Leviton GFCI nuisance tripping and circuit analysis

[gar]

190120-2411 EST

tersh?

On the surface, quick analysis, your fault LED circuit of post #175 does not make operational sense.

.

 

[junkhound]

does not make operational sense

Agree, piecemeal with erroneous interpretation is borderline diy..

fwiw, here are 2 waveforms, > 1/2 amp on turn off on a common mode commercial CT, both wires thru CT in same direction. 500 us/div sweep
Lower pic is differential mode 300 mA 60 Hz rms, 120 Vac 60 Hz rms.

Top is 60 mA/div, common mode, both wires thru commercial CT at turn off, lower pic is differential mode showing about 800 mA inrush DM

 

[tersh]

190120-2411 EST

tersh?

On the surface, quick analysis, your fault LED circuit of post #175 does not make operational sense.

.

Why. In the following circuit. It's isolated to the red led only.

This is with flash that makes the red led on above not visible:

As you can see clearly. There is no more chip, no more SCR, no more diodes and resistors. They were all removed. The reset circuit was even gone. I tried to use the SCR from the reset circuit to the main but it was fried too (probably because the chip is defective already). The diodes from pin 5 was even cut. So its the isolated led circuit, and it turns on.

1. You can see the Collector of the NPN transistor connecting to the 1M ohm resistor and to ac supply red.
2. You can see the Emitter connecting directly to the ac supply black.
3. You can see the Base connecting directly to the 10k ohm resistor and to the ac supply black
4. You can see the Collector connecting to the 4.7k ohm resistor and then direct to the LED terminal
5. The other LED terminal connect to the ac supply black.

The red led lights up. What is wrong with this diagram which is based on above??

I took up 5 years course in electronics engineering. I spent 6 years in college. One year more for failed subjects.

It's frustrating I couldn't even understand all this now.. Lol

 

[tersh]

190120-2352 EST

tersh:

Troubleshooting is a process of performing experiments or tests, making logical decisions from information obtained, making new logical decisions, and repeating as necessary. In electrical circuits it means knowing the basic operation of the devices you are working with. You did a pretty good job of this, but cut a little too much instead doing other experiments first.

I think that you proved that the false signal was getting into the IC input by shorting 1,2, and 3. Possibly it was not good performance of the IC. The long lead to pin 3 could be the problem. Shorting the current transformer at the transformer instead of at the IC pins might be an indicator of whether the long lead is a problem. The current transformer might be a problem if it does not properly difference the forward and return currents at high frequencies.

A new test I have thought of is to supply AC power to the output terminals and the load. Thus, no current flows thru the current transformer, but we have voltage transients to the board circuitry and the current transformer. If no tripping, then likely it is an apparent current imbalance of the current transformer. Then one could try low pass filtering at the current transformer output, but probably best done at the IC input.

It appears you are giving up, but you really shouldn't because it is a good learning experience. There are things about electronics that you could learn and be of benefit to you.

.

Here is an interesting and illogical results of the experiments you suggested above.

1. When I shorted the current inbalance sense coil. There is no more tripping of the shaded pole motor after switching for more than 30 times.

2. When the ac supply was put at the load. Guess what. There is tripping! (one out of 4 which you can see in the youtube video shared earlier)

3. With the ac supply still put at the load and with the current inbalance sense coil shorted, there is NO tripping even after more than 30 switchings! Normally it trips less than 7 switchings.

So what in the blazes is in the sense coil where it's not current inbalance that can trip it when ac supply put even at the load (here ac power doesn't pass through the sense coil).

Transients or capacitive coupling occurring right at the sense coil?

I don't have any other instruments to try them on and don't have the skills for it.

Tomorrow my uncle will go to the United States. I will let him bring the following:

I'd like my uncle in Ohio to send it via USPS ground to gar for more expert analysis especially making separate low pass filter for the sense coil etc. which I don't have the equipments, knowledge or skills to do right. I'm asking international assistance for this problem. If gar doesn't want it. Then it can be shipped to Junkbound or to Winnie or any electrical engineers with instruments, knowledge and will share all results.

Thank you.

 

[tersh]

Here is an interesting and illogical results of the experiments you suggested above.

1. When I shorted the current inbalance sense coil. There is no more tripping of the shaded pole motor after switching for more than 30 times.

2. When the ac supply was put at the load. Guess what. There is tripping! (one out of 4 which you can see in the youtube video shared earlier)

3. With the ac supply still put at the load and with the current inbalance sense coil shorted, there is NO tripping even after more than 30 switchings! Normally it trips less than 7 switchings.

So what in the blazes is in the sense coil where it's not current inbalance that can trip it when ac supply put even at the load (here ac power doesn't pass through the sense coil).

Transients or capacitive coupling occurring right at the sense coil?

I don't have any other instruments to try them on and don't have the skills for it.

Tomorrow my uncle will go to the United States. I will let him bring the following:

I'd like my uncle in Ohio to send it via USPS ground to gar for more expert analysis especially making separate low pass filter for the sense coil etc. which I don't have the equipments, knowledge or skills to do right. I'm asking international assistance for this problem. If gar doesn't want it. Then it can be shipped to Junkbound or to Winnie or any electrical engineers with instruments, knowledge and will share all results.

Thank you.

The frequency of tripping when the ac supply was supplied to the load side with nothing in the input side is about the same. Tripping once in 2 to 5 switches jus like in the video. But it's difficult to do this continuously because when it trips, you can no longer reset it due to the circuit no longer have any power. So you have to switch the wires to the input, reset it, then change it to the load side again. When it trips, you have to do it again.

Whatever. Since if you put the ac suppy to the load, with nothing at input terminal. And it trips. It's definitely not the sense coil getting current inbalance because you have nothing at the input side. There is no load at input, it's open.

Therefore one very logical possibility now (after pondering on the unexpected result) is the when the inductive kick forms capacitive coupling in the long cooper path at middle leading to the sense coil. It needs the sense coil length to produce greater inductor loading to cross the threshold and produce enough voltage enough to either cause signal inbalance at pin 1,2,3 or directly affecting the SCR. What do you think?

Also I mentioned threshold because if you short the sense coil with ac supply at load side. It won't trip at all. Only when the sense coil was not shorted and ac supply at load would it trip with same frequency as when ac supply is at input side.

If you would say the inductive kick is directly affecting the sense coil, can this occur? Or should it need the long cooper path to inject it to the sense coil to raise the voltage, transient or whatever that can make pin 1,2,3 produce non-null result? or directly triggering the SCR? What is more likely of these two?

 

[tersh]

The frequency of tripping when the ac supply was supplied to the load side with nothing in the input side is about the same. Tripping once in 2 to 5 switches jus like in the video. But it's difficult to do this continuously because when it trips, you can no longer reset it due to the circuit no longer have any power. So you have to switch the wires to the input, reset it, then change it to the load side again. When it trips, you have to do it again.

Whatever. Since if you put the ac suppy to the load, with nothing at input terminal. And it trips. It's definitely not the sense coil getting current inbalance because you have nothing at the input side. There is no load at input, it's open.

Therefore one very logical possibility now (after pondering on the unexpected result) is the when the inductive kick forms capacitive coupling in the long cooper path at middle leading to the sense coil. It needs the sense coil length to produce greater inductor loading to cross the threshold and produce enough voltage enough to either cause signal inbalance at pin 1,2,3 or directly affecting the SCR. What do you think?

Also I mentioned threshold because if you short the sense coil with ac supply at load side. It won't trip at all. Only when the sense coil was not shorted and ac supply at load would it trip with same frequency as when ac supply is at input side.

If you would say the inductive kick is directly affecting the sense coil, can this occur? Or should it need the long cooper path to inject it to the sense coil to raise the voltage, transient or whatever that can make pin 1,2,3 produce non-null result? or directly triggering the SCR? What is more likely of these two?

Another thing. I noticed the design of the American Leviton has the chip so close to the sense coil (almost above each other). If this can still trip 1 out of 4 time. Then the long cooper path to the sense coil both contributing to increased capacitive coupling is not true.

In the china gfci, the chip is opposite side to the sense coil. So maybe the cause of nuisance tripping is different for different GFCI manufacturers? Maybe in the Leviton the cause is the C3 near the SCR as gar noticed in his model?

Is there a rule or implementing rules and regulations that focus on capacitive coupling within circuit? For example, the SQUID or Superconduting Quatum Interference Device takes into account capacitive coupling in the circuit?

If the china gfci has the chip close to the sense coil. Maybe tripping would not occur anymore or is there a unique explanation that a sense coil can directly contribute to transients? bad design coil? chip? what do you think?

 

[gar]

190121-1111 EST

tersh:

Since you have now run the experiment where no current flows thru the current transformer this may imply that there is capacitive coupling between the current transformer primary leads and the secondary. If this is what is happening, then it would be because the current transformer lacks an electrostatic shield between primary and secondary.

I could try some experiments on your circuit.

About your LED fault circuit. There just has to be some adequate current limiting to the transistor base emitter circuit. If not the transistor would be destroyed.

.

.

 

[tersh]

190121-1111 EST

tersh:

Since you have now run the experiment where no current flows thru the current transformer this may imply that there is capacitive coupling between the current transformer primary leads and the secondary. If this is what is happening, then it would be because the current transformer lacks an electrostatic shield between primary and secondary.

How are you supposed to put electrostatic shield between the primary conductor and secondary winding?

I could try some experiments on your circuit.

About your LED fault circuit. There just has to be some adequate current limiting to the transistor base emitter circuit. If not the transistor would be destroyed.

In the above. You can clearly see that transistor Emitter is directly connected to the ac supply black. And the Base is connected to the ac supply black via a 10k ohm resistor. Why. How should they be connected? But the red led lights up when the circuit is damaged. So it means it is working and the connection is really such:

When the pcb is working. Something is suppressing the red led from lighting up. Any idea how the suppression is done.. via signal to the base or from above to collector?

 

[tersh]

190121-1111 EST

tersh:

Since you have now run the experiment where no current flows thru the current transformer this may imply that there is capacitive coupling between the current transformer primary leads and the secondary. If this is what is happening, then it would be because the current transformer lacks an electrostatic shield between primary and secondary.

I could try some experiments on your circuit.

You could when you get them in a week. It's on its way.

Here's a recap how the top works.

The plug metallic terminals were connected to the pcb at the lower left and right side. A green led is connected to them in series with a 200k ohm resistor (in the pcb, the green and red labelling were reversed because the china labeler probably confused what the words mean). Now look at this. ​

When the contacts were closed, and you put ac supply to input, the load side is energized and the green LED were lit. When the contacts were opened, the top portion and bottom portion is totally disconnected except the switches and red fault led. However, when you put ac supply at load side to test if the sense coil would trip. And the contacts were opened. You would still get a green LED on.
I initially got confused when I first tried this where you asked me to put the ac supply at load. I heard a click and still see the green light. And I continued pressing the motor switch. Later I realized the click was the contacts opening from the trips. And even when the contacts were opened (from the trips). The green led won't turn off because you are applying power to the load side.

Now about the red LED. It's connected as the red LED in the lower part of the circuit shared earlier. In an isolated circuit, the red light stays on. But something is suppressing it so the red led isn't lit in normal operation.

I included the blue and red button for the Test and Reset terminals.

I also included the grounded neutral sense coil and current inbalance sense coil to help determine the composition of the toroids and know if there is electrostatic shielding or not and how it compares to US sense coil toroids and whether this is the sole cause why even when ac supply is supplied to the load side, the gfci trips with same frequency as when it's put in line side. This proved that it was not current inbalance causing the gfci to trip.

Remainder. Do not disconnect pin 1 because it will immediately destroy the SCR and chip. Before they were destroyed. I could see the Red LED turning on or blinking in synchronization to the solenoid clicking so fast and heating up (insulation smoking). It's so mysterious how it could turn on when circuits get damaged.

It seems the chip can self destruct if Rset between 1 and 2 was not put or removed. I thought it was only the SCR destroyed. This was why I cut the SCR in the 4th test circuit. The red light didn't turn on. I resoldered the SCR. Note the SCR was so sensitive. I used used small tweezer to move the SCR of other 1st to 3rd pcbs, and the cover comes up exposing the metal inside the SCR. It may not survive great heat that's why I took effort to reconnect the 3 legs in the 4th working test pcb. The solder may not look nice but there is continuous conduction of the 3 legs to rest of circuit and it's working perfectly.

That's all information I have now. I just don't have access to any oscilloscope for any detailed analysis of both the mysterious red light and exact source of capacitive coupling. And even if I go back to my electronic engineering laboratory, they won't let me touch it because I'm no longer a student, but alumni And my main interest now is the Higgs field and Large Hadron Collider.

Thank you very much for helping.

About your LED fault circuit. There just has to be some adequate current limiting to the transistor base emitter circuit. If not the transistor would be destroyed.

.

.

 

[tersh]

The last topic I'll inquire is the exact reason for the nuisance tripping of this GFCI outlet.

So far it has nuisance trip on:

1. phone chargers
2. orange juice squeezer machine
3. rice cooker
4. refrigerators
5. Dental spray
6. some others

The Siemens-2 pole GFCI breakers don't nuisance trip on any of them. Zero nuisance tripping, not even once. All GFCI outlet must be designed like the Siemens.

I think nuisance tripping can be considered a design flaw. It is not leakage current (at least for the particular model mentioned). This was because when you supply power to the load side with the input side open (as gar suggested to do). It still had nuisance tripping on the waterpik motor. Only when you short the sense coil input that it stops. Gar theorized it could be capacitive coupling between the primary and secondary of the sense coil. If that is simply the cause, then all manufacturers must use sense coil like Siemens which doesn't nuisance trip.

My motor and meiji outlet sample are already with gar. Hope he can examine it and confirm whether it is simply the sense coil capacitive coupling or others are involved. Also in case he didn't encounter any nuisance tripping with the supplied water motors. Then it could be US electricity is cleaner? But far out considering the Siemens don't nuisance trip while it does all the time. I don't have access to any oscilloscope and last I handled it was 25 years ago.

Also gar, please help this electronics puzzle.

consider this:

now the following resistors are shorted:

The complete voltage will drop across 1st 2 ohm resistor which will be 3V.Voltage across other resistors will be 0.

So in the circuit. Consider that when base was not triggered. And current doesn't flow from collector to emitter. The LED turns on (during fault condition).

When base was triggered. And current can flow from collector to emitter. Then just like the 3 resistor analogy, the current won't go to the LED and the LED turns off. Does this mean the base is constantly being turned on and this is what suppressed the red LED from turning ON? In my 2 test units. The red LED is always on because the chip is defective or the diode opened.

Consider the full circuit:

​

​

​

Consider also when the chip is defective and the red LED turns on, the reset isn't pressable. So it's like its detecting something in pin 5 of the FM2141 (or the Fairchild equivalent of it)?

Once I know the reasons for it. Then all my questions are answered then I'd no longer ask any questions. So many thanks for it.

 

[tortuga]

Interesting thread indeed, my limited understanding of Philippines 60Hz power is you have a 230V L-N or a Y connected 415Y240 volts ( or 400Y230 volt) or a 120/240 system. I have a copy of the Philippine NEC somewhere and the main amendments are 230V and 415Y240 are a standard voltage in 240.6. I believe this statement is mistaken:

<snip>I just realized something. Since the Philippines centertap is the one grounded (at least in the capital), the line to soil is only 120v, not 240v. </snip>

Sure possibly so if your in Manila you might have 120/240 single phase but by the NEC ( as amended in the Philippines ) you must assume you have just 230V single phase, with 230V to ground, and need devices rated as such.
I would assume any breaker used in the Philippines should be rated 415 L-L and 240V L-N
Similar to a 480Y277 system in the mainland US.
Here is an example I found of a GE plug in breaker manufactured for such a system :
https://i1.wp.com/myphilippinelife.com/wp-content/uploads/GE_breaker_label.jpg?w=1000&ssl=1
or here http://apps.geindustrial.com/publibrary/checkout/GEZ-7927?TNR=Application%20and%20Technical%7CGEZ-7927%7Cgeneric (the correct breaker likely is a TQL215GF)
It is interesting to see such a hybrid Euro / North American system, I would not be surprised if we see 415Y240 or 400Y230 volt gradually become a optional utility voltage here eventually.
Cheers

 

[tersh]

When gar received the waterpik motor, shipping has damaged the plastic switch so he couldn't turn it on. Hope he can make it run when he has time.

When I asked other electrical engineers in the US about all those reports of GFCI outlet nuisance tripping. They either believed there was actual leak or inductance motor can induced some current inbalances of some sort. None believed or commented it was capacitive coupling in the circuit itself.

If it can be proven that nuisance tripping is caused by bad design where capacitive coupling can induce nuisance tripping. Don't you consider it as defective product that needs to be either recalled or taken out of shelves?

Nuisance tripping can be a nuisance and dangerous. What if it tripped on your electric fan and you thought it was nuisance tripping when in fact there was an actual current leak already.

I mentioned defective product because it is possible to manufacture one that never nuisance trip. The legendary Siemens 2-pole GFCI breakers. It was so eerie for mere 5mA to power so many circuits without nuisance trip. I had to use GFCI tester with the ground leg connected to the soil just to test and make sure the Siemens were not defective. They were not and it tripped during the soil test.

GFCI outlets must be designed like the Siemens GFCI breakers, never tripping.

For the present tripping Leviton and related GFCI outlets. If it can be proven they were caused by capacitive coupling in the circuit. Do you consider it as just tolerance or do you think they must be recalled and those faulty GFCI outlets taken out from the store shelves?

 

[mopowr steve]

Yes!

 

[victor.cherkashi]

maybe it's time reconsider 5mA threshold. As far as I know the ground fault protection in EU residential is 30mA at 230V. it's huge difference. do they have more durable people ?

Sent from my ONEPLUS A6013 using Tapatalk

 

[ptonsparky]

maybe it's time reconsider 5mA threshold. As far as I know the ground fault protection in EU residential is 30mA at 230V. it's huge difference. do they have more expendable people ?

Sent from my ONEPLUS A6013 using Tapatalk

IFIFY.

 

[kwired]

maybe it's time reconsider 5mA threshold. As far as I know the ground fault protection in EU residential is 30mA at 230V. it's huge difference. do they have more durable people ?

Sent from my ONEPLUS A6013 using Tapatalk

More to it than just 5 and 30 mA. We are using it on 120 volts to ground they are using it on 240 volts to ground. We have mostly used it in the past on 5-15 and 5-20 receptacle outlets - those have a tendency to have missing EGC pins on the plugs. I don't know how common it is to see missing grounding pin on EU equipment, my guess is not that common though.

 

[tersh]

More to it than just 5 and 30 mA. We are using it on 120 volts to ground they are using it on 240 volts to ground. We have mostly used it in the past on 5-15 and 5-20 receptacle outlets - those have a tendency to have missing EGC pins on the plugs. I don't know how common it is to see missing grounding pin on EU equipment, my guess is not that common though.

In Asia (or at least in my building). I am using 5mA on 240v phase to phase (US centertapped). The Siemens GFCI breakers ae so reliable that all my lights are put on load side to it and no nuisance tripping of any kind and no worries of sudden lights out. The only time the Siemens GFCI breakers tripped was when there was actual live wire to concrete short in the wires leading to the washing machine and attic. That's the only time they tripped. While the Meiji Leviton chipped GFCI outlets, they tripped even on phone chargers and orange juice machine. It tripped not because of leaking current but something wrong with the GFCI circuit (which hopefully gar can share us the exact mechanism of the fault when he fixed the plastic switch of the test motor and can test it on the meiji sample unit I sent him. Please gar as I don't have any oscilloscope and can no longer test it any further). So nuisance tripping GFCI outlets must either be recalled or taken out of store shelves. We must aim for GFCI products that doesn't nuisance trip like the Siemens breakers.

 

[kwired]

In Asia (or at least in my building). I am using 5mA on 240v phase to phase (US centertapped). The Siemens GFCI breakers ae so reliable that all my lights are put on load side to it and no nuisance tripping of any kind and no worries of sudden lights out. The only time the Siemens GFCI breakers tripped was when there was actual live wire to concrete short in the wires leading to the washing machine and attic. That's the only time they tripped. While the Meiji Leviton chipped GFCI outlets, they tripped even on phone chargers and orange juice machine. It tripped not because of leaking current but something wrong with the GFCI circuit (which hopefully gar can share us the exact mechanism of the fault when he fixed the plastic switch of the test motor and can test it on the meiji sample unit I sent him. Please gar as I don't have any oscilloscope and can no longer test it any further). So nuisance tripping GFCI outlets must either be recalled or taken out of store shelves. We must aim for GFCI products that doesn't nuisance trip like the Siemens breakers.

Siemens breaker VS Levition receptacle nuisance tripping issues has nothing to do with 5mA vs 30 mA trip point and everything to do with unexpected noise, kickback, etc. that they didn't design to filter against in some way.

I occasionally run into GFCI receptacles that trip from something happening on line side of the device. This has everything to do with interference from such things and nothing to do with trip setting.

 

[tersh]

Siemens breaker VS Levition receptacle nuisance tripping issues has nothing to do with 5mA vs 30 mA trip point and everything to do with unexpected noise, kickback, etc. that they didn't design to filter against in some way.

I occasionally run into GFCI receptacles that trip from something happening on line side of the device. This has everything to do with interference from such things and nothing to do with trip setting.

gar theorized it was lack of dielectric shielding in the sense coil itself that allows capacitive coupling between primary and secondary.. but how does it translate to current inbalance in the input side of the chip?

Also does that mean the Siemens sense coil has dielectric shielding? Can they actually do this to tiny transformers? In big toroidal, only my medical grade isolation toroidal transformer has dielectric shielding. But it's just lying in the box because it was supposed to be used on the fridge that nuisance trip on the Leviton clone outlets but never on the Siemens (not even once).

I'm still concerned about GFCI outlets because I am still using 8 pieces of back in wet areas as backup to the whole house GFCI breakers that protect all circuits. But it's so hassle for the GFCI outlets to trip every now and then. People who are so used to seeing them trip and just tried to press the Reset button may mask real leakage or hide it and this is dangerous. That's why all GFCI products must be designed like the Siemens GFCI breakers, completely and totally resistant to all kinds of nuisance tripping.

 

[kwired]

gar theorized it was lack of dielectric shielding in the sense coil itself that allows capacitive coupling between primary and secondary.. but how does it translate to current inbalance in the input side of the chip?

Also does that mean the Siemens sense coil has dielectric shielding? Can they actually do this to tiny transformers? In big toroidal, only my medical grade isolation toroidal transformer has dielectric shielding. But it's just lying in the box because it was supposed to be used on the fridge that nuisance trip on the Leviton clone outlets but never on the Siemens (not even once).

I'm still concerned about GFCI outlets because I am still using 8 pieces of back in wet areas as backup to the whole house GFCI breakers that protect all circuits. But it's so hassle for the GFCI outlets to trip every now and then. People who are so used to seeing them trip and just tried to press the Reset button may mask real leakage or hide it and this is dangerous. That's why all GFCI products must be designed like the Siemens GFCI breakers, completely and totally resistant to all kinds of nuisance tripping.

I wouldn't claim the Siemens breaker is bullet proof when it comes to nuisance tripping, you just witnessed specific interference situations that trip a Leviton receptacle but don't trip the Siemens. There very well may be something out there that can cause undesired trip of the Siemens.

I personally haven't seen any such nuisance trips for any GFCI circuit breakers but have seen a lot of GFCI receptacles subject to nuisance trips, some more frequently than others. Don't know enough about how they design to mitigate such things to be able to comment much further though. One thing with receptacles I have noticed though is you are generally better off to get P&S, Levition, or maybe Hubbell brands than to get cheap discount brands at a big box store - they must put some extra care into lowering nuisance trips into their designs.