We’ve had a few posts recently about blown up gear due to power problems. Here is a basic metering procedure that should give you a good indication of the safety of a standard North American receptacle (NEMA 5-15R or 5-20R).

Standard disclaimers: 

  • Performing this test can give you reasonable assurance that a receptacle is properly wired and functioning, however it does not guarantee that the receptacle is good – there are still possible fault conditions that may not be revealed with this test.
  • Performing this test requires sticking conductive things – multi-meter probes – into an unknown electrical supply. This has the potential to create an electric shock hazard if you touch the metal tips of your meter probes (duh), and if things are really wrong, you may cause your multi-meter (especially if it’s a cheap one) to explode.

Proceed at your own risk (though testing is a lot less risky than plugging gear into an unknown receptacle).

Anatomy of a Receptacle

An electrical circuit is a loop. Power comes out of something, goes through the device you’ve plugged in, and then has to return to the source. If this loop is broken, no current will flow. In the NEMA 5-15 series receptacle, power comes out of the hot wire, flows through your device, and returns to the neutral conductor. The hot terminal is shorter than the neutral terminal. This keying is intentional to make sure that the device can’t be plugged in backwards.

The third terminal is the Ground terminal. This is a safety path back to the service panel that may conduct away fault currents in the event of a device fault.

What’s Supposed to Happen

When things are wired correctly and working normally, approximately 120 volts and up to either 15 or 20 amps of current, depending on the circuit’s capacity, are available for doing work. This energy flows out of the hot wire, through the connected device (the load), and returns to the neutral wire. As such, the hot and neutral wires are considered current carrying conductors.

The ground wire is not intentionally a current carrying conductor – if more than a small amount of current is flowing on the ground wire, that indicates a fault.

Testing Procedure

Set your multi-meter to AC Volts. Be sure that your multi-meter is capable of measuring line voltages. It is strongly recommended to get a meter that is rated Category III at 600 volts or better.

DO NOT USE THIS RECEPTACLE!

  1. Insert one probe in the hot terminal, and the other in the neutral terminal. It doesn’t matter which one goes in which terminal for the AC measurements we will be doing. You should read between 110 and 125 volts AC. If you read higher than this range, DO NOT USE THIS RECEPTACLE. If you read lower than 110 volts, that indicates the circuit feeding this receptacle is heavily loaded, or there is a lot of wire between the service panel and where you are metering. It may be possible to use this receptacle, but you may encounter brown-out conditions which can also be hard on gear.
  2. Measure between the hot terminal and the ground terminal. You should read between 110 and 125 volts AC. This reading should be roughly the same as your previous reading between hot and neutral. If you read outside this range, DO NOT USE THIS RECEPTACLE.
  3. Measure between the neutral and ground terminal. You should read between 0 and 5 volts AC. If you read a bit more than 5 volts, the circuit may still be safe to use, but that indicates either the circuit supplying this receptacle is heavily loaded, or a long way from the service panel. If you read substantially more than 5 volts, DO NOT USE THIS RECEPTACLE.
  4. Plug a high-current but inexpensive device such as a 500 watt PAR can or heater into the receptacle, and repeat all 3 measurements. What you should see is the voltage between hot and neutral should drop slightly – a volt or two, the voltage between hot and ground should drop similarly, and the voltage between neutral and ground should increase. If any of these values deviate more than a volt or two from your no-load testing, the receptacle is pretty soft. If the voltage between neutral and ground doesn’t increase when tested under load, that likely indicates a “cheater” receptacle, where the neutral and ground terminals are connected together inside the receptacle box, and the receptacle should not be used.

Common Faults

  • Mis-wired receptacles are incredibly common – even in new, nice looking commercial construction. If you read values other than what is listed above, the receptacle is likely mis-wired: either a hot-neutral swap, or a hot-ground swap. Note that it is not possible to test for a neutral-ground swap using a multi-meter.
  • Loose wiring. Since a multi-meter draws almost zero current when testing, it is possible for a receptacle to test normally, but under load, the loose connection opens up. Depending on which wire is loose and where the fault lies, the voltage may either go down below 120 volts, or it may even go up to 208 or 240 volts! This is why it is important to test your receptacles under load. Putting a heavy load on a circuit like a big light bulb or heater causes the receptacle and supply circuit to reveal their true colors.
  • Wrong voltage supply. Virtually every building has multiple voltages in use – 120v, 208v, 240v, 277v, 480v, 600v, and others. It is possible to have a receptacle fed by a supply other than 120 volts – incredibly these dangerous situations can stay around for months or years.

Test, test, test! Get in the habit of unloading your multi-meter before you unload your truck. If the power is suspect, pack up and hit the road. “The show must go on” is a stupid motto. Expensive gear can be damaged, and people can be hurt. It’s just not worth it.

22 COMMENTS

  1. [SIZE=2][QUOTE=Jay Barracato;bt759]Oh yeah- If the circuit has a GFCI on it, measuring hot to ground should trip the GFCI breaker.[/QUOTE]It actually shouldn’t. Most typical meters have an input impedance of several MΩ. I looked up a Fluke 117 – a good choice for our industry, and the AC volts impedance is listed at greater than 5MΩ. At 120 volts, that’s .024mA – far less than the 3mA that is allowed for a GFCI.

    Some meters have a “Low-Z” function where the input impedance is much less – like 3KΩ. These ranges are useful for loading down ghost voltages, and this mode will indeed trip a GFCI (current at 3KΩ is 40mA), but this mode is special purpose and not found on most multi-meters.
    [/SIZE]

  2. Thank you very much, TJ. This is something I’ve needed to know for a very long time and appreciate the effort. I suspect that I’ll find very interesting results when testing some of the places I think might have power issues (i.e. my school’s outlets…. wired from back in 1970 or so with a few hundred computers on the same service as our auditorium).

  3. [QUOTE=Bennett Prescott;bt762]Hey TJ, my Fluke 117 has a “lo-Z” AC voltage autoranging measurement mode which they list as about 3KΩ.

    [URL=”http://www.fluke.com/fluke/usen/Digital-Multimeters/Fluke-117.htm?PID=55996″]Electrical Multimeter | Fluke 117 Electrician[/URL][/QUOTE]

    Yes it does, as well as the normal Hi-Z AC volts range one click to the right of the “Off” position.

    The Fluke 117 is designed as an industrial meter, where sometimes the “finesse” of a regular high-impedance meter may give wrong readings – induced voltages from very long wires can be picked up by a high-impedance meter, however in many cases that is the wrong answer as there is no current behind that voltage – hence the term ghost voltage. A low-impedance mode draws enough current to minimize these ghost voltages.

    Note that the Low-Z situation is the minority of measurement scenarios, and until fairly recently this function wasn’t found on too many meters. Since the very beginning of digital volt meters, one of the most prized benefits is the higher-impedance characteristic of digital meters compared to older analog meters. In electronic situations, rather than industrial or power distribution ones, the less power siphoned away by the meter means the meter won’t interfere [much] with the lower power electronic circuit, and will give a more accurate reading by not influencing the circuit under test.

    It’s horses for courses. If you’re trying to eliminate ghost voltages, the Low-Z mode is useful. For what we are trying to do, it’s the wrong tool. Low-Z mode WILL trip GFCIs. Since Low-Z mode only draws maybe 40 – 50 mA of current, this really isn’t enough to get a good idea of the solidity of the circuit; since we intend to use the circuit to nearly full capacity, we need 5 or 10 amps of load at least to get a good idea if the circuit will hold, and if testing a 400A company switch, ideally quite a bit more than that.

    The Fluke 117 is really a great meter though, and is my top recommendation for what we need to do. It’s rugged, fairly small, will last forever, and is purpose-built for line voltage measurement, and has appropriate safety features. A lot of the cheapo $20 – $75 meters on the market are really shoddy, and can actually be dangerous, since they lack the safety features of the better meters.

    Dave Jones of the EEV Blog (eevblog.com) has done a number of great videos on multi-meters, and why you don’t want a cheap one. Here are a couple:
    [URL=”http://www.eevblog.com/2010/06/15/eevblog-94-near-death-multimeter-experience/”]EEVblog #94 – Near Death Multimeter Experience[/URL]
    [URL=”http://www.eevblog.com/2010/05/05/eevblog-84-high-energy-multimeter-destruction/”]EEVblog #84 – High Energy Multimeter Destruction[/URL]
    as well as a review of the venerable Fluke 117:[URL=”http://www.eevblog.com/2010/02/08/eevblog-60-fluke-117-multimeter-review-and-teardown/”]
    EEVblog #60 – Fluke 117 Multimeter Review and Teardown[/URL]

  4. [QUOTE=Max Warasila;bt763]Thank you very much, TJ. This is something I’ve needed to know for a very long time and appreciate the effort. I suspect that I’ll find very interesting results when testing some of the places I think might have power issues (i.e. my school’s outlets…. wired from back in 1970 or so with a few hundred computers on the same service as our auditorium).[/QUOTE]I’m glad you’ve found it helpful.

    Existing wiring is a strange beast, and there really isn’t any correlation between building age (within reason), and wiring quality. Any wiring done in the 1970’s or more recently is very similar to modern methods and materials, and as long as it was installed well, is probably every bit as good as a new building. In fact, brand new buildings are arguably more dangerous than structures that have been around for a while, since you might be the first person to use a particular receptacle, and therefore the first to find a wiring error.

    There can be interaction between different loads in a power system – air conditioners causing brownouts, triac dimmers inducing noise, etc., but as long as the service is sized adequately and has over-current protection, these things shouldn’t affect the safety of a circuit.

    My article deals only with electrical safety, and not power quality or interference. With today’s pro gear, there really isn’t any reason to have trouble with noise in a system. If you are having trouble, that would indicate a piece of gear with a “pin-1 problem” – a device’s inability to reject noise coming into the chassis ground.

  5. [QUOTE=TJ Cornish;bt765]I’m glad you’ve found it helpful.

    Existing wiring is a strange beast, and there really isn’t any correlation between building age (within reason), and wiring quality. Any wiring done in the 1970’s or more recently is very similar to modern methods and materials, and as long as it was installed well, is probably every bit as good as a new building. In fact, brand new buildings are arguably more dangerous than structures that have been around for a while, since you might be the first person to use a particular receptacle, and therefore the first to find a wiring error.

    There can be interaction between different loads in a power system – air conditioners causing brownouts, triac dimmers inducing noise, etc., but as long as the service is sized adequately and has over-current protection, these things shouldn’t affect the safety of a circuit.

    My article deals only with electrical safety, and not power quality or interference. With today’s pro gear, there really isn’t any reason to have trouble with noise in a system. If you are having trouble, that would indicate a piece of gear with a “pin-1 problem” – a device’s inability to reject noise coming into the chassis ground.[/QUOTE]

    I just have a feeling that we get swing at least 5 volts both above and below 120VAC. I’m excited to be able to measure it.

    Of course, I’ll have to get an appropriate multimeter first… I wonder if I can scavenge the cash for a good one. Any cheaper suggestions in the meantime (I know, dangerous, but I thought I’d ask and see what I get back anyway)?

  6. [QUOTE=Max Warasila;bt766]I just have a feeling that we get swing at least 5 volts both above and below 120VAC. I’m excited to be able to measure it.

    Of course, I’ll have to get an appropriate multimeter first… I wonder if I can scavenge the cash for a good one. Any cheaper suggestions in the meantime (I know, dangerous, but I thought I’d ask and see what I get back anyway)?[/QUOTE]
    Any meter from a reputable company that has a Category III 600v rating or better should be OK. Flukes are the standard, but as we are not concerned about absolute measurement accuracy – whether it is 119.5 volts or 120.3 volts doesn’t really make any difference, we just need a safe meter that is easy to use for basic AC measurements.

    In the Fluke 110 series, I would avoid the 113 for sure – it requires too many button presses to get into the right mode. The 114 or higher are OK, though be aware of the Low-Z vs regular AC volts modes as mentioned before.

    Auto-ranging is nice (but be sure to know what range you’re in when interpreting answers), but I am not a huge fan of automatic AC/DC selection (one of the reasons I don’t like the Fluke 113). Backlit screens are really handy, and other ergonomic features like rubber edges, a good stand, and nice feeling probes are nice to have.

    I’ve seen meters at Home Depot badged as Milwaukee and Greenlee. These are likely to be safe meters (though check for the Category III or IV rating). Some of the ergonomics are “interesting” on a few of the models. Agilent meters are also high quality, but will likely be out of your price range.

    You may also want to look for a used Fluke. Meters generally last a long time; a 5 or 10 year old meter if well cared for is likely just as good as new.

  7. [QUOTE=Jay Barracato;bt772]Now you have me wondering if it is my meter (Extech) or the quality of the GFCI’s in my classroom (the GFCI’s I meter the most often).[/QUOTE]
    You should be able to meter your meter. Find a second meter and do a resistance mesurement of your first meter set to AC Volts. If I recall correctly, GFCIs [U]must[/U] trip at 5mA leakage current, but [U]must not[/U] trip at less than 3mA, to reduce nuisance tripping. Assuming my memory is correct, either your meter is Low-Z (which would make a lousy lab volt-meter), or your GFCI is defective.

  8. One more thing that’s VERY important to know when checking any outlet. Be on the lookout for something I call an RPBG (Reverse Polarity Bootleg Ground) outlet. It’s a combination of a bootleg ground AND reverse polarity. This sometimes occurs when an ungrounded outlet has been improperly “upgraded” to a grounded one in a older (pre 1970’s) building.

    The real interesting thing about an RPBG outlet is that while it measures as perfectly OK using a standard 3-light tester or even a voltmeter (on a branch extension it can behave just like a correctly wired outlet with more than 1-volt G-N), the ground and neutral are sitting at 120-volts while the hot is sitting at 0-volts. So any gear plugged into a GFCI mis-wired outlet will appear to operate normally, but its chassis is now at 120-volts above earth ground.

    The very dangerous part is that the chassis of the gear plugged into a GFCI has a 20-amp 120-volt low-impedance connection, and if you touch it and a ground at the same time you can die. Also, if you interconnect it with a piece of gear plugged into a properly grounded outlet, the full circuit breaker capacity of the branch circuit will flow between the gear through the XLR or USB cables. That’s a serous meltdown.

    [IMG]http://howtosound.com/images/RPBGSokolFault.JPG[/IMG]

    The best way to find an RPBG outlet is to use a standard Non Contact Voltage Tester (NCVT) to see if the ground contact in the outlet is hot. See my article on RPBG outlets in EC&M Magazine at [URL=”http://ecmweb.com/contractor/failures-outlet-testing-exposed”]Failures in Outlet Testing Exposed | Contractor content from Electrical Construction & Maintenance (EC&M) Magazine[/URL].

  9. Thanks for your comments, Mike. The load test in the procedure I outlined should reveal the reverse polarity bootleg ground condition you describe, as measuring between neutral and ground – since they are jumpered in that scenario – will read 0 volts under load. That condition fails the outlet, whether it also has a hot – neutral swap or not.

  10. In the UK the numbers are L>N 220-250V, L>E same as L>N, E>N less than 15V everything else re testing with a load and meter impedences applies just the same. One major difference is our standard domestic plug has a fuse built in this is anything up to 13A rating it’s worth making sure that the fuse in the appliance plug is of the correct rating G

  11. Over the decades, nominal US line voltage as been trending upward. Post WW II it was 110V then 115V, later 117/118V then later it was 120V now it’s drifting towards 125V. In the last few NEC books it’s been 125V and most UL tested units are marked 125V. In my very old city neighborhood I sometimes see 124V and most of the time 123V.

    So you need to re-think calling 125V the maximum.

  12. [QUOTE=Kevin Graf;bt812]Over the decades, nominal US line voltage as been trending upward. Post WW II it was 110V then 115V, later 117/118V then later it was 120V now it’s drifting towards 125V. In the last few NEC books it’s been 125V and most UL tested units are marked 125V. In my very old city neighborhood I sometimes see 124V and most of the time 123V.

    So you need to re-think calling 125V the maximum.[/QUOTE]
    As far as I can tell, line voltage was standardized by ANSI C84 in 1954 to 120v +/- 5%, which translates to 114v – 126v (commonly rounded to 125v). I am not aware of any changes to this standard since then (though surely it took a while for distribution to implement this), and haven’t found any documentation indicating measured general trends towards voltages higher than 120v nominal. Any recent trends would likely be to more closely approach the nominal 120v standard with the rise of better utility voltage regulation in the last couple decades (which could be an increase if previously that location tended towards under-voltage).

    The NEC refers to 125v/250v devices – not because of a trend towards that as the nominal voltage, but to indicate the high-side of the tolerance band as the rating for that device.

    A number of devices such as induction motors have other voltages such as 115v or 117v stamped on the nameplate, but this number was designed to indicate the voltage at the device through branch circuit wiring, allowing for a little voltage drop from the nominal 120V at the supply. This doesn’t necessarily imply that there ever was a formal standard (post-war anyway) of 117v. This rating was meaningful, as an under-voltage condition is hard on induction motors; when the supply voltage is low, the motor’s current draw [U]increases[/U] – generating heat, unlike a simple resistive load where current decreases when voltage is reduced.

    Your “very old” neighborhood with a 123v-124v supply is not particularly indicative of a new trend, since it’s both very old, and still well within tolerance.

    Measuring 126 or 127 volts doesn’t necessarily indicate a problem, so if it makes you feel better you can likely consider that a safe voltage; but an over-voltage condition – especially on the load test – is a strong indication of a high-impedance neutral conductor somwehere, so be careful.

  13. “”Measuring 126 or 127 volts doesn’t necessarily indicate a problem, so if it makes you feel better you can likely consider that a safe voltage;”

    You might want to add to measure the other leg at the same time, if its 113, 114 you have a problem. Never think 126, 127 is safe. There are issues.

  14. [QUOTE=Alan Sledzieski;bt834]” power comes out of the hot wire, flows through your device, and returns to the neutral conductor. ”

    Interesting, you do know what AC stands for?[/QUOTE]Since the neutral conductor is (supposed to be) bonded to ground – effectively anchoring it at the same potential as you and me, yes – power flows out of the hot wire no matter what its instantaneous phase rotation is.

    If you’d like, you can pretend to be a super hero and claim that you are shocking the receptacle 50% of the time instead of the other away around, but in a shocking fight, normally the electrical equipment wins and the human loses.

  15. [QUOTE=Jay Barracato;bt772]Now you have me wondering if it is my meter (Extech) or the quality of the GFCI’s in my classroom (the GFCI’s I meter the most often).[/QUOTE]

    I have recently posted about having had two oddball electrical situations (metering GFCI’s popping the circuit, and older wall warts working one way but not when inverted) in my classroom. I have decided whatever is happening it is most likely due to having multiple cheap GFCI outlets on the same circuit in my classroom. Each 20 amp circuit may have as many as 10 GFCI outlets.

  16. I haven’t heard of line-side issues causing trouble with GFCIs, but I suppose it’s possible. Maybe the more important part is the [U]cheap[/U] GFCI outlets…

    My Bridgeport milling machine with VFD that is stationed near my service panels will trip one particular arc-fault breaker supplying lighting to the floor above my shop, but only if I have lights on up there. I’ve been meaning to order the line filter for the VFD, but haven’t gotten around to it.

  17. [QUOTE=Jay Barracato;bt856]I have recently posted about having had two oddball electrical situations (metering GFCI’s popping the circuit, and older wall warts working one way but not when inverted) in my classroom. I have decided whatever is happening it is most likely due to having multiple cheap GFCI outlets on the same circuit in my classroom. Each 20 amp circuit may have as many as 10 GFCI outlets.[/QUOTE]

    one of my dubious gfci’s went up in smoke the other day. It wasn’t in use, just all of a sudden started pouring out smoke and then tripped. I am still waiting on the electrician, but it is unlikely they will know any more than simply replace the faulty unit.