Unit conventions (SI versus others)

[Dr.D]

As missing a prefix can send your calculations off by hundreds of millions of years!

That right there points to one of my principle objections to SI as it is commonly (improperly) used. Proper SI says all lengths are in meters, and that works just fine. Right now, I'm dealing with a problem concerning machine parts with all dimensions in millimeters (yes, I can multiply by 10^3). The difficulties come when calculating mass moments of inertia, with density in kg/m^3, linear dimensions in mm, so suddenly there is a factor 10^15 that drops into everything. What a mess! If it was all dimensioned in inches with weights in pounds, then the MMOI calc would give lb-s^2-in in the usual way, with no massive fudge factors required. But its so much better, or so they tell me ...

 

[OmCheeto]

While doing a search on the Delisle scale, the Wikipedia information has this conversion chart I found interesting. Here's a scaled down image of it I made.

View attachment 228147

Doh! I guess I read through that section too fast. From that chart, the Delisle scale is STILL BACKWARDS!

The Celsius scale, like the Delisle scale, originally ran from zero for boiling water down to 100 for freezing water. This was reversed to its modern order after his death, in part at the instigation of Swedish botanist Carl Linnaeus and the manufacturer of Linnaeus thermometers, Daniel Ekström.​

 

[russ_watters]

That right there points to one of my principle objections to SI as it is commonly (improperly) used. Proper SI says all lengths are in meters, and that works just fine. Right now, I'm dealing with a problem concerning machine parts with all dimensions in millimeters (yes, I can multiply by 10^3). The difficulties come when calculating mass moments of inertia, with density in kg/m^3, linear dimensions in mm, so suddenly there is a factor 10^15 that drops into everything. What a mess!

This is nonsense. If all problems 'should' be done in the base units, there would be no reason for the prefixes to even exist. It's as if you are trying to prove your car is better than ours by beating ours with a baseball bat before comparing them. It says more about the chip on your shoulder than about the ease/difficulty of using the units.

If it was all dimensioned in inches with weights in pounds, then the MMOI calc would give lb-s^2-in in the usual way, with no massive fudge factors required.

Are you calling the prefix a fudge factor? Is 12 inches per foot a fudge factor? What if you did the problem in in^3 and you find the catalog for the part you need is in ft^3. Can you do that "fudge" in your head?

Caveat: Every industry has conventions for unit choices and rolled-up constants for common calculations. It may seem easier once you have them memorized, but it is only because they are memorized. A couple of days ago I had to dissect the rolled up constant 6356 for showing my work in a report. It took me an absurd amount of time.

But its so much better, or so they tell me ...

If you drive your car as actually recommended instead of driving it on the highway at 60mph in first gear, it will definitely work much better.

 

[russ_watters]

A bit more on that last example:
cfm x in wc / 6356 = horsepower

I literally used this equation for years thinking 6356 included a constant of nature* (like 1.082 does...) before I realized it was entirely comprised of IP unit conversions. And so I didn't realize the equation I was using was simply airflow x pressure = power. No conversions inherrent!...

...until I did a project in SI and found that m^3/s x Pa = Watts

It was a revelation! Physics makes sense as applied to engineering!

Two minor caveats:
-You of course need to know the named units. But it is as if ft-lb=horsepower, but without the conversion factor. Still easier.

-Can't avoid conversions with time: Often the flow rate is expressed per hour. But again, that issue is worse in IP. I once showed my boss a calculation with ton-hours per year as the unit and he kicked me out of his office! (it has three different units of time in it).

*An not for nothing, but SI is also designed to eliminate conversion factors with some constants of nature as well.

 

[jim hardy]

As a fellow who still keeps Whitworth wrenches
and uses Imperial unit steam tables
and considers himself practical
I agree with Dr D - we should be fluent in both systems.

i appreciate some of the simplicities of the Metric system
but i curse my Oldsmobile with its fasteners that are SAE on the drive train and Metric on the body.
The eye can pretty easily discern the 1/16 inch increments in SAE bolt heads , but a mm being only ~2/3 of that fools me most every time.

We EE's have it easy because our units are already SI (Amps Volts Ohms Farads Henries).
but I've always pitied ME's because their Ohm's Law involves local gravity, Reynolds Number, viscosity, density, a square root, and a discharge coefficient.

So i think you guys ought to ease up on Dr D . His point is simply that we should be multilingual if we wish to consider ourselves well rounded.

When nobody can read cursive anymore how will we proof-read printed versions of the Bill of Rights ?

Magnetics is a place you really need to be fluent - there's still Oersteds and Gilberts and Maxwells floating around. Every specialty develops its own shorthand with implicit 'fudge factors' .