"If you cannot fix it with a multimeter,
The modern engineer is somewhat fascinating. Whenever he opens his bag the first instrument he pulls out is the fanciest digital multimeter the world produces - yet 9-tenths of the features he will never use. Ok, ok! 8-tenths! But that was said that under duress!
Have a look at the older engineer and the first instrument is more than likely a good ol' fashioned AVO or similar moving coil meter. But why? Is it because he has not kept up with modern technology? Does he fear advances in digital techniques? Well, there are a few like this but the main reason is the moving coil meter can tell a lot more to the trained individual than many modern fancy digital types. Personally speaking, the old "analogue" meter is held on to for dear life. And no, not another soul is allowed to use it!
Modern digital multimeters (DMMs) and digital test instruments do have their place but they have some inherent weaknesses that moving coils don't. DMMs take time for the reading to reach the value being read. This stops any possible flicker being noticed. Some even have "stabilising" and/or averaging features built in to stop sudden changes from upsetting the reading. There are more advanced instruments that deliberately filter the incoming signal to stop spikes and transients from giving "incorrect" values. To the electrician, this is perfect. To the power quality investigator, this is a serious flaw!
DMMs are generally designed to sample the incoming signal at the mains frequency of 50 or 60Hz. The problem is if the sampling done by the meter is fractionally different in frequency to the mains, which is highly possible, then the reading can "wobble" at a rate which is the difference between the sampling and incoming frequency. The problem is further extended when meters designed for one frequency are used on another. The travelling power quality expert needs to be specifically made aware of this. You cannot buy a DMM in the USA and expect it to be accurate in Europe.
There are DMMs that "lock" themselves onto the incoming frequency (or sub-harmonic of the incoming frequency) with the aid of a phase locked loop. Although readings will not "wobble" they are even more prone to missing a spike, transient, or flicker (and the serious flaw referred to earlier even more pronounced by the absence of decent readings).
There is a world renowned make of DMMs. Although they have made other superb hand-held test instruments, their DMMs are the most frustrating to use when locating a fault as the readings will almost always be incorrect. Don't ask me to prove it. I've done that so many times I'm now nauseated by it, in fact, so much so I've even coined a phrase which states "the fact your meter actually indicates what is presented on the leads is sheer ....."
One redeeming feature of the older analogue meter is it places a higher load on the circuit being tested and therefore gives a clearer indication that a circuit has in fact been switched off before commencing with work on it. DMMs have a habit of showing a circuit to be live when the circuit has no load on it and is fed in a way where Live has a reasonable level of capacitive coupling to other Lives.
Ok, so we've slated the poor DMM. They do, however, have one redeeming feature. They tend to be a lot more forgiving when accidentally connected to a voltage for which they are not adjusted (which is why the analogue meter is not entrusted to anyone else). There is a rule which has personally worked well. The cheaper the DMM, the better for use with finding problems. Anything other than a perfect, stable sinewave tends to shoot the reading in all directions - a quick indication there is a serious problem requiring a deeper look.
DMMs are also very good when trying to find out whether a panel or overhead line is alive or not, without actually having to connect to the panel or line. Having a high input impedance means it only takes very little capacitive coupling to get a reading.
Want to know if an overhead line is live? Take a prong in each hand holding them on the plastic body with your finger tips (don't touch the metal, you'll short out the signal). Stand below the line and spread your arms right-angles to the direction, but in the same plane, as the line. The voltage read on the meter will depend on the voltage on the line, the distance you are below it, and the impedance of the meter, but well over 100V can be realized.
Want to know if a cable is live without having to cut through it? Simply hold on to the one prong (the metal this time) and wrap the other lead around the cable in question. If live, there is likely to be a voltage (if you cannot wrap the lead around the cable, some large crocodile clips with a 'flat' construction will offer a reasonable coupling to the cable).
Two warnings about the above examples. This only works on AC therefore DC cables cannot be tested like this. This type of testing cannot be used when safety is an issue. Very cheap DMMs tend to have very low impedance inputs and, by not reading a voltage, could give the user a false sense of security thinking the line is dead and can be worked on.
But for the person wanting to ascertain it is a general power failure and not just a pole mounted breaker that has tripped, being able to 'test' whether an overhead line is live or not could go a long way to saving time!
I will never forget when doing my apprenticeship in two-way radio and coupling a transceiver to what appeared to be a solid 24V supply when installing it in a truck. Everything appeared fine... until the transmit button was pressed. Suddenly the indicator relay started ticking!
The classic mistake was we never loaded the wire as part of the test to see if it was sufficient for the job we had intended for it. Worse still, we simply relied on a multimeter to tell us where to find, in this case, 24V DC. The same happens when testing 240V circuits. Testing for the 'solidity' of a test voltage should go well beyond the loading of a multimeter, especially a digital whose input impedance can be many megohms.
If a measurement is being called into question, then load the point being measured with a suitable lamp, or at least a resistor that will dissapate a watt or two when connected.