A friend from my old hackerspace was in grad school for electrical engineering. He had a professor who would ask, when something went wrong with a student project, “Have you fully characterized the system?” It’s a good, if lofty, goal, but it also became an inside joke around the hackerspace because YOLO was our MO about 95% of the time. Head crashes on the 3D printer – “not fully characterized”. Forgot to take out the trash last weekend? Was the system fully characterized?
It’s maybe also the difference between theory and practice: In theory, there’s no difference between theory and practice, and all systems can be fully characterized. But in practice, it’s hard to fully characterize a system that you don’t yet fully understand.
Case in point: we have nine small saplings growing in our front yard, and I have to water them. It’s boring moving the hose from tree to tree, so I thought I’d take a length of hose, stopper it at one end, and drill enough holes in it so that it could irrigate all of the trees at once. I kinda characterized the system: I figured out how much water flows per minute through our hose, and divided that up into a reasonable outflow in my mind, and drilled holes that ended up being way too large.
Why? Because a length of hose has a resistance to flow, and the water came pouring out of the first few holes, while the last few were dry. It wasn’t a constant pressure system like I thought it would be. I hadn’t even thought that the drag in the hose would matter, so there was no way I would have tried to measure it. But how would I characterize this resistance anyway? You could make a hose with too-large holes and measure the falloff. (Oops, that’s exactly what I did.)
In retrospect, professional drip irrigation systems always have holes that are tiny relative to the pipe diameter, which avoids this pressure-drop phenomenon, which means that they don’t have to worry about characterizing the hose resistance. So that’s what I ended up doing. I cut the hole size in half, and later widened up some of the downstream holes until it looked about right. Not even close to fully characterized, but it works.
So now, in addition to the engineer’s “have you fully characterized the system?”, I have the hacker’s “can you avoid characterizing parts of the system?” in my mind. And a holey chunk of hose in the trashcan.
Supercon News
Just briefly, in case you missed it: Tickets are on sale now for Supercon Ten, and we’ve extended the call for participation by another two weeks. If you’re a Hackaday fan, you owe it to yourself to join us at our annual gathering.
” I cut the hole size in half”
you have got to teach me that!
It’s like cutting it longer. A very useful skill to have.
Next most useful is to have more unused hose in the basement.
hose story reminds me of a system that was fairly well characterized but i didn’t understand it
hydronic heating system…i knew the Gallons Per Minute i desired, and i had a table showing the resistance (head?) for so many feet of pipe and so many count of elbows and tees and for each radiator at that GPM rate. and i had a good idea already the pump i wanted to use (everyone starts with the taco 007, i guess?). and the pump had a datasheet showing that at the resistance i had summed up, the pump would run at an asymptotically infinite-looking GPM (the curve cut off at like 10x my desired rate, and i was beyond that). i knew that wouldn’t happen but i shrugged and clicked buy instead of continuing my analysis.
but the obvious reality i eventually came to understand is that as i moved along the curve on the pump’s datasheet, the GPMs would increase…and that meant i would move along the curve on all of the pipe/fitting/radiator datasheets, and their resistance would also increase. and they’d find a happy medium in fact only like 1.5x my desired GPM. the two were related, and the relation was well-characterized. i just didn’t bother to look at it.
luckily in the parts of that system design that were more touchy, i was more rigorous, and it works well in practice :)
I had a similar experience with hydronics. Iterative load line analysis for the win, in the end.
Residential, the 007 pump is nearly universal, as it covers the majority of standard heating loops (about 50MBtu/hr or so, figuring 5GPM and 20F drop, at constant operation, and down to about 3 to 5MBtu/hr with reasonable cycles at the low end). A 003 is useful for a few things (smaller radiant floor, one or three 5 to 10MBtu kick heaters, a single bathroom with heated towel warmers, and the like), but anything larger than 007 is really in the commercial, or more dosh than I deal with specialty, space.
Whoops… I’m old…. MBtu/hr is thousand BTU/hr, or just under 300 Whr, for those that don’t speak outdated units.
Why use BTUs instead of kW?
‘Merica… The UK mostly gave up British Thermal Units long ago, but still used here. Heating/cooling is one of the few subfields I work in where I default to non-SI, as it’s how the equipment is specified and characterized. Lets not get into other field specific energy units used here like the ton, or the multitude of horsepowers (really, there are an Imperial butt-load of horsepowers. The wikipedia article doesn’t do justice, nor does is add clarity, to the merriment involved when dealing with a `horsepower’ specification that crosses industries)
Because 1 BTU increases the temperature of one pound of water (which is also also 1 pint of water) by one degree Farenheit. It is a convenience unit for HVAC calculations and it is the one you find on equipment in the US.
Similarly a “ton” of air conditioning is the cooling power of melting 2000 lbs of ice over 24 hours. Back in the ice mining days that was a more convenient measure of cooling power than kW. Now it is what US HVAC equipment makers uses to indicate nominal cooling power for rough comparison purposes.
Take the Feynman approach and explain the system to others.
is that the ai generated rambler on youtube?
You alphas and your 67 you don’t even know who Richard Feynman is
What is this world even coming to
What no 3D printed esp-32 controlled dynamically-sized nozzles?
Right? It’s Hackaday, overengineering is part of the appeal.