Some after-thoughts...
Always I come up with a better answer after I study on it for a while....
I use a 4000 count Fluke for field troubleshooting, on instrumentation and computers, where 10 micro-volts IS important. Industrial scales are 0-20mV full scale. I once built a scale with 750,000 (3/4 million) pound capacity. The resolution was 10 pounds. I could tell if someone was out walking on the weigh bridge.....
The comment about
general use is spot on, +/- 5% is usually more than sufficient.
My bench meter for calibration is an American Reliance 4-1/2 digit. Both it and the Fluke are "True RMS" meters. RMS is important if you are working with "non-symmetrical" waveforms. For D-C and 50/60 Hz line frequency, standard calibration is quite acceptable.
For the hobbiest, over-range tolerance and internal fusing is important. When the LCD display blows up in your face, the glass shards can draw blood. Not if, but when. It
will happen, eventually. The more careful you are, the longer it takes. But it will happen.
High current capacity isn't really important; If your circuit is pulling 7-1/2 amps, light meter leads can get you burned. Use a shunt resistor, say 0.1 ohm and read the milli-volts across it.
Light-weight "pocket" meters are very handy, especially in the field, but I really don't trust them on anything over about 24 volts. Also, look for the cushioned case, with moisture seals. Conductive dust will zap you just as quick as a wet meter. And, you never know when. At least if the meter is wet you know to back out and dry it out.
For power circuits, especially dealing with motors and large inductive loads, an analog meter is more useful. And safer. I keep a Simpson 260, and a vintage Hickock for working on motor controls. Also an (very) old Micronta. From Radio Shack, back when they were still a division of Allied Radio. (before Tandy) But, it stays on the bench. The motion of the needle is useful working with large caps, as well.
Know well that a digital, especially newer models, WILL lie to you on inductive circuits. I have seen it both ways; telling me power was there when it wasn't. Or more important, the full line voltage WAS'T there when it was. At 480 volts, and high horsepower, that is definately not acceptable.
A digital will also tell you a glass fuse is good when it isn't. Use a test light, and NOT one of those neon lamp gizmos. Those things are just plain dangerous.
It has to do with input impedence. The digitals have input impedences in the multi-meg ohm region. Wonderful to keep down loading on sensitive circuits. But, there are places where the 20Kohm per volt loading makes a difference to the better.
And, digitals have quick resolution of RMS, if that is an issue. True RMS analog meters are pricey and slow to stabilize. And sensitive to temperature, as they normally use the "Joule Heating Effect" on a resistor to drive a milli-volt thermocouple.
Since I work with such a wide variety; heavy electrical, analog instrumentation, and computers (analog computers DO still exist), for tool-box use I have a Simpson(260/5), the Fluke(73), and a Square D "Wiggington" industrial voltage tester. The Ideal brand "Wiggy" is nicer and has a fuse tester built in. BUT, I've taken 480 phase to phase across the chest because their leads are not as well made as the Sq.D. (That zap is probably why my mind is so far out in left field
)
The old man that taught me to troubleshoot at my first mill job carried 2 of the 220 volt shop lamps in series with 3 leads. DC cranes operate from 260-280 volts. That's a little
heavy handed for the hobbiest but it was what worked for the situation.
Bottom line, there is no single "best" meter. Look at quality more than price. Sometimes it's hard to tell. And unusual conditions will ruin the best of the best. Airline baggage holds are not usually pressurized. When I traveled for Wang, I had to replace my Fluke two-three times a year. The low pressure ruined the LCD display.
This probably caused more confusion than help. But, I do hope it helps.
Bill