Video: Why NCV is dangerous when used without proper knowledge for wrong task

[Cymaphore]

Hi,

for the people around here it is probably pure comedic in value how I use a couple of meters to do some clumsy NCV probing of mainly USB cables - but you can see my new BM789 in action :-)



After I got the impression, that many people seem to see EF as some sort of improved voltage testing, I could not resist to make a video about it.

Probably you'll notice instantly which YouTubers from the EEVblog community I regularly watch, due to some gestures I picked up without noticing before.

i am not quite pleased with the result, too much irony and too pompous, and I think I regularly point to facts without proper explanation why it is that way...

But I uploaded it now, so be it.

Used in the video:

  * BRYMEN BM789
  * RS PRO S2
  * ANENG A3008
  * APPA 605
  * BENNING duspol digital

What do you think?

Best regards,
Martin

[RayRay]

It's been my experience as well (dealing with NCV testers both standalone and in meters) that they can also detect low DC and beep intensely as if it's high voltage (regardless of the specs claimed by the manufacturer) and yes, some of em definitely have better sensitivity than others, and rarely, you might encounter a completely faulty one, however, when used correctly, they're not even remotely useless! The primary purpose of NCV testing is to protect a person from possible shock/electrocution when dealing with high voltage stuff, for example, if you've switch off the breaker for particular room because you need to change an electrical outlet or a lighting switch in it, and you wanna verify it's safe to work on before you begin, and in that regards (again, when used correctly) an NCV tester is very useful and practical! Firstly, you're suppose to check the sensitivity of an NCV tester when you get it (if it's really bad, then you should get a better one and not rely on it!) and secondly, for safety, you're suppose to always test the NCV tester on a live outlet before using it for production (literally minutes before each time) so you'd know for a fact it's still in good working condition, because otherwise, you'd literally be risking your life (if you'd just presume it's still in good condition when it's not, it'd give you a false reading and you could get electrocuted!) so again, it should be used responsibly, and not blindly. It's also possible to get a reading with multimeter in addition (for extra safety) but IMO, it's not really needed when used correctly. Btw, "feature" and "future" are pronounced differently 

[Cymaphore]

My pronunciation leaves lots of room for improvement.... :-)

you wanna verify it's safe to work on before you begin, and in that regards (again, when used correctly) an NCV tester is very useful and practical!

When the circuit in question runs next to many others for some length or next to a cable under load, NCV usually thinks it's on. Similar to the DMM, where you would switch to LoZ to kill the phantoms.

That's why the EN Standard for voltage testers demands a relatively low input impedance (usually under 200k based on voltage ratings if I remember correctly) and mandatory ability to function even without batteries.

Because it has to give a rock solid answer and leave no room for guessing or things to go wrong. You don't want people to guess about results, especially when they are on the clock.

Even the capacitive test (single probe life detection) tend to false negatives, if you don't hold it correctly.

So why not just recommend the good old two pole voltage testers with these features as additions instead? It always gives reliable answers and leaves little room for misuse or interpretation.

[Kleinstein]

It's been my experience as well (dealing with NCV testers both standalone and in meters) that they can also detect low DC and beep intensely as if it's high voltage (regardless of the specs claimed by the manufacturer) and yes, some of em definitely have better sensitivity than others, and rarely, you might encounter a completely faulty one, however, when used correctly, they're not even remotely useless!

The NCV is not decting DC, but it is detecting common mode capacitive coupled AC residue found with essentially all SMP due to the EMI caps.
I would consider the NCV nearly useless, espeically with only a simple yes / no reading. One of the few good uses is to check for hidden wires in the wall, trying to follow a wire. This needs good sensisitity and at least some bar graph to pinpoint the maximum.

Chances are NCV may do more harm than good. There is a danger from not decting a life cable and not all users are aware the NCV is so poor reliability. The problem is in the principle, and the meter quality makes little difference - the best implementation is not to have NCV at all at a meter.
NCV is at best a first guess - before touching use a real meter.

[_Wim_]

In the company I work for the safety department forbids to use NCV testers, both for internals and external contractors. Only "proper" meters are allowed.

[Brumby]

I believe there is a place for NCVs - and even then, only when used by someone who understands the limitations and risks.

Many moons ago I worked for a friend at a Tandy store (how Radio Shack landed on Australian shores).  He was the manager and was very focussed on sales.  He didn't have a big technology knowledge base ... outside of what he needed to know to sell a product.  I did, however, and that was something I had to keep in check, otherwise I could waste a lot of time.

One day, a customer came in looking at the NCV they had for sale and I served them.  They started asking questions that indicated they knew nothing and when they started venturing down the path of getting me to educate them, I baulked.  I quickly backed away and told them it was not my place to train them and that I was not prepared to sell the item to them.  They looked disappointed, but left without issue.  I then told my boss what had happened and the action I took - which he quietly accepted.

I could not, in all good conscience, let that customer loose around anything electrical bigger than a 9V battery.

[bc888]

Thank you for the video Martin. I enjoyed it and you made your point quite capably. I appreciate that you shared your findings. My main useage, and to me it's important is this. 2 months ago, I used the Eevlog BM789  NCV to map the hidden wires behind the wall in my house so I could saw the drywall out and install a wall safe.

I prefer the new NCV method of finding the non-visible wires and their orientation before I saw into the wall much more than my traditional method,. Which is typically to guess where I believe the wiring will be, and then saw right through live wiring. OK, it was much more common when I was younger and had never heard of a NCV checker. The 789 Dave sold me has a fantastic NCV meter, better than anything I've used before.

Now, a smarter person than I would have turned off the breakers beforehand. However, the breakers being unmarked and a wife using the electricity along with all of the programmable electronics in the home makes that rather problematical.

[threephase]

It is an interesting video, to some extent I would agree with you that a NCV function on a multimeter is not all that useful in terms of what a multimeter is usually used for in the electronics industry. NCV testers can have their uses though.

With regards to verifying that something is safe to touch, then yes the two pole voltage tester you showed is the ideal choice and is always what I use. I would add in a caveat though, in that I follow a specific test procedure for verifying something is safe to touch of prove-test-prove. Just testing with a two pole probe without this procedure doesn't show beyond doubt that the circuit is safe to touch.

i.e. I test my voltage probe on a known source first, usually a proving unit, that matches the circuit I am working on in terms of AC / DC and the voltage level. I then test what I intend to work on and afterwards test the voltage probe on my known source again.

An NCV tester can be used as an addition to the above procedure, especially in the UK where some installations can have a combined neutral-earth connection and a break in this connection in certain locations can fool a two pole voltage probe in showing that the supply is dead when it is still actually live.

I actually carry a proving unit that can test both my two pole voltage tester and my NCV tester.

An NCV tester is also good for finding breaks in cables where there is no externally visible damage and yes as other have stated can find cables buried within walls, as long as there is no earthed metal capping around them, or they aren't inside conduit.

[AVGresponding]

It is an interesting video, to some extent I would agree with you that a NCV function on a multimeter is not all that useful in terms of what a multimeter is usually used for in the electronics industry. NCV testers can have their uses though.

With regards to verifying that something is safe to touch, then yes the two pole voltage tester you showed is the ideal choice and is always what I use. I would add in a caveat though, in that I follow a specific test procedure for verifying something is safe to touch of prove-test-prove. Just testing with a two pole probe without this procedure doesn't show beyond doubt that the circuit is safe to touch.

i.e. I test my voltage probe on a known source first, usually a proving unit, that matches the circuit I am working on in terms of AC / DC and the voltage level. I then test what I intend to work on and afterwards test the voltage probe on my known source again.

An NCV tester can be used as an addition to the above procedure, especially in the UK where some installations can have a combined neutral-earth connection and a break in this connection in certain locations can fool a two pole voltage probe in showing that the supply is dead when it is still actually live.

I actually carry a proving unit that can test both my two pole voltage tester and my NCV tester.

An NCV tester is also good for finding breaks in cables where there is no externally visible damage and yes as other have stated can find cables buried within walls, as long as there is no earthed metal capping around them, or they aren't inside conduit.

This pretty much says it all, though I would add three common examples of cable that a NCV won't work on; FP, MICC, and SWA.

It's a continuous battle to keep apprentices from doing stupid things, and sad to say there are plenty of sparkies that set a bad example by not proving their test gear.

[threephase]

Yes, pretty much any cable with an earthed outer screen or armour could present problems for NCV detection. I don't work on FP and MICC rarely, but SWA is usually an issue, unless the installer forgets to earth the armour of course. Then NCV works a treat on single core cables.

[AVGresponding]

That really winds me up, as does leaving a floating conductor (a frequent occurrence when people can't find a 3-core Klik lead and use a 4-core, they just cut the black at both ends. So much for the regs).

[joeqsmith]

I thought you did a good job with the video.   

I'm not an electrician and would have little need for the feature.   I do have a couple of other types of non-contact voltage meters.   I have a very old electrostatic meter that's pretty interesting to look at.  It uses a mechanical chopper.  After reading the patent, I spoke with the designer.  I was curious why it had been developed and they explained some of the industry applications.  Interesting stuff.   

After several people asked, I bought a non-contact Fluke meter that can measure voltage.  The patents for this are a pretty good read.  I thought about making my own to see if I could improve on what they had but lost interest before I even got started.     We ran a lot of tests on it.  This isn't my video but gives a basic overview of the meter.       

[Cymaphore]

After several people asked, I bought a non-contact Fluke meter that can measure voltage.  The patents for this are a pretty good read.  I thought about making my own to see if I could improve on what they had but lost interest before I even got started.     We ran a lot of tests on it.  This isn't my video but gives a basic overview of the meter.

Would be very interesting to see a review from you about it. How sensitive is it on ghost voltages?

Personally I would not see much of a use case for me, since it's lacking the loz/load function like unfortunately all the fork-style voltage testers I know of.

I prefer the new NCV method of finding the non-visible wires and their orientation before I saw into the wall much more than my traditional method,. Which is typically to guess where I believe the wiring will be, and then saw right through live wiring. OK, it was much more common when I was younger and had never heard of a NCV checker. The 789 Dave sold me has a fantastic NCV meter, better than anything I've used before.

Personally I would recommend using a detector designed for that task (like the Bosch GMS 120). These devices usually give you a much better indication. They can also detect metal in general and sometimes even other material.

An NCV tester can be used as an addition to the above procedure, especially in the UK where some installations can have a combined neutral-earth connection and a break in this connection in certain locations can fool a two pole voltage probe in showing that the supply is dead when it is still actually live.

These TNC-installations up to the socket are a pure plague and still widespread in older installations in germany as well.

But for that problem I rather have a 2-pole voltage tester with capacitive test function and continuity. By my experience, the capacitive test is very reliable when facing broken N, PE or PEN connections and works much better than NCV on lines without load.

[madires]

I would consider the NCV nearly useless, espeically with only a simple yes / no reading. One of the few good uses is to check for hidden wires in the wall, trying to follow a wire. This needs good sensisitity and at least some bar graph to pinpoint the maximum.

Yep, I use it to check for cables before drilling holes into a wall.

[joeqsmith]

Would be very interesting to see a review from you about it. How sensitive is it on ghost voltages?

I suggest you watch this video instead.  He clearly demonstrates in very simple terms all the problems and just how dangerous  the meter is. 


After he release that video, I started having people ask me about running it.  There are three parts to it.     

[Fungus]

I believe there is a place for NCVs - and even then, only when used by someone who understands the limitations and risks.

Yep.

If it gives a positive reading, it's probably right and you need to be careful.

A negative reading? Doesn't mean much unless the situation is very clear and the same cable gives a positive reading when you toggle the switch on/off. It certainly doesn't mean the cable is safe to touch.

[Brumby]

I believe there is a place for NCVs - and even then, only when used by someone who understands the limitations and risks.

Yep.

If it gives a positive reading, it's probably right and you need to be careful.

A negative reading? Doesn't mean much unless the situation is very clear and the same cable gives a positive reading when you toggle the switch on/off. It certainly doesn't mean the cable is safe to touch.

This pretty much sums up the way I use these things.  Used for initial indication, but backed up by more positive checking before potentially safety compromising actions are taken.

[tggzzz]

I could not resist to make a video about it.

I'm not going to spend 15 minutes of my remaining life watching to see if there is something I don't know.

Is there any chance of you indicating the content via a few "bullet points" indicating what I need to understand? If they pique my interest, I will watch.

Dave does that with his videos, to great effect. Bonus: words are found by search engines, speech in videos isn't.

[Cymaphore]

If it gives a positive reading, it's probably right and you need to be careful.

You always need to be careful. And it regularly gives false positives by my experience. So it's maybe right or maybe wrong, unless you use an actual voltage tester.

I'm not going to spend 15 minutes of my remaining life watching to see if there is something I don't know.

Is there any chance of you indicating the content via a few "bullet points" indicating what I need to understand? If they pique my interest, I will watch.

Dave does that with his videos, to great effect. Bonus: words are found by search engines, speech in videos isn't.

I demonstrated how NCV gives mostly wrong readings detecting dangerous voltage levels when there are none, while having a hard time finding dangerous voltage levels on a live cable. And I explained why you should never use NCV for voltage testing, but you can use it to identify circuits.

I thought my initial comment and the video description would give that away well enough. And I'm not a professional YouTuber, I only made a clumsy demonstration on clear deficiencies to point out why it's not permitted to use it for that purpose (at least in Europe).

I suggest you watch this video instead.  He clearly demonstrates in very simple terms all the problems and just how dangerous  the meter is. 


After he release that video, I started having people ask me about running it.  There are three parts to it.     

Thank you, I should have a closer look at your older videos!

If I would send you over some BENNING duspol digital, would you be willing to test it the same way? Not that I distrust the VDE testing performed on it, but I would love to see it put to it's limits.

[joeqsmith]

VDE and other agencies will typically test the products for safety (61010).   Not being an electrician,  I have little interest in that and the tests I run have nothing really to do with safety.    I am closer to the EMC standards (61326) but even then I have my own ideas on how I want to test them. 

For a good meter, it can take me about a week to fully run one.  The switch life cycle test alone is a few days which ties up my equipment.  This is why you won't see me running a lot of meters.   Looking at cheap meters from the off brands is hardly worth the effort.  I also tend to look at brand new products.   We ran into a snag where one meter was damaged prematurely which caused a bit of head scratching.   Later bought a brand new one and repeated the test.  I also then repeated the test on the old meter after repairs and it too did very well.   So, lesson learned.     

I really like the Fluke 189 and of course, they are no longer made.   I was interested in seeing how one would hold up and recently bought a really beat up one just to test.  It held up really well.  I also looked at a trash picked Fluke 77 which did poorly.  However,  the data for these two meters is not documented in the spreadsheet.   

I think I could get that non BLE version of the Benning but based on the feedback you have received (or lack of it), I doubt many people would be interested.  I do get a fair number of people wanting to see some SANWA products ran.   Maybe. 

[tggzzz]

I'm not going to spend 15 minutes of my remaining life watching to see if there is something I don't know.

Is there any chance of you indicating the content via a few "bullet points" indicating what I need to understand? If they pique my interest, I will watch.

Dave does that with his videos, to great effect. Bonus: words are found by search engines, speech in videos isn't.

I demonstrated how NCV gives mostly wrong readings detecting dangerous voltage levels when there are none, while having a hard time finding dangerous voltage levels on a live cable. And I explained why you should never use NCV for voltage testing, but you can use it to identify circuits.

I thought my initial comment and the video description would give that away well enough. And I'm not a professional YouTuber, I only made a clumsy demonstration on clear deficiencies to point out why it's not permitted to use it for that purpose (at least in Europe).

Thank you.

Your initial comment said "problems". Your elaboration said "problems X, Y, Z". Big difference.

[rsjsouza]

I also use a NCV more or less the same way as Fungus mentioned: go/no-go immediately before and after cutting power to the circuit, either via a breaker or a light switch. I also have a simple dual pole voltage tester to take with me, especially to weird places where access is limited as it is quite portable. However, it also does not give information about if a wire is live or not if you simply plug them between neutral and live - I have seen and been exposed to weird installations where a live at 240 or 380V was properly cut off by its breaker but, due to imbalances on a crap installation, the neutral was in a weird high voltage of around 100V. The tester was happily showing zero volts between live and neutral, but a test between live and ground and neutral and ground is what saved my bacon.

Also, for the regular outlet or lighting fixture around the house, I take my old and trusty neon screwdriver. Yes, people freak out about it but, as long as it is properly inspected and well built, it reliably shows if a wire will in fact shock you.

[Cymaphore]

Since I gave an overly long explaination what I think should not be used as a safety-related tester...

Here my list of things about what I think a good two-pole voltage tester should include:

• Fully automatic operation, no interaction of the user must be able to suppress or disable HV alert output
• Basic operation with empty and/or removed batteries
• Redundant indication by LEDs or similar
• Single-pole voltage testing (capacitive one-probe phase testing), in case PE and/or N are broken
• Continuity with acustic output
• Switchable load to test for phantom readings
• Strong thick cables, probes not removeable and not pluggable
• CAT IV 600V or better, compliance with the latest relevant standards, based on area of use (for example EN61243-3)
• Tested and approved by a reputable institute (UL, VDE, etc.)

Useful additional features:

• Support for 1200V DC
• Rotating field detection / Phase sequence
• NCV to help in identification of unlabeled cables

Variety of Devices that more or less match these criteria (not including outdated products, just a selection):

• BENNING duspol digital and duspol expert (also sold as Gossen Metrawatt duspol)
• Beha-Amprobe 2100-BETA and 2100-GAMMA
• Fluke T90, T110, T130
• HT Instruments HT10
• Metrel MD1160