Mr. George and His Units
by Paul McGoldrick
In my first year of college we had, as potential engineers, to study outside my chosen field of electronics. We had technical drawing, mechanical engineering, structural engineering, fluid engineering, and the heavyweight of the lot: Power engineering. All of those areas were actually of use in my career, although I never dreamed they would be at the time, but it was power engineering where I was most intimidated.
The knowledge gained was of considerable use in later life as I spent years hanging around high-voltage power conversion, filament generators, mercury-arc rectifiers, and blower equipment in RF transmitters, but the lecturer responsible for the course terrified me…
Mr. George was a tall man with a large beer belly and he was a natural haranguer. Work came back from him covered in red ink and the ultimate threat was sometimes at the top of a page, "See me!" It took me many months to realize that he never remembered who he had asked to see him and that his bark was decidedly louder than his bite. But, then, he would never have allowed me to write what I just wrote and get away with it. "What do you mean louder? Quantify!" OK, so maybe his bark was 40 dB louder than his bite.
He was all about definitions, limits and units. He would insist that a mathematical calculation have a parallel calculation of units so that when you ended up with a result you were able to justify it with the correct unit. And it worked; if you came up with a voltage, but with the unit in amperes, you could be pretty sure you screwed up at least once. And he was hard about the use of words: "What do you mean a lamp bulb? Bulbs grow in the garden!"
But amid all the exclamation points that got hurled at us it was the precise use of units that so much got embedded in me, and still does today.
I was in e-mail correspondence with a reader over a couple of weeks period about some of the units that get misused today, and in some cases they are so profligate that upcoming engineers believe them to be correct; I suspect that unlike my time as both a student and a college lecturer less care is taken by those that now teach.
Our whole system of units and abbreviations hinges directly from SI Units (Le Système International d'Unités) where they are divided into base units, derived units, and supplementary units. The base units are the quantities of length (meter - m), mass (kilogram - kg), time (second - s), electric current (ampere - A), thermodynamic temperature (kelvin - K), amount of substance (mole - mol), and luminous intensity (candela - cd.) Note that the unit names all have lower-case initial letters but where a unit is named after a person its symbol uses upper-case, as in A, K and derived units in electronics like the volt (V), farad (F) or Henry (H).
Additionally there are rules from the likes of the IEEE in the US and the IEE in the UK that suggest abbreviations to be used. Among the abused in this arena are things like PNP (should be pnp), N-channel (should be n-channel), AC (should be ac), DC (should be dc). I personally correct several hundred units a week in submitted manuscripts although we have a policy at analogZONE of not correcting vendors' press releases in any manner, units included.
But the worst abuses are things that the computer engineers have slipped into our vernacular. What does MSPS stand for? Mega-samples-per-second according to the abusers of it although the IEEE specifically advises that units like these should be spelled out. The SI/conforming engineer in me reads MSPS as Mega-Siemen-peta-Siemen (peta is 10^15) which is a really interesting unit and, of course, its companion KSPS is Kelvin-Siemen-peta-Siemen.
How about kbps? kilo-barn-pico-second? (barn is an "allowable" unit in SI for nuclear physics representing the tiny area of 10^-28 m²) KBPS? Kelvin-Bel-peta-Siemen?
You get the idea. But how do you express a number that isn't a thousand at all but, of course, is 1024?
In 1998 the International Electromechanical Commission (IEC) tried to stop these latter abuses by - and I am not making this up - introducing the kilobinary which when used for a number of bytes was to be abbreviated as a kibibyte and using Ki as the symbol. Larger units, the Mi and Gi, were more readily adopted but most engineers just look at the word kibibyte and think they're reading the brand name of some dry cat food.
So, just spell it out instead of abusing it. And if this comes back to
me with red ink from Mr. George on it, I hope it says, "Glad you listened!"