Simple GFI tester question

[Smart $]

... Is there some other possibility that I haven't thought of?

Would the receptacles happen to be feed by an old two wire system? The receptacles do not have to have a ground attached to them to work, but the tester does.

Didn't read the whole thread and not going to, but...

If using a receptacle/gfci combo tester, shouldn't it indicate an open ground???

 

[kwired]

I did use the trip button and it worked. The one thing I didn't do and should have was pay attention to how the GFI recept was wired. I didn't pay careful attention because the 3-prong tester showed me "correct" for the wiring throughout the circuit. Given the limitations of those testers though, I was working on a dangerous assumption. It won't happen again.

Considering that it could have been a GFI that required more than 4-6 mA to trip, I find it frustrating at times that we really don't have many options for comprehensive GFI testing like the installation testers they use in other countries which give a good indication of how much leakage a GFI/RCD needs before it trips (some use a sweep test, others use timed, preset mA leaks). The only tester I've seen here that does it is the Greenlee 5708I which has 3, 5, 6, 7 and 10 mA test settings, but costs upwards of $300 and doesn't really do anything other than test GFIs.

Did the (alleged) downstream outlets lose power when the GFCI was tripped?

I'm imagining the circuit diagram in my head, but I'm thinking that you can wire a GFCI outlet on a 2-wire circuit with a bootleg ground, and the GFCI will test as having propery polarity and ground using a 3-light tester. I'm pretty sure you can trip it with an external GFCI test button on a "fancy" 3-light tester, and it's own built-in test button will trip properly. I don't think this wiring error will cause a current imbalance and trip the GFCI prematurely, but I'll have to draw this out and trace the circuit before I'm confident enough to postulate a hypothesis and set up a test.

I'm also thinking this GFCI outlet could even be mis-wired with a RPBG condition (Reverse Polarity Bootleg Ground) and it would still test as OK with a 3-light tester as well as the external and internal GFCI test buttons. And there would be no way to check for this "hot ground" condition without an external earth reference (DMM or solenoid tester on a copper pipe or test wire back to the panel bonding point or NCVT such as a VoltAlert). However, the GFCI in the branch circuit should protect consumers from an RPBG chassis shock resulting from a hot ground since that would create an unbalanced load condition and still trip at 5 mA fault current or thereabouts. There may be one mis-wiring condition that would appear to be wired correctly at first glance and check out properly with internal and external testers, but would actually disable the GFCI function. However, I'll have to draw this diagram out completely and set up a test to be sure.

While I've got to believe that no licensed electrician would do a bootleg ground on a GFCI knowingly, there's a real possibility that a DIY homeowner or plant maintenance guy could find what he thinks is a non-grounded outlet that's actually protected by an upstream GFCI receptacle or panel GFCI, and decide he's going to "ground" it with a bootleg jumper.

I'm not bringing this up as an electrician's or inspector's code issue, just as troubleshooting intel since there's a LOT of strange wiring problems in older homes and churches where I've wondered just what were they thinking? With enough knowledge of how forced wiring errors test using your meters, you can get a jump on troubleshooting real world errors. Although, I've got to admit I've seen a lot of wiring errors that I would have a hard time imagining.
As I say in my sound installers classes "You can't make this stuff up"! :slaphead:

If I understood correctly his tester did indicate proper polarity and would not trip the GFCI that the outlets are apparently on. Bootleg ground ahead of the GFCI would respond as if everyting is fine - wiring is new enough that is not too likely but not impossible I guess.

What needs confirmation is whether or not the outlets in question are indeed on load side of GFCI, and that it is a class A GFCI. If not a class A GFCI then the tester was not useless you found a problem of not using proper GFCI and the use of test button alone would not have revealed that.

 

[kwired]

Didn't read the whole thread and not going to, but...

If using a receptacle/gfci combo tester, shouldn't it indicate an open ground???

It will not indicate an open ground if the ground was bootlegged from the neutral pressing test button should still put imbalance across the sensing device and trip if everything else is ok, and this has been brought up in the thread.

 

[Smart $]

It will not indicate an open ground if the ground was bootlegged from the neutral pressing test button should still put imbalance across the sensing device and trip if everything else is ok, and this has been brought up in the thread.

Only if the "bootleg" was "pirated" from line-side of gfci device. If bootlegged from load side of the GFCI, the imbalance created by the tester would NOT (normally) initiate a trip.

Sorry if already discussed, but it's easier to type a short reply than read the entire thread

 

[kwired]

Only if the "bootleg" was "pirated" from line-side of gfci device. If bootlegged from load side of the GFCI, the imbalance created by the tester would NOT (normally) initiate a trip.

Sorry if already discussed, but it's easier to type a short reply than read the entire thread

Yes it was alreasy mentioned - and that it needs to be done on line side of GFCI. Load side will likely result in *nuisance* tripping.

*to me it is doing what it is supposed to and not really nuisance tripping.*


Anything else you want me to read for you

 

[Smart $]

...
Anything else you want me to read for you

That should suffice... but thank you for offering

 

[PetrosA]

Did the (alleged) downstream outlets lose power when the GFCI was tripped?

If I understood correctly his tester did indicate proper polarity and would not trip the GFCI that the outlets are apparently on. Bootleg ground ahead of the GFCI would respond as if everyting is fine - wiring is new enough that is not too likely but not impossible I guess.

What needs confirmation is whether or not the outlets in question are indeed on load side of GFCI, and that it is a class A GFCI. If not a class A GFCI then the tester was not useless you found a problem of not using proper GFCI and the use of test button alone would not have revealed that.

I didn't check each and every bath, but at least one bathroom plus all the outlets (four) in the master bath went dead once the GFI tripped. I'm really kicking myself for not noting whether the line and load was reversed and for not keeping that GFI for further testing. I'm considering getting either an Ideal circuit analyzer or Amprobe INSP-3 which would at least give me the option to check at class A and B levels. Any recommendations anyone?

 

[jmsokol]

I'm considering getting either an Ideal circuit analyzer or Amprobe INSP-3 which would at least give me the option to check at class A and B levels. Any recommendations anyone?

Peter, In addition to the INSP-3 and Ideal SureTest, I also have the Extech CT-70 and have compared all three testers both on my test bench and in the field. I still have my "antique" Woodhead Ground Loop Impedance Tester (GLIT for short) with a big analog meter, which still works great but offers much less intel than a modern load tester. Guess that makes me an antique as well...

I would leave the Extech CT-70 out of the running since it has too many screens to go through for info, and the information is too spread out for a quick evaluation. Also, there's no primary screen indication of a Bootleg Ground, which is important when testing a lot of rental property outlets.

That leaves the Amprobe INSP-3 and the Ideal SureTest. Both units seem to have similar accuracy since we've tried them out on the same outlets and they read close to the same numbers. However, I did not confirm their accuracy with an independent measurement, which is what I would do using a calibrated meter and known resistance load. But I've gotta believe their accuracy is close enough for field work. (I used to build nuclear missile guidance systems in a previous life, so calibration procedures were drilled into me...)

Both the SureTest and INSP-3 provide similar info, but the SureTest is a little quicker to use since it gives you a pass/fail indicator on the first screen that's pretty clear. The INSP-3 gives you more data on the first page, but you have to really work to interpret what it all means. For instance, while either tester finds a bootleg ground, the INSP-3 just blinks the actual impedance reading number, while the SureTest actually has 3 lights on screen that show up like a cube tester, and the ground "light" blinks a big F for false ground if it determines there's too low of a impedance between the Neutral and Ground (less than 15 feet of wire). So a wiring fault that could be overlooked on the INSP-3's smaller display is really obvious on the SureTest display. Also, the fluorescent display of the SureTest is easier to see than the INPS-3, especially at a distance, so that could be helpful to somebody with less than perfect vision (like me).

Both the Amprobe INSP-3 and Ideal SureTest have a nice carrying case with a short "euro" power plug, and both seem equally rugged. I would make the SureTest my first choice, then the INSP-3 as a second choice, and finally the CT-70. All of them will get you there, but the SureTest is the easiest one to drive.

Finally, remember that none of these GLIT's can tell the difference between a CPBG (Correct Polarity Bootleg Ground) and an RPBG (Reverse Polarity Bootleg Ground) outlet. To know the difference you'll either need to use a DMM or Wiggy from the outlet ground contact to a known earth connection such as a copper water pipe, a long extension cord ground that's plugged into a known-good outlet, a test wire run back to the service panel bonding point, or easiest of all a NCVT such as a Fluke VoltAlert or Amprobe VoltProbe. I really think using an NCVT as a quick check to discover any "hot" grounds followed up with an intelligent load tester such as an Ideal SureTest would be a very solid outlet check and quick to perform.

 

[PetrosA]

Mike, thanks for the in-depth answer. Personally, I don't think I'd mind the screen on the Amprobe and I like the fact that it's made in the US (according to Grainger), but... I talked with the local Ideal rep today who talked me out of the 61165 arc fault version as they're still trying to work out the bugs it has with various brands of AFCI breakers. He is checking to see if the 61164 GFI tester also has the ability to detect the load on the circuit under test like the 61-165 (this function isn't available on the Amprobe). This would put the two in a very similar price range for me and my local supply house carries the Ideal. I'll see how things progress and go from there. Thanks!

 

[hurk27]

This post was at Mike's post #40 before I relized there was another page, ( have got to pay attention to the pages LOL)

While yes you can have a neutral to ground connection if it is on the line side of the GFCI device, because a GFCI does not care how it is wired ahead of it, but this could never happen on the load side as all GFCI's have a second current transformer in them that inject a 120hz signal that will allow the electronics to detect a grounded neutral, and because of this a GFCI will immediately trip as soon as any power is applied to the circuit, this came about back in 1996 as UL revamped their requirement for safer GFCI's that will also not reset if they go bad and they have to remove all power from any receptacle on the load side, including the neutral, in the above diagram that kwired posted in post #32 he is not showing the neutral contact or the second 120hz current coil and it also does not show that even the receptacle contacts will be disconnected as he has the contacts after the GFCI device for simplicity, but they are there, I have a good diagram of how the newer GFCI's are wired let me see if I can find it.

But you are correct even a GFCI can be fooled into thinking that it is wired correctly if it is miss wired ahead of the device, and like you have so clearly been pointing out is or can be a very big problem, the good thing is even if the GFCI is wire RPBG it will still detect current flowing on any other path if it is not returning through both the hot or neutral in the device so it will still trip if someone were to get between the ground and earth, they might receive a shock but the GFCI will limit it to the set .005 amps and should protect them, but someone who has a bad heart or other health condition it might not be enough.

Check out this attachment it is a great paper on how GFCI's function and even shows the how they are wired internally

 

[jmsokol]

but this could never happen on the load side as all GFCI's have a second current transformer in them that inject a 120hz signal that will allow the electronics to detect a grounded neutral

Hurk....

Wow! Take a breath and let me review your post line by line and read the attachment.

But I am puzzled by your reference to 120 Hz signal as I'm not aware of that coming from a GFCI. What am I missing, or is that contained in the attachment?

 

[hurk27]

@ Mike again on the post 40, after reading a little more, I agree as I said above yes the GFCI can be RPBG wired ahead of it and still function as if it was wired correctly, and as you pointed out any current flowing in any other path that do not take it back through the current coil in the GFCI will cause it to trip if it exceeds the .005 amps, but in that post you made the commit that you thought there was a way that would defeat the GFCI from detecting a ground fault, and I don't think that is passable as any current that flows from the hot through the current coil must flow back through through same current coil or the coil will develop a voltage and trip the electronics, and the fact that all GFCI devices newer then 1996 will have the 120hz injection coil that will keep anyone from boot legging the ground from the neutral any where down stream from a GFCI device whether its a breaker or receptacle device, this does not not hold true to AFCI's as they do not have the 120hz injecting coil and that is why if you have a grounded neutral on a AFCI circuit it will not trip until a load is present.

I will say you are on top of the rest of it, and have offered some very great advice. again look over that PDF I attached to the previous post as it is a good reference for those who would like to know how a GFCI works, the writer is a very well known electronics publisher who has written many thousands of books on electronics and is most well know for his "Sam Facts" booklets that many of us used to repair anything from radios and TV's and even had a series on CB radios, and are highly sought after as they included all schematics for many of these electronic appliances.

 

[hurk27]

Hurk....

Wow! Take a breath and let me review your post line by line and read the attachment.

But I am puzzled by your reference to 120 Hz signal as I'm not aware of that coming from a GFCI. What am I missing, or is that contained in the attachment?

sorry I get carried away some times LOL

Yes it is in the atachment, and the next post was kind of before I seen this post

 

[hurk27]

also another little feature that newer GFCI's are required to have is they can not be reset if even there is a reversed line side neutral and hot, but this depends upon having a correctly connected EGC conductor to the GFCI device and would not apply on a two wire circuit, you can reverse the load side neutral and hot and it will still reset, but another feature is in you mix up or cross lets say put the load neutral on the line side neutral terminal and the line neutral on the load side terminal it will also not reset, but I have found a few GFCI receptacles that were made in China that have got past customs as they are clearly boot legged UL stickers on them that do not have these required features and I sent them to UL to investigate them.

 

[ELA]

RE: 120 Hz current injection

RE: 120 Hz current injection

You might find this link interesting. About midway down it talks about a how else this is commonly done.

http://www.idea2ic.com/GFI/GFICs.html

I came across this after having reverse engineered a Leviton 7599 GFCI and finding that the 200:1 coil was not connected as shown in the LM1851 doc that is commonly viewed.
I measured the voltage across that coil and it was not 120Hz so that is what got me looking for how else it worked.

 

[jmsokol]

You might find this link interesting. About midway down it talks about a how else this is commonly done.

http://www.idea2ic.com/GFI/GFICs.html

This is most excellent! OpAmp theory I understand, so this schematic and explanation will plug a lot of holes in my GFCI theory. Thanks very much...

I'm happiest when I'm learning something...