot & Neutral + a round ground slot) and a 20 amp pass through capacity for any 20 amp receptacles farther down the line (1 vertical slot, 1 T slot + ground).

    GFCI's sense the current in the hot (black) conductor and compare it to the current in the neutral (white) conductor normally via a differential transformer. any imbalance between them (Normally > 5 milliamps) will cause them to open, however some older units may have a higher threshold as they met an older standard which if my aging gray cells are still working I think was on the order of (> 50 milliamps)

    Note that GFCI's don't open on over current only an imbalance in hot and neutral current, The circuit breaker or fuse in the power panel is what handles over current conditions.

    GFCI's work best on resistive loads not inductive or capacitive loads. switching speed controllers frequently synthtize a sine wave (1 to 100 hz) by PWM rectified AC at a high frequency (5 - 40 KHz) across the motor windings. The inductance of the motor filters the high frequency components so the motor responds to the low frequency components for its speed. However there is always capacitance from the motor winding to its case and the higher the switching frequency the more energy (remember Xc decreases with frequency) to the GFCI this looks like a ground fault since some of the Hot current is being returned via ground vice the Neutral wire. This is the noise you will see on the lines if you put a O'scope on them.

    This is a greatly simplified explanation but hope it helps.

    1. Thanks for the refresher on this. I have seen most of it in various places. I understand how they're supposed to work, but am far from expert in ways they might not work properly.

      I started down this road trying to troubleshoot a very different problem: random GFI interruptions. One thing I'd never done is just replace the outlet with a new one - I had replaced it with one from the other side of the shop, so probably out of the same box from the same manufacturing lot sold wholesale to the electricians who installed them. I've since put in two new units and believe the 5 year old one is bad.

      Then there's the not-so small matter of the lightning strike we had last August. It hit a tree in the front yard and from there induced massive surges in wiring in the house. OTOH, the failure mode in that article you mention is the opposite of what we're seeing: those were failed in the "ON" position and never reset. Ours faults apparently at random.

  8. It would be very time consuming, but one 'test' would be to unplug everything on the suspect GFI, and then sit and wait. If it never random trips, you have a downstream device (one of the unplugged) causing it.

    I see your reasoning for having a GFI due to the coolant system for machining. I've never had this problem, so I'm in the dark here.....

    1. Also, there simply isn't enough room in here to list everything we've done. We've been puzzling over the random GFCI popping almost as long as the shop has been there. Maybe not all 4-1/2 years, but most of it. I can't tell you the number of things we've changed and the problem goes away for a month or two, then a random power open hits. When the power company replaced that transformer, the instances of the lights varying dropped way down, and the number of random GFCI openings went down, but not to zero.

  9. I thought I'd add a comment from the Little Machine Shop page for the lathe.

    "We have no expectation that any HiTorque machine will work on a circuit with a Ground-Fault Circuit Interrupter (GFCI). The National Electric Code has a specific exception to the requirement that receptacles in a garage or work area have GFCI: "Exception No. 2: A single receptacle or a duplex receptacle for two appliances that, in normal use, is not easily moved from one place to another and that is cord-and-plug connected"."

    Hmmm A 250 pound lathe on top of 35 pounds of steel cabinet sounds like "not easily moved" to me!

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

  11. SiG, some gfci outlets trip when power is removed, say a momentary 4000v circuit breaker on the pole opens and resets, may be the culprit. An example is the gfci included cords used on jobsites. I know for a fact that those need to be reset when first plugged in. Your outlets may suffer the same issue?

    Tom 762