Electricians: measuring an single appliance's earth leakage current?

[Howardlong]

It could be both, i.e. the surge protector triggered by surges from switching inductive loads and the leakage current of all the EMI filters (in SMPSUs) added up. If you have plenty of larger SMPSUs it would make sense to distribute them across multiple RCDs to keep each group's leakage current below the RCD's trigger current. Having multiple RCD zones also prevents a totally dark house/lab/workshop.

The consumer unit only has a single RCD for all socket outlets.

That's a standard design for simple and inexpensive surge protectors, often used in power strips or adapters.

OK, thank you.

[nightfire]

Do PAT testers actually measure AC earth leakage current?

more expensive models will,but as its not a requirement for pat testing you'd have to double check the particular models specs

Here in germany, if you do appliance testing according to VDE 701/702, it is AC+DC you have to measure also.
Therefore may depend on model of PAT tester you get which modes it supports.

[Monkeh]

These surge protectors show 80MΩ at 250V, but only 300kΩ at 500V, while all of the other appliances, splitters and power strips were tens of MΩ if not into the GΩ.

What were you expecting a 470V MOV to do at 500V other than become conductive?

[mikeselectricstuff]

On a related subject - this cheap (£27 Uk delivered) clamp-meter is excellent for investigating earth leakage -
0.1mA resolution on AC, 1mA on DC
https://www.aliexpress.com/item/1005006582824997.html

Do you have one?

yes ( actually two, one kept in car as emergency DMM).

[Howardlong]

These surge protectors show 80MΩ at 250V, but only 300kΩ at 500V, while all of the other appliances, splitters and power strips were tens of MΩ if not into the GΩ.

What were you expecting a 470V MOV to do at 500V other than become conductive?

Er, think about it. How would I have known it had a 470V MOV in there? I'm not telepathic, perhaps you are.

I'd identified the surge protector was behaving substantially differently to the dozens of other tests I'd done on other appliances and PDUs. The tester I have tests at a variety of discrete voltages, including 50, 100, 250, 500 and 1000V.

When I removed the surge protector, the trips stopped happening.

It was only then that I took the semi-potted surge protector apart to see what was inside.

[Howardlong]

On a related subject - this cheap (£27 Uk delivered) clamp-meter is excellent for investigating earth leakage -
0.1mA resolution on AC, 1mA on DC
https://www.aliexpress.com/item/1005006582824997.html

Do you have one?

yes ( actually two, one kept in car as emergency DMM).

How well do they behave when tracing earth leakage with both L+N tails in the clamp? Are the results consistent?

[amyk]

These surge protectors show 80MΩ at 250V, but only 300kΩ at 500V, while all of the other appliances, splitters and power strips were tens of MΩ if not into the GΩ.

What were you expecting a 470V MOV to do at 500V other than become conductive?

Er, think about it. How would I have known it had a 470V MOV in there? I'm not telepathic, perhaps you are.

I'd identified the surge protector was behaving substantially differently to the dozens of other tests I'd done on other appliances and PDUs. The tester I have tests at a variety of discrete voltages, including 50, 100, 250, 500 and 1000V.

When I removed the surge protector, the trips stopped happening.

It was only then that I took the semi-potted surge protector apart to see what was inside.

What do you expect a surge protector to do when subjected to a voltage much greater than normal mains?

[Andy Chee]

These surge protectors show 80MΩ at 250V, but only 300kΩ at 500V, while all of the other appliances, splitters and power strips were tens of MΩ if not into the GΩ.

Properly functioning RCDs normally trip at 30mA.

300kΩ at 500V is only 1.66mA.

So logically it must be something else.

[Monkeh]

These surge protectors show 80MΩ at 250V, but only 300kΩ at 500V, while all of the other appliances, splitters and power strips were tens of MΩ if not into the GΩ.

What were you expecting a 470V MOV to do at 500V other than become conductive?

Er, think about it. How would I have known it had a 470V MOV in there? I'm not telepathic, perhaps you are.

I'd identified the surge protector was behaving substantially differently to the dozens of other tests I'd done on other appliances and PDUs. The tester I have tests at a variety of discrete voltages, including 50, 100, 250, 500 and 1000V.

When I removed the surge protector, the trips stopped happening.

It was only then that I took the semi-potted surge protector apart to see what was inside.

You didn't mention any other surge protectors. Yes, many lesser types will have 560V or greater MOVs, but it's quite expected to see things start to conduct on a 500V IR test.

It sounds far more like you're hovering on the threshold of an overly sensitive RCCB with too many circuits on it.

[Howardlong]

You didn't mention any other surge protectors.

I did, and I put it in bold.

[Howardlong]

These surge protectors show 80MΩ at 250V, but only 300kΩ at 500V, while all of the other appliances, splitters and power strips were tens of MΩ if not into the GΩ.

Properly functioning RCDs normally trip at 30mA.

300kΩ at 500V is only 1.66mA.

So logically it must be something else.

It may be a combination of things, or an accumulated combination of things, however, a previously repeatable trip scenario no longer happens.

[Howardlong]

These surge protectors show 80MΩ at 250V, but only 300kΩ at 500V, while all of the other appliances, splitters and power strips were tens of MΩ if not into the GΩ.

What were you expecting a 470V MOV to do at 500V other than become conductive?

Er, think about it. How would I have known it had a 470V MOV in there? I'm not telepathic, perhaps you are.

I'd identified the surge protector was behaving substantially differently to the dozens of other tests I'd done on other appliances and PDUs. The tester I have tests at a variety of discrete voltages, including 50, 100, 250, 500 and 1000V.

When I removed the surge protector, the trips stopped happening.

It was only then that I took the semi-potted surge protector apart to see what was inside.

What do you expect a surge protector to do when subjected to a voltage much greater than normal mains?

I don't expect it to trip an RCD when a combination of other appliances are turned on.

[Andy Chee]

These surge protectors show 80MΩ at 250V, but only 300kΩ at 500V, while all of the other appliances, splitters and power strips were tens of MΩ if not into the GΩ.

Properly functioning RCDs normally trip at 30mA.

300kΩ at 500V is only 1.66mA.

So logically it must be something else.

It may be a combination of things, or an accumulated combination of things, however, a previously repeatable trip scenario no longer happens.

In which case, I think it more pertinent to measure the earth leakage of your system, rather than the MOVs in one particular device.

[themadhippy]

Properly functioning RCDs normally trip at 30mA.

30mA is the point at which it must trip,it may trip at less than that,anything down to 15ma is acceptable.

[Monkeh]

You didn't mention any other surge protectors.

I did, and I put it in bold.

.. that's you mentioning this surge protector (and from another post, others of this type). Not other surge protectors.

Even if this device is leaking slightly at normal operating voltages, it's not significant enough to be a root cause.

Have the circuits protected by the RCCB undergone insulation resistance testing, not merely the appliances?
Has the RCCB been tested to ascertain its tripping characteristics in practice?
Have all factors been accounted for with these trips, such as routinely connected appliances which are not double pole isolated or are operated by other members of the household who may not recognise the relevance of them? (I don't drink hot drinks - nobody else in the household grasps electricity. Imagine how long it took me to figure out a kettle was the cause of nuisance trips when the answer to "what were you doing when the power went off?" was "nothing".)

[madires]

BTW, the usual leakage current of an ATX PSU is roughly 0.5 mA (at 230 V). Some have 1 mA. And the max limit is 3.5 mA based on IEC 60950-1.

[mikeselectricstuff]

On a related subject - this cheap (£27 Uk delivered) clamp-meter is excellent for investigating earth leakage -
0.1mA resolution on AC, 1mA on DC
https://www.aliexpress.com/item/1005006582824997.html

Do you have one?

yes ( actually two, one kept in car as emergency DMM).

How well do they behave when tracing earth leakage with both L+N tails in the clamp? Are the results consistent?

Not bad, but it is a bit sensitive to external fields, like if you have 32A charging the car on a nearby breaker, so ideally need to turn off other breakers & test leakage on one circuit at a time. 

[AVGresponding]

Presumably when the SMPS inside these appliances is plugged in or turned on, it's enough to send short spikes enough to wake up the surge protector, and for it to cause an RCD fault. Possibly over time these MOVs degrade to some degree as they absorb repeated stresses.

Yes, they do degrade over time.

Properly functioning RCDs normally trip at 30mA.

300kΩ at 500V is only 1.66mA.

So logically it must be something else.

Properly functioning RCDs normally trip at 30mA.

30mA is the point at which it must trip,it may trip at less than that,anything down to 15ma is acceptable.

Typically 18-25mA is the goal the manufacturers shoot for. These can also degrade over time, and become inconsistent.

With the plethora of things that use bloody class Y caps, it's no surprise that an older installation with one main RCD could be pushed a bit hard.

How well do they behave when tracing earth leakage with both L+N tails in the clamp? Are the results consistent?

Not bad, but it is a bit sensitive to external fields, like if you have 32A charging the car on a nearby breaker, so ideally need to turn off other breakers & test leakage on one circuit at a time.

You just need to get your clamp over the main earth to measure your total earth leakage current. And as Mike says, you should test one circuit at a time when chasing down issues.

[Howardlong]

You just need to get your clamp over the main earth to measure your total earth leakage current. And as Mike says, you should test one circuit at a time when chasing down issues.

The problem with this approach is that it doesn't catch any earth leakage that doesn't come down the "main earth". That's why it's recommended to clamp both L and N at the same time.

Another problem is that I believe that the MCBs I have in my CU are single only pole.

Legally speaking, ISTBC, but I believe that anything inside the CU needs a qualified electrician here in the UK.

[Andy Chee]

Another problem is that I believe that the MCBs I have in my CU are single only pole.

Not a problem. 

If your MCB only switches L and omits the N connection, then that MCB is NOT an RCD/RCBO, and can be ignored insofar as measuring earth leakage is concerned.

Can you highlight the breaker that is "tripping" from your surge protector?  Is it even in the photo?

[mikeselectricstuff]

Legally speaking, ISTBC, but I believe that anything inside the CU needs a qualified electrician here in the UK.

It's a complicated mess of building regs and "competent" persons schemes, (where "competent" is defined as paying a sub to one of the cartels like NIC/IEC or NAPIT).
In reality, if you actually know what you're doing then it's highly unlikely to ever be an issue unless you ever sell the house & get asked for paperwork, in which case you just claim ignorance about any work that may have been done in the past by some electrician who's now long-gone,  and just offer to get an EICR done to show that the installation is safe.

Historical aside - they changed wiring colours from red/black to brown/blue around the time they brought in the Part P regs on domestic electrical works  (though there was some overlap, about a year or so ISTR). There's a reason there's plenty of ebay listings for "old colours" twin & earth cable.

[mikeselectricstuff]

Another problem is that I believe that the MCBs I have in my CU are single only pole.

Not a problem. 

If your MCB only switches L and omits the N connection, then that MCB is NOT an RCD/RCBO, and can be ignored insofar as measuring earth leakage is concerned.

I think there may be some confusion between both L+N physically passing through the RCD/RCBO, vs. 2-pole switching of both L+N.
 
Most RCBOs (at least in UK) only switch the Live, with Neutral just passing through for residual current measurement. This means that any neutral-earth leakage will still be there when it's turned off, and a N-E short on that circuit will trip any upstream RCD

[mikeselectricstuff]

You just need to get your clamp over the main earth to measure your total earth leakage current.

Unless some of that leakage isn't leaking to the mains earth wire. 

[themadhippy]

Legally speaking, ISTBC, but I believe that anything inside the CU needs a qualified electrician here in the UK.

your not replacing the consumer unit,nor adding or making changes to the existing installation and the rules do say

Non-notifiable work
2.7 Regulation 12(6A) sets out electrical installation work that is notifiable. All other electrical
installation work is not notifiable – namely additions and alterations to existing installations outside
special locations, and replacements, repairs and maintenance anywhere

[Andy Chee]

Another problem is that I believe that the MCBs I have in my CU are single only pole.

Not a problem. 

If your MCB only switches L and omits the N connection, then that MCB is NOT an RCD/RCBO, and can be ignored insofar as measuring earth leakage is concerned.

I think there may be some confusion between both L+N physically passing through the RCD/RCBO, vs. 2-pole switching of both L+N.
 
Most RCBOs (at least in UK) only switch the Live, with Neutral just passing through for residual current measurement. This means that any neutral-earth leakage will still be there when it's turned off, and a N-E short on that circuit will trip any upstream RCD

Indeed.  When I Googled the manufacturer of the breakers shown in the posted image above, the breakers I found only had a single physical screw terminal on the line and load sides.

So unless I Googled the incorrect model breakers, the ones shown in the posted image above CANNOT be RCD/RCBOs because they do not have both L+N.

In other words, RCD/RCBOs must have L+N physically passing through them.  They may switch L only, or switch both L+N, but they MUST have both L+N physical connections.

I have attached a photo; this is NOT a RCD/RCBO.



In contrast this is a RCD/RCBO.  Note the extra neutral connection.  Also the big tell-tale is the TEST button that all RCD/RCBOs have.