A popular path for shock is: power wire, defective appliance, person, dirt/concrete(*), and back through ground bond to power company.
(*) Also metallic pipes which often run into the dirt.
There used to be actual ground current detectors. They did not work well. There are too many possible paths back to the service dirt-bond.
The "GFI" actually detects the -difference- of the current in the two load-power wires. Mine do not even look at the green wire; it's just there to drain the appliance cabinet.
Same as my bottle water company. The slip says to drop 3 full bottles, with a space where the driver notes how many empty bottles he picked up. If 3 and 3 all is well. If 3 and 2, they have a "leak". (3 and 4 is also a "leak", and they may wonder why.)
If my hedge trimmer pulls 5.000 Amps on the black wire, and returns 4.990 Amps on the white wire, there is a 0.010A leak of current. It could go many places. Maybe I have a 12K resistor around the GFI. Maybe the cord is cracked and 0.010A is zinging one blade of grass. But maybe the 0.010A is going through ME to dirt and back. That would be bad. The GFI assumes worst-case and cuts-out.
Note that the leak path may NOT involve dirt; but there are very few real cases which don't involve a "ground wire" at least tangentially.
And these things break the load-carrying wires, not the Ground/PE wire.
Therefore the UK's "Residual Current Breaker" name is probably more accurate. Here in the US we cling to Yankee Doodle and "GFI". (Which is what National Semi called it.)
It appears that *all* these devices are based on the same original chip design, now made many places in several variants.
GFI
http://www.idea2ic.com/GFI/LM1851.pdf
http://www.3c-test.com/de/?m=Type&a=download&id=189
My bottle water company will let me slide for one bottle short; if I keep holding-out I get the nasty letter to give them bottles or bucks. Quicker if 3-shy than 1-shy. Likewise the GFI will allow 0.1A leakage only for a very short time (<0.5 sec), and 0.005mA for a longer time (~7 secs) will trip it. I believe this time-curve is why US GFIs are nominally 5mA and UK RCBs are nominally 30mA: they picked different points on the same curve.
Yes, we call them RCDs here (Residual Current Device) - but I have never felt that « Residual » was a good way of expressing the action of the trip. It seems to me that the French language gets much closer with « Interrupteur différentiel »
Interrupteur differentiel LEGRAND, 30 mA 40 A AC | Leroy Merlin
not all GFI/RCB circuits are identical. I always thought that these were symmetrical devices and that is does not matter what is input or output. I was WRONG.. I wired accidentally some chinese (TOMZN brand) RCB/automatic fuse combo' (called RCBO) thinking that input was bottom output was top, neat for the wiring.
upon hitting the test button I heard the relay turn off, but then smoke and a dying smell came out. I had to find the reason for this. turns out the sensor circuit is a small toroid sense transformer coupled to a low power amplifier that triggers a thyristor. the thyristor is thought to trigger the release coil to turn itself off. if the RCB is wired in reverse, the thryristor and coil stay powered, and the thryristor dies a smelly death.
upon hitting the test button I heard the relay turn off, but then smoke and a dying smell came out. I had to find the reason for this. turns out the sensor circuit is a small toroid sense transformer coupled to a low power amplifier that triggers a thyristor. the thyristor is thought to trigger the release coil to turn itself off. if the RCB is wired in reverse, the thryristor and coil stay powered, and the thryristor dies a smelly death.
Indeed WRONG. Makes a big difference! (I did not know it was smelly.)
US market brand-name receptacle GFIs come with tape over the "Load" terminals, so you won't wire to them unless you need to (to feed downstream outlets with GFI protection).
US fusebox GFI/breakers in all the common US fuseboxes only go in ONE way. The box has a bus(es). The breakers snap onto the bus. That's how they get power. (With few odd exceptions) you can't run the power "backward".
I won't argue the bus against DIN-rail. There's much in favor of DIN. I am unsettled about the need to wire the "bus" in the field, because I have not worked in such boxes. I know there are pre-made buses for jumpering all the breakers in a group, but I gather they are also wired with bits of cable.
Being able to wire a GFI/RCB "backward" sounds like a drawback to me, but maybe you only do it once.
I was able to "Do-It-Myself" a RCBO/GFCI tester by using a small E14-type 7W/230V light bulb, a regular insulated bulb socket, 2 electrical wires coming from the socket and a 3-pole power plug. I connected one wire coming from the bulb to the GND from the mains plug and the remaining wire to one of the 2 remaining poles (one pole from the socket remains not connected).
When plugging the plug into the wall socket the RCBO/GFCI will trip if everything's working fine, otherwise the light will get powered ON and will remain lighten.
I found 3 defective RCBO breakers by using this "tester" and I replaced them with good ones. In case it matters, the defective RCBO's were CHiNT 20A 30mA.
P.S.: A 7-8 KOhms resistor should also be able to trip a RCBO or GFCI. Actually, the test button from the RCBO itself has a resistor inside to perform this test.
When plugging the plug into the wall socket the RCBO/GFCI will trip if everything's working fine, otherwise the light will get powered ON and will remain lighten.
I found 3 defective RCBO breakers by using this "tester" and I replaced them with good ones. In case it matters, the defective RCBO's were CHiNT 20A 30mA.
P.S.: A 7-8 KOhms resistor should also be able to trip a RCBO or GFCI. Actually, the test button from the RCBO itself has a resistor inside to perform this test.
Last edited:
the increased number of switching powr supplies and their EMI capacitors to ground or acrosss the transformer from primary to secondary, adds up to the total earth leakage current.
if there are some solar panel inverters operating, the grounded panels can create a big ground current. especially when it rains, as these panels are galvanically coupled to the mains. I bought myself a clamp current meter in the mA range, just clamp it over the cable and the difference of the current going in and out is indicated, very handy to find smaller ground currents. (UT251 from uni-T).
if there are some solar panel inverters operating, the grounded panels can create a big ground current. especially when it rains, as these panels are galvanically coupled to the mains. I bought myself a clamp current meter in the mA range, just clamp it over the cable and the difference of the current going in and out is indicated, very handy to find smaller ground currents. (UT251 from uni-T).
An UNI-T clamp meter I have myself too, but I only used it for phase/null measurements and only few times to do some in-house checks.
I prefer to connect the scope between Null and Ground to check the current-flaw and the existing noise. This way I've seen that vacuum cleaners and other electrical appliances needs to be OFF while listening to audio equipment, unless really good EMI/RFI filters and HUM-X-like ground breakers exist in the audio chain.
I prefer to connect the scope between Null and Ground to check the current-flaw and the existing noise. This way I've seen that vacuum cleaners and other electrical appliances needs to be OFF while listening to audio equipment, unless really good EMI/RFI filters and HUM-X-like ground breakers exist in the audio chain.
Odd that this thread popped up. Twice in two months a GFI has failed. Both, on autopsy, the major damage is the trip-solenoid which releases the contacts. It could be that the cut-out is not really cutting out.
FWIW: under US documentation, the *only* consumer-level "test" allowed is the TEST button on the device. No other test is considered valid/safe. (But I agree with you: a many-mA resistor should trip it.)
On a different thought: the US's lower voltage, higher current, and general casual approach to electricity means that we often have over-heated sockets. One documented problem is Glowing {electric} Contacts: socket contact is made only on a micro-point and that goes red-hot. It was thought that it is also a "noisy" contact and the ragged voltage drop could be inferred from ragged current. This gave us the "AFI" Arc-Fault Interrupter. However these things trip-out on vacuum-cleaners and other universal motors. This gets into the Politics of UL and the device makers, so I guess I'll stop there. But yes, you don't want several types of common electrical appliances where your audio can see them.
FWIW: under US documentation, the *only* consumer-level "test" allowed is the TEST button on the device. No other test is considered valid/safe. (But I agree with you: a many-mA resistor should trip it.)
On a different thought: the US's lower voltage, higher current, and general casual approach to electricity means that we often have over-heated sockets. One documented problem is Glowing {electric} Contacts: socket contact is made only on a micro-point and that goes red-hot. It was thought that it is also a "noisy" contact and the ragged voltage drop could be inferred from ragged current. This gave us the "AFI" Arc-Fault Interrupter. However these things trip-out on vacuum-cleaners and other universal motors. This gets into the Politics of UL and the device makers, so I guess I'll stop there. But yes, you don't want several types of common electrical appliances where your audio can see them.
Some close-up pics with the intestines of the defective GFI, please?
I've had very good results with Gewiss and Schneider RCBO's, in 5 years none failed (apartment and house).
Twice more current could damage the sockets indeed, however...at least you're much safer there at 120V instead of 230V here.
I've had very good results with Gewiss and Schneider RCBO's, in 5 years none failed (apartment and house).
Twice more current could damage the sockets indeed, however...at least you're much safer there at 120V instead of 230V here.
> Some close-up pics with the intestines of the defective GFI, please?
Attached.
My eye was caught by the swelling on the fine-wire coil with "A9" on the rivet. While some damage was done when breaking-in, some parts escaped, I believe this is the coil which latches the contact arm.
FWIW: the other (top) end has the differential current transformer. Two doughnuts, one red wire, one green core(*). The 20A current carrying paths pass up through the cores and out to the contact arms. The clearance across the 120V is quite slim, but reasonable. I note that the solder-blobs on the PCB where 20A is carried from side-screws to CT have been slathered with a Silicone caulk. And noteworthy: THIS particular GFI (and not past ones) has extra words about wet locations, implying I should not be using it in an outdoor box. (But it was sold pre-installed in an outdoor box!)
(*) I do not know if EUR has the second toroid. In recent years (decade+?), US GFIs will trip if N touches G. They detect this by putting a small RF signal on the lines. If connected with 20 feet of standard cable, they say "OK". If shorted together at the device (to fake a "ground") they trip. I did not know this, and had repeated 'trouble' with a GFI 4 feet from the fuse panel (shorter than typical wires).
Attached.
My eye was caught by the swelling on the fine-wire coil with "A9" on the rivet. While some damage was done when breaking-in, some parts escaped, I believe this is the coil which latches the contact arm.
FWIW: the other (top) end has the differential current transformer. Two doughnuts, one red wire, one green core(*). The 20A current carrying paths pass up through the cores and out to the contact arms. The clearance across the 120V is quite slim, but reasonable. I note that the solder-blobs on the PCB where 20A is carried from side-screws to CT have been slathered with a Silicone caulk. And noteworthy: THIS particular GFI (and not past ones) has extra words about wet locations, implying I should not be using it in an outdoor box. (But it was sold pre-installed in an outdoor box!)
(*) I do not know if EUR has the second toroid. In recent years (decade+?), US GFIs will trip if N touches G. They detect this by putting a small RF signal on the lines. If connected with 20 feet of standard cable, they say "OK". If shorted together at the device (to fake a "ground") they trip. I did not know this, and had repeated 'trouble' with a GFI 4 feet from the fuse panel (shorter than typical wires).
Attachments
Interesting, didn't knew that. Here in EU I've recently tested few RCBO's, but I only shorted GND with NULL in the wall-outlet, not inside the outbox. Makes lot of sense, because some people don't want to pay for a real grounding with less than 3-Ohms at their homes and prefer to connect NULL with GND inside their home outbox (AFAIK it's forbidden, but some still doing it). For GFI's not sure how this will help, but it's probably a good thing.
Thank you for the close-up pic!
Last edited:
More thoughts on AFCI's (Arc Fault Current Interrupter). They have a micro-processor that looks for the signatures of bad connections and insulation failures. The thought was that older homes often have these types of wiring problems. But older homes also have older appliances which have similar signatures. But this meant frequent call-backs for electricians, unhappy home owners. Then we add in the cost of these breakers to any old home electrical projects. All this meant a reluctance to do any electrical upgrades to older homes. So some communities removed the requirement of AFCI's on retro electrical projects.
* * * * * * * *
Note that a US community is not required to follow NEC code, they can add or remove sections if the so wish.
If the N wire fails (I have seen it happen), and anything "on" is in the outlet, the G wire, the box, and the appliance goes full "live" (230V??).
Several traditional US wiring shortcuts have been banned for this reason.
The code looks mostly at repair processes. We used to take a 120V+120V line to feed dishwasher and disposal, common N. But if anybody disconnects N under the sink to repair one device, the other device would be full 120V.
Code tries to pretend N does not come off by itself, I I have seen it more than once. Screws are not always full tight. (Aluminum is a curse in 1970s houses.) Some large loads can burn-off a downsized N wire or connector.
Checking for N-G shorts is not part of a GFI's job, but it was a cheap upgrade to safety. Especially as when in 120V/240V work we can often use one N wire for two loads (if it would stay tight).
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.