Open Service Neutral GFCI/RCD Detector

[winnie]

As Adamjama says, a clamp meter uses the basic mechanism of detecting the magnetic flux around a wire to measure current.

I would not trust a random clamp meter for general AC use designed to detect single mA of leakage current; when I use these I am happy if they are accurate to the nearest 0.1A.

A google search for leakage current sensor brings up many options; I have never used these and could not give you any advice as to which to pick. I looked at one and it's specs suggest that it has about 0.5mA of error measuring 5mA of leakage on the 40mA scale. https://www.calright.com/product/reed-st-9809-ac-leakage-current-tester/

Regarding your question about the output of the sense toroids in the GFCI, they have lots of turns in the sense coil (1000 in the schematic you posted), so the output current of the coil is a small fraction of the sensed current. This very small current is passed through a resistor (the 1.1 meg resistor in the schematic) to give a voltage which is easy to sense. In the schematic you gave, 4mA of leakage current would give a voltage of about 5V across Rsns.

-Jon

 

[Tony S]

Loosing the supply neutral along with an earth fault? South Africa has an optional supply accessory that monitors the neutral and any earth leakage fault. Not many are fitted as illegal abstraction of power is common place.

 

[tersh]

As Adamjama says, a clamp meter uses the basic mechanism of detecting the magnetic flux around a wire to measure current.

I would not trust a random clamp meter for general AC use designed to detect single mA of leakage current; when I use these I am happy if they are accurate to the nearest 0.1A.

A google search for leakage current sensor brings up many options; I have never used these and could not give you any advice as to which to pick. I looked at one and it's specs suggest that it has about 0.5mA of error measuring 5mA of leakage on the 40mA scale. https://www.calright.com/product/reed-st-9809-ac-leakage-current-tester/

Regarding your question about the output of the sense toroids in the GFCI, they have lots of turns in the sense coil (1000 in the schematic you posted), so the output current of the coil is a small fraction of the sensed current. This very small current is passed through a resistor (the 1.1 meg resistor in the schematic) to give a voltage which is easy to sense. In the schematic you gave, 4mA of leakage current would give a voltage of about 5V across Rsns.

-Jon

Good you mentioned it because I was just looking at this $707 Fluke 360 series leakage clamp meter.

https://www.amazon.com/Fluke-FLUKE-...sr=8-1-fkmr0&keywords=fluke+360+leakage+clamp

It describes:

"The Fluke 360 Series of true-rms leakage current clamp meters help users detect, document, record and compare leakage current readings over time as a means of preventing unplanned downtime, and identifying intermittent GFCI and RCD trips, all without taking equipment off line. The Fluke 368 FC and 369 FC have large diameter jaws (40 mm for the 368 FC, 61 mm for the 369 FC) for work with large conductors. The clamp’s jaw is fully shielded to accurately capture very small leakage signals, and to minimize external electromagnetic interference. "

If it's not fully shield, it can't really pick up in the 1mA resolution? Why?

This is after I watched this video about chassis leakage tests with a professional instrument and the Fluke at:

https://www.youtube.com/watch?v=TZzy2cWihGY

This is my personal clamp meter.

You think the Fluke or Reed leakage meter can accurately detect any leakage current of refrigerators??

To review something about Ampere's law with Maxwell's addition:

Does the cancellation of the magnetic flux occurs in the toroid by each wire contributing to it. So does the fact the two wires are so close together already made their fields cancel and nothing for the toroid and sense coil to pick up? Note the subtle distinction between this two. Thanks.


​

 

[winnie]

To review something about Ampere's law with Maxwell's addition:

Does the cancellation of the magnetic flux occurs in the toroid by each wire contributing to it. So does the fact the two wires are so close together already made their fields cancel and nothing for the toroid and sense coil to pick up? Note the subtle distinction between this two. Thanks.

It is almost certainly a combination of both effects.

1) What does it mean 'when two wires are close together'? Close is a very subjective measurement The fact is that the fields would only strictly cancel if the two wires took up the exact same space, which is physically impossible. So when someone says 'two wires close together' what they are really doing is making an approximation: when the spacing between the wires is small (say 1/10) the distance from the wires to where you are making your measurement, then the measured flux is approximately zero. But it is never _exactly_ zero since the wires don't actually overlap.

2) Any residual field from each wire will create some flux in the toroid. If the two wires are carrying balanced current, then this should cancel out.

(Note, there is a way that you can get perfect canceling of the fields without overlap: by using 'coaxial cable'. Here the symmetry means that any external fields cancel, and the only field left is between the conductors, or resulting from residual current. There is no good reason to use coax for 60Hz power circuits, though you might use it in a very very accurate residual current measurement instrument.)

-Jon

 

[winnie]

You think the Fluke or Reed leakage meter can accurately detect any leakage current of refrigerators??

That is what they are sold to do, but as I said I have no experience with these tools. If you buy such a tool, I'd suggest running some controlled tests (eg. by creating a well defined leakage current through a resistance), and then making a bunch of reference measurements. How do the readings change with different cables? Different orientation of the probes and the cables? Etc.

Buy from a reputable supplier who will accept a return if the instrument doesn't meet its specifications.

-Jon

 

[tersh]

I recently tested one of the RCD's in my flat (apartment) board that fed the ring circuit, it tripped at 17mA. Supposedly too low but I'm more than happy with the result.

We can get 5mA RCD's but they are dammed expensive. I used them for my photographic darkroom, electricity, water, chemicals and wet hands aren't a good combination, I've an aversion to electric shocks after getting 10kV up my arm. The shock didn't do the damage, hitting the concrete wall backwards knocked me out.

How did you measure it to be 17mA? What brand of clamp meter did you use (if you use such)? I found the $707 Fluke so expensive and the $100 Reed not tested. Is yours accurate?

How expensive are these 5mA RCDs? What are the brands you have inquired?

 

[Adamjamma]

Good you mentioned it because I was just looking at this $707 Fluke 360 series leakage clamp meter.

https://www.amazon.com/Fluke-FLUKE-...sr=8-1-fkmr0&keywords=fluke+360+leakage+clamp

It describes:

"The Fluke 360 Series of true-rms leakage current clamp meters help users detect, document, record and compare leakage current readings over time as a means of preventing unplanned downtime, and identifying intermittent GFCI and RCD trips, all without taking equipment off line. The Fluke 368 FC and 369 FC have large diameter jaws (40 mm for the 368 FC, 61 mm for the 369 FC) for work with large conductors. The clamp’s jaw is fully shielded to accurately capture very small leakage signals, and to minimize external electromagnetic interference. "

If it's not fully shield, it can't really pick up in the 1mA resolution? Why?

This is after I watched this video about chassis leakage tests with a professional instrument and the Fluke at:

https://www.youtube.com/watch?v=TZzy2cWihGY

This is my personal clamp meter.

You think the Fluke or Reed leakage meter can accurately detect any leakage current of refrigerators??

To review something about Ampere's law with Maxwell's addition:

Does the cancellation of the magnetic flux occurs in the toroid by each wire contributing to it. So does the fact the two wires are so close together already made their fields cancel and nothing for the toroid and sense coil to pick up? Note the subtle distinction between this two. Thanks.


​

You would not run both cables through the meter and expect to get a good idea of leakage or amps... you would try to separate the wires to run two meters, one for forward and one for reverse current, comparing the difference... however many times they only try to since the neutral current to see what the reverse current leakage is... even I know that... kinda like checking for swr ... simple analogue meter has one needle.. better meter has two needles and you compare the two.. even better metertells youin digits your swr is x...lol

 

[tersh]

You would not run both cables through the meter and expect to get a good idea of leakage or amps... you would try to separate the wires to run two meters, one for forward and one for reverse current, comparing the difference... however many times they only try to since the neutral current to see what the reverse current leakage is... even I know that... kinda like checking for swr ... simple analogue meter has one needle.. better meter has two needles and you compare the two.. even better metertells youin digits your swr is x...lol

Adamjamma, what is the best RCD brand you have encountered, one with the best features?

 

[winnie]

You would not run both cables through the meter and expect to get a good idea of leakage or amps... you would try to separate the wires to run two meters, one for forward and one for reverse current, comparing the difference

I would recommend against this. You end up with the 'difference of two large numbers' problem. A small % error in your measurements (totally reasonable since no meter is perfect) leads to a huge % error in your leakage measurement.

-Jon

 

[tersh]

As Adamjama says, a clamp meter uses the basic mechanism of detecting the magnetic flux around a wire to measure current.

I would not trust a random clamp meter for general AC use designed to detect single mA of leakage current; when I use these I am happy if they are accurate to the nearest 0.1A.

A google search for leakage current sensor brings up many options; I have never used these and could not give you any advice as to which to pick. I looked at one and it's specs suggest that it has about 0.5mA of error measuring 5mA of leakage on the 40mA scale. https://www.calright.com/product/reed-st-9809-ac-leakage-current-tester/

Regarding your question about the output of the sense toroids in the GFCI, they have lots of turns in the sense coil (1000 in the schematic you posted), so the output current of the coil is a small fraction of the sensed current. This very small current is passed through a resistor (the 1.1 meg resistor in the schematic) to give a voltage which is easy to sense. In the schematic you gave, 4mA of leakage current would give a voltage of about 5V across Rsns.

-Jon

How about these RCDs without any electronics.. so it's 1:1 instead of 1:1000? So the output of these RCD sense coil are the real leakage current such as 30mA? But are these enough to drive a relay that doesn't use any electronics too? Would you personally prefer one with or without electronics?

 

[tersh]

How about these RCDs without any electronics.. so it's 1:1 instead of 1:1000? So the output of these RCD sense coil are the real leakage current such as 30mA? But are these enough to drive a relay that doesn't use any electronics too? Would you personally prefer one with or without electronics?

I mean these RCDs above Winnie. Thanks.

 

[winnie]

I don't know enough of the details of those RCDs to comment.

I could imagine many variations on the theme, with lots of turns in the sense coil to get higher voltage, few turns to get higher current, etc.

It looks as though at least some RCDs function by having the sensed residual current drive an electromagnetic trip mechanism.

My gut feeling is that the electronic RCD will be more sensitive and more predictable, but also more prone to damage from things such as transient overvoltage on the supply line.

-Jon

 

[tersh]

I don't know enough of the details of those RCDs to comment.

I could imagine many variations on the theme, with lots of turns in the sense coil to get higher voltage, few turns to get higher current, etc.

It looks as though at least some RCDs function by having the sensed residual current drive an electromagnetic trip mechanism.

My gut feeling is that the electronic RCD will be more sensitive and more predictable, but also more prone to damage from things such as transient overvoltage on the supply line.

-Jon

In principle. It is also possible to create 5mA GFCI without electronics even if it would increase the size that can't fit receptacle form factor anymore or let's say make it plug-in (to existing receptacles). So if you were to do a thesis on plug-in GFCI (with 5mA tripping) but without using electronics, do you think you can do it?

 

[tersh]

This may help:

https://www.youtube.com/watch?v=Yg6G5VUSsWA


What they are saying is that when the neutral breaks, everything in the structure that is tied to the panel ground bar will be energized relative to the soil outside the structure (remote earth).


That voltage to the soil will vary based on two thing:

1. How well balanced your 120 volt loads are between the two hot wires, more imbalance more voltage.

2. The resistance of the ground rods. Higher resistance earth rod means higher voltage, lower resistance means lower voltage because neutral current can get back to the source easier.





100% Correct!

But here is the twist, if you bond the rebar in the concrete to the panel ground bar, you energize both the rebar and the toaster case. Because both are energized at the same equal potential, the voltage between them is actually zero volts protecting the user. This is how bonding works and why its done on pools, operating rooms, ect. But both the rebar and toaster case will still be 120 volts to the earth outside.

Let's get on topic about this open service neutral issue.

Someone told me last week that if the local electrode ground rod is property put with 2 rods 6 feet apart, then even if the neutral was cut at service, your appliance won't have the current going to neutral, then to neutral/ground bond at main panel and go to the metal chassis (because it cant go out the service entrance). This is because the local electrode ground rod won't make it happen? Is it true? Or not??


By the way, the reason I didn't inquire anymore about open service neutral was because I found out that the Siemens First Surge SPD only used 240v to operate, it doesn't need the 120v line to neutral voltage. I found this out after I told the electrician to remove the ground/neutral lead of the Siemens First Surge from the neutral bar in the office building. And it still works (Both green lights still on and no audible alarm to indicate fault condition, the 2nd manual below says light would blink if something is wrong) see https://www.downloads.siemens.com/d...aspx?pos=download&fct=getasset&id1=BTLV_43430 and https://www.downloads.siemens.com/d...spx?pos=download&fct=getasset&id1=BTLV_43434:

I know by removing the neutral/ground wire from the neutral bar. The L-G, N-G protection modes is gone. So there is only the L-L protection mode (imagine two 150v MCOV SPD in series). I will only tell the electrician to return the ground/neutral lead to the neutral bar after I got further confirmation from Siemens, etc. that the neutral won't suddenly get used (or switch to 120v, can this happen?). Btw.. a 65 veteran EE in the US taught me how to use the split phase Siemen First Surge in the panel. Had to spend one month discussing with him before buying the $240 Siemens First Surge.


So you see, the last 120v device in the entire building was deactivated just to avoid the problem of open neutral where the metal cases could become energize (Note: our office building has neutral connection to the utility transformer centertap while the parent home doesn't have any neutral (ground wire) connection to the utility centertap).


In Europe, I found out there are RCDs (or Pluggable RCDs) that used power to close the contacts during normal usage, meaning if you remove the source of power, it becomes open and your appliance would lose power. This can solve the problem of open service neutral (at least there). But I wonder if there is 120v version of these available for use in the US? You could use this before the appliance so in the event neutral at service was lost. Your appliance would lose the power and can't make the dreaded live to neutral to main panel neutral/ground bond to metal chassis situation. Also in the US, inspectors may not allow these normally open RCDs to put before the main panel?

There are easy solutions to the problem of open service neutral, but why the US doesn't allow its citizens or make available 120v RCDs that is normally open to handle it?

In any case, do not rely on bonding- someone could lean against the building while standing on soil and get hit. Key is to remove the fault no matter that scenario.


[/FONT][/COLOR][/LEFT]




Honestly, I doubt the lower VPR makes a difference. What you are protecting against is line to line surges (over 230 volts anyways) and line to ground voltages that are MUCH higher then 120 or 230 volts. Insulation in most appliances can tolerate over 1kv before break down. Whether the MOVs are rated 120 or 230 volt line to ground will not make a difference as both will start shunning and clamping voltage well before it can damage line to ground insulation.

 

[winnie]

Plug in GFCIs which detect (and operate with) open neutral conditions are available, and I believe required by OSHA.

However the simple 'open the circuit with loss of power' approach would not reliably detect an open service neutral in the US. Because of the numerous 120V loads on services in the US, devices connected to the neutral bus would likely remain powered if the service neutral opened.

The voltage would become unstable, and connected loads might be damaged. So a relay which detects the unbalanced voltage and opens would do the trick. These are available and not prohibited, but I know of no-one actually using them in a residential service.

-Jon

 

[tersh]

Plug in GFCIs which detect (and operate with) open neutral conditions are available, and I believe required by OSHA.

However the simple 'open the circuit with loss of power' approach would not reliably detect an open service neutral in the US. Because of the numerous 120V loads on services in the US, devices connected to the neutral bus would likely remain powered if the service neutral opened.

The voltage would become unstable, and connected loads might be damaged. So a relay which detects the unbalanced voltage and opens would do the trick. These are available and not prohibited, but I know of no-one actually using them in a residential service.

-Jon

That is why either one main RCD that can disconnect power to the entire 120v must be put before the main panel, and disconnects when neutral is loss. Or every single device with access to neutral must have its own RCD that can disconnect when neutral is loss. In the US, this may not be practical.. but if you have only 2 or 3 120v gadgets, this is feasible. Also in the US you can easily tell if neutral is loss because light would go out or flicker.. but if the 2 or 3 gadgets don't have visual indicators.. then the more RCDs that can disconnect with loss of power required. Is this pluggable GFCI that can disconnect with loss of power available in the US? What brands carry them?

 

[winnie]

You miss the point. In the US you have numerous 120V loads, connected between both 'hot' lines and 'neutral'. These loads act as a voltage divider, and will create a neutral point even if the service neutral is lost. This neutral point has unstable voltage, so that the voltage between 'hot' and 'neutral' will be anything from 0 to 240V depending on the connected loads. But you can't reliably detect the loss of neutral as a loss of 120V power.

 

[tersh]

You miss the point. In the US you have numerous 120V loads, connected between both 'hot' lines and 'neutral'. These loads act as a voltage divider, and will create a neutral point even if the service neutral is lost. This neutral point has unstable voltage, so that the voltage between 'hot' and 'neutral' will be anything from 0 to 240V depending on the connected loads. But you can't reliably detect the loss of neutral as a loss of 120V power.

I was familiar with it after watching several youtube videos before about some loads getting much more voltage from divider action and loss of it. But if you connect each load to an RCD or GFCI that disables load contacts during loss of power.. even a light bulb.. it can disable all power to all appliances during lost of service neutral.. isn't it? But of course it's not practical to install 30 pcs of RCD/GFCI in a home. I guess that's why they don't manufacturer GFCI outlets that disconnects load power when neutral was loss.

I can now perfectly understand why our country doesn't use neutral as 120v.. because of this very complex issues about installing it the right away which can't be taught to millions of homes immediately. In 240v line to line. Any electrician just connects the load to it and that's it. We don't have Mike Holt schools in the country.

Why doesn't the US just use 240v too line to line and omit using the neutral as 120v? What's with 120v? It requires thicker wires from more current. Our 240v appliance has thinner wires. Rest of world use 220-240v. Maybe USA uses 120v to avail of lower 150v MCOV Surge Protective Devices to take advantage of lower let thru voltage (related to the VPR) for better surge protection?

 

[ptonsparky]

I was familiar with it after watching several youtube videos before about some loads getting much more voltage from divider action and loss of it. But if you connect each load to an RCD or GFCI that disables load contacts during loss of power.. even a light bulb.. it can disable all power to all appliances during lost of service neutral.. isn't it? But of course it's not practical to install 30 pcs of RCD/GFCI in a home. I guess that's why they don't manufacturer GFCI outlets that disconnects load power when neutral was loss.

I can now perfectly understand why our country doesn't use neutral as 120v.. because of this very complex issues about installing it the right away which can't be taught to millions of homes immediately. In 240v line to line. Any electrician just connects the load to it and that's it. We don't have Mike Holt schools in the country.

Why doesn't the US just use 240v too line to line and omit using the neutral as 120v? What's with 120v? It requires thicker wires from more current. Our 240v appliance has thinner wires. Rest of world use 220-240v. Maybe USA uses 120v to avail of lower 150v MCOV Surge Protective Devices to take advantage of lower let thru voltage (related to the VPR) for better surge protection?

.

More likely the surge protection was designed for the voltage we use.

 

[kwired]

If you use some device that drops out anytime power is lost and have a distribution system that experiences frequent short drops in voltage - you may find that rather inconvenient as it will leave the entire house without power every time there is such a brief loss of voltage.