'Noisy' Loop Impedance Measurements with Fluke 1652?

[JohnW2]

Hi folks,

A few weeks back, I went through a period of a week or two during which almost every attempt I made to measure the L-E loop impedance on any circuit in my house (or at the origin of the installation) with my Fluke 1652 resulted in an "Err 5" message, which apparently means "mains is too noisy- accuracy may be degraded". When that happens, there's a button one can press to see the measured value ('despite the noise') and the reading I got when I did that was sometimes credible, sometimes not. There was never any such problem with L-N loop measurements.

As I said, that problem only persisted for a week or two. Since then (a good few weeks) it has always worked as usual, giving credible (and consistent) readings, and never any error message.

What is this "noisy mains" all about? What sort of 'noise', and why should it appear for a while and then go away? Has anyone else experienced something like this?

Edit: I should perhaps have added ... in the several years I've been using 1652s in my house, I have never experienced this before.

Kind Regards, John

 

[CBW]

Hi John,

Is this worth a read? I know it mentions 60Hz (US), but still might help explain?

 

[JohnW2]

Hi John, Is this worth a read? I know it mentions 60Hz (US), but still might help explain?

Thanks. I understand what electrical noise is, and how it can arise. However, I don't know what sort of noise upsets the meter's loop impedance measurements nor why (given that nothing here changed) the issue suddenly appeared (for a week or two) and then spontaneously disappeared, after several years during which it had never previously happened.

Furthermore, although the manual calls it 'noisy mains' the fact that the issue showed itself only in relation to L-E (not L-N) loop measurements might suggest that there was a problem with the 'earth' (theoretically TT but, in reality, seemingly the neighbour's TN-C-S, via a water pipe!), rather that with 'noise' on either/both of the live conductors.

Kind Regards, John

 

[RandomGrinch]

Sorry, probably not helpful in determining a cause, but a search led me here:

Fluke 1653 Error 5

Did an earth loop test on a water heater today in an office block at the factory where I work. The Fluke 1653 came up with Err5. When I checked this in the manual, I found that it stood for Error 5-excessive noise. When I checked a few other parts of the area, they too came up with the same...

www.diynot.com

...and the last comment was quite interesting:

My KT64 has a display symbol 'noise'. The manual states that when doing EFLI testing and this displays you can get a reading by turning off ATT (anti trip technology). ATT is to stop RCD trip during test. Not sure if Fluke has this feature?

There are no settings for disabling the anti-trip on the Fluke, and if this is to be believed, the "Err 5", is a common occurrence:

1653 fluke error

Hi guys I use a fluke 1653 mft for testing at work. We work on mainly TT systems so when zs testing it is on no-trip. I don't have no instructions for this tester and more and more increasing I am getting a error 5 when zs testing. Can anyone tell me what a error 5 is and what a error 4 is as...

www.electriciansforums.net

 

[JohnW2]

Sorry, probably not helpful in determining a cause, but a search led me here: .... .and the last comment was quite interesting:

Thanks.

There are no settings for disabling the anti-trip on the Fluke, and if this is to be believed, the "Err 5", is a common occurrence:

It is true that the 1652 offers nothing other than 'no trip' loop impedance measurements, but other models in the range (I think including the 1653) do.

However, as you say, from what you've discovered, it sounds like quite a common problem - even though, as I said, in several years of using 1652s its the first time I have experienced it. Even stranger is the fact that, for a week or two, it happened almost every time I tried, on many different circuits in my house, but then stopped doing it as suddenly as it has started. That suggests that, although I am unaware of anything having changed in the house during that period, it seems that something, somewhere, had changed for that few days.

@EFLImpudence ... I think you've been using a 1652 for a while-have you exp[evidenced this?

Kind Regards, John

 

[EFLImpudence]

.. I think you've been using a 1652 for a while-have you exp[evidenced this?

I never encountered the problem.

Although did take advantage of a company once who were selling two 1652s really cheap because they had error 5 faults.

There are no settings for disabling the anti-trip on the Fluke, and if this is to be believed, the "Err 5", is a common occurrence

You can do high current tests with the 1652.

 

[JohnW2]

I never encountered the problem.

Thanks. As I said, nor had I (in several years use), until that brief (1-2 week) period recently, during which is was happening nearly all the time

Although did take advantage of a company once who were selling two 1652s really cheap because they had error 5 faults.

I would think that was silly of them. As far as I can see (and presumably as I have experienced) unlike the "Err 1 message of doom", Err5 does not usually mean that there is anything wrong with the machine. Did you buy one (or more) and, if so, did you experience any Err5 problems with it?

You can do high current tests with the 1652.

1652B and 1652C can - but, unless I need a serious talk with Mr Specsavers, a 1652 (like mine) cannot).

If you show me a photo of your 'function selector knob' , I'll show you mine

Kind Regards, John

 

[EFLImpudence]

I would think that was silly of them. As far as I can see (and presumably as I have experienced) unlike the "Err 1 message of doom", Err5 does not usually mean that there is anything wrong with the machine. Did you buy one (or more) and, if so, did you experience any Err5 problems with it?

Yes, bought both; still have one and sold the other for more than the price of the two.

No, I have never encountered the problem.

1652B and 1652C can - but, unless I need a serious talk with Mr Specsavers, a 1652 (like mine) cannot).
If you show me a photo of your 'function selector knob' , I'll show you mine

It's not a question of what your knob looks like. Mine is the same as yours.

The three-wire loop test is an automatic non-trip test function and the two-wire L-N test is the standard high current loop test; with the N lead connected to Earth it is a high-current Earth loop test. See page 27 of manual.

 

[JohnW2]

Yes, bought both; still have one and sold the other for more than the price of the two. ... No, I have never encountered the problem.

Lovely! That show (per what I suggested) how silly they were in "inviting themselves to be ripped off" by not understanding what Err5 does (and does not) mean!

It's not a question of what your knob looks like. Mine is the same as yours. The three-wire loop test is an automatic non-trip test function and the two-wire L-N test is the standard high current loop test; with the N lead connected to Earth it is a high-current Earth loop test. See page 27 of manual.

My goodness, one learns something (sometimes after years of being unaware) every day - many thanks!

I had always wondered why L-N loop measurements were essentially instantaneous, whereas the L-E ones took several seconds, but this is clearly the explanation. It is true that, as you say, and although it does not go onto say anything about actually doing this, the manual says:

You can use this to measure:
..... • L-PE loop measurement. This is a way of making a high current, 2-wire loop measurement. It cannot be used on circuits protected by RCDs because it will cause them to trip.

So, if I connect the meter's N input to E rather than N, then an "L-N" test will then conduct a (presumably 'instantaneous') high-current L-E test?

If so, is that actually useful ('useable') in practice? In other words, does it have enough time to produce an accurate reading in the the fairly small number of milliseconds before an RCD trips? If not, and given that almost all circuits are now required to be RCD-protected, this facility would seem to be of very limited usefulness (applicable only to the few remaining non-RCD circuits or measurements at the CU, upstream of any RCDs)?

Whatever, many thanks again for bringing this to my attention. I'm amazed that I have been using the machines for so many years (and probably reading bits of the manual hundreds of times) without becoming aware of this!!

Kindest Regards, John

 

[EFLImpudence]

So, if I connect the meter's N input to E rather than N, then an "L-N" test will then conduct a (presumably 'instantaneous') high-current L-E test?

Not sure what is "instantaneous" in this regard.

If so, is that actually useful ('useable') in practice? In other words, does it have enough time to produce an accurate reading in the the fairly small number of milliseconds before an RCD trips?

Well, it obviously dates back to before the RCD was so commonplace. Possibly better for measuring Ze. You could compare the results.

Not an expert but I had assumed it gave a more accurate reading than the no-trip test although I think the Fluke no-trip test is quite accurate compared with other meter manufacturers; one of whom in particular frequently seems to cause problems.

If not, and given that almost all circuits are now required to be RCD-protected, this facility would seem to be of very limited usefulness (applicable only to the few remaining non-RCD circuits or measurements at the CU, upstream of any RCDs)?

Quite.

Whatever, many thanks again for bringing this to my attention. I'm amazed that I have been using the machines for so many years (and probably reading bits of the manual hundreds of times) without becoming aware of this!!

You are welcome. As you say, though, you likely have rarely actually needed it.

 

[JohnW2]

Not sure what is "instantaneous" in this regard.

I simply mean that the reading (for an L-N measurement) appears almost before one's finger has had time to get off the 'Test' button, whereas there is a delay of a good few (sometimes many) seconds before a L-E reading appears.

Well, it obviously dates back to before the RCD was so commonplace. Possibly better for measuring Ze. You could compare the results. Not an expert but I had assumed it gave a more accurate reading than the no-trip test although I think the Fluke no-trip test is quite accurate compared with other meter manufacturers; one of whom in particular frequently seems to cause problems.

The implication presumably must be that 'high-current' tests are more accurate than 'no trip' ones, otherwise there would be no point in their existence, but I don't know how they compare.

I've never really looked into it, but I don't really know how the 'no trip' tests are done. On the face of it, if the test is not going to trip a 30mA RCD, the current through the E has to be limited to 15mA, implying an effective ~15,333Ω impedance inserted in the path. However, that would mean that to measure, say, an EFLI of 1Ω the machine would have to distinguish between, say, 15,333Ω and 15,334Ω - and any measurement dependent upon "a very small difference between very large numbers" is a mathematician's nightmare (in terms of potential accuracy). Have you any idea how it's done?

You are welcome. As you say, though, you likely have rarely actually needed it.

Although you don't say so explicitly, I presume you are saying that (as I expected), a high-current measurement on an RCD-protected circuit (i.e. a test which will trip the RCD) will not result in (any, or an accurate) reading? If so, it's not so much a question of 'need' as of 'what is possible'.

The point is that RandomGrinch introduced material suggesting that the Err5 issue only arises in relation to 'no-trip' measurements, so that the way to avoid it is not to use 'no trip' tests. However, if that is something which is increasingly impossible to do, it does not really represent much of a 'solution' to the issue!

Kind Regards, John

 

[EFLImpudence]

I've never really looked into it, but I don't really know how the 'no trip' tests are done. On the face of it, if the test is not going to trip a 30mA RCD, the current through the E has to be limited to 15mA, implying an effective ~15,333Ω impedance inserted in the path. However, that would mean that to measure, say, an EFLI of 1Ω the machine would have to distinguish between, say, 15,333Ω and 15,334Ω - and any measurement dependent upon "a very small difference between very large numbers" is a mathematician's nightmare (in terms of potential accuracy). Have you any idea how it's done?

I believe (don't know where I read it - might be wrong) that the fluke pulses for very short periods of time so that the RCD doesn't have time to react rather than apply very small currents.

Although you don't say so explicitly, I presume you are saying that (as I expected), a high-current measurement on an RCD-protected circuit (i.e. a test which will trip the RCD) will not result in (any, or an accurate) reading? If so, it's not so much a question of 'need' as of 'what is possible'.

Yes, it will trip the RCD.
I meant since RCDs are everywhere now you might never have the opportunity to use the high current test except as I said for Ze.

The point is that RandomGrinch introduced material suggesting that the Err5 issue only arises in relation to 'no-trip' measurements, so that the way to avoid it is not to use 'no trip' tests. However, if that is something which is increasingly impossible to do, it does not really represent much of a 'solution' to the issue!

I don't know. One can always bypass the RCD.

 

[JohnW2]

I believe (don't know where I read it - might be wrong) that the fluke pulses for very short periods of time so that the RCD doesn't have time to react rather than apply very small currents.

I hadn't though of that but (now you've mentioned it) it does sound like an 'obvious' way to do it. However, if so, it presumably must do so 'thousands of times' (to get n average?), since it can easily take 10-20 seconds.

Yes, it will trip the RCD. I meant since RCDs are everywhere now you might never have the opportunity to use the high current test except as I said for Ze.

Agreed. I will try to see if it even attempts to produce an L-E reading before the RCD trips.

I don't know. One can always bypass the RCD.

One can, but it would be a little tedious to have to do that to undertake an EFLI measurement.

Kind Regards, John

 

[JohnW2]

I believe (don't know where I read it - might be wrong) that the fluke pulses for very short periods of time so that the RCD doesn't have time to react rather than apply very small currents.

A new development/update.

As you (EFLI) will know, when undertaking 'no trip'loop impedance measurements with a 1652, by pressing the appropriate button one can choose between a display of Ze + PFC or Ze + Re. I have always assumed that, in the latter case, it measures L-N and well as L-E loop impedance and then displays the difference as "Re". Is that your understanding?

Today (with Ze+PFC chosen), I started getting the Err5 again, for the first time in quite a while. However, I discovered (essentially 'by accident') that if I select the Ze+Re test/display, I don't get the Err5 - and with things as they are at present, that's consistent - always Err5 with Ze+PFC, but never with Zs+Re at the moment (at least, on one circuit).

Any thoughts?

Kind Regards, John

 

[EFLImpudence]

As you (EFLI) will know, when undertaking 'no trip'loop impedance measurements with a 1652, by pressing the appropriate button one can choose between a display of Ze + PFC or Ze + Re.

Yes - it would actually be PEFC. It must assume you would know that to save a display digit.

I have always assumed that, in the latter case, it measures L-N and well as L-E loop impedance and then displays the difference as "Re". Is that your understanding?

I don't know how it does it but it would be "the difference plus half of L-N" wouldn't it?

Does it display similar with the two-wire test? I can't remember - or have never noticed.
I note there is a "PSC" display so it must show that even if you are measuring L-E. (I haven't looked at the instructions). It wouldn't know.

Today (with Ze+PFC chosen), I started getting the Err5 again, for the first time in quite a while. However, I discovered (essentially 'by accident') that if I select the Ze+Re test/display, I don't get the Err5 - and with things as they are at present, that's consistent - always Err5 with Ze+PFC, but never with Zs+Re at the moment (at least, on one circuit).

Interesting.

I was thinking that the machine would do both calculations but just display the one selected - perhaps not.
So what would be the difference between the two?

Any thoughts?

Not really apart from above. I have never had an Error 5.