Gears to grind

My roommate and I share a love for coffee. Yes, caffeine, that ubiquitous and societally accepted drug, is near and dear to both of us, and like all good addictions, the fix comes with a ritual. For the making of a good cuppa, freshness is key all throughout the process, and the closer you get to your mouth, the more it matters. No one wants, for example, to drink day(s)-old coffee. For starters, you want something freshly brewed, and to go one step up from there, you want beans that are freshly ground. (Ideally roasted recently as well–I’ve roasted green coffee beans at home on a few occasions before, but that’s a topic for another day.) Of course, if you really want to preserve all the delicious aromatic molecules that confer subtle flavor and scrumptious scent upon the coffee bean, you want to use a grinder that doesn’t abuse (read: heat) the beans too much. Yes, the friction created by those “blade” grinders amounts to high speed bashing and degradation of those banana tones (hey, maybe there are banana tones in there if your coffee’s fruity?), so real coffee gourmands will tell you to go hard or go home. As in slowly crush your coffee beans against a hard (ceramic) surface into the grounds with which you brew. This is what burr grinders do. I’ve got a manual one that works great! It’s made in Japan and does wonders for your grounds. It just takes a cool minute (or two) of cranking…by hand…to go through enough beans to make yourself a cup. Some people just don’t have the patience. Even though the amount of time it takes is nearly equivalent to the amount of time it takes to bring water to boil in an electric kettle, I’ll confess there are times when I lament the inconvenience. And I’ve been known to ride bicycles from Brooklyn to the far reaches of the Bronx on a whim. Twice in a week. Partially for spite. Again, I digress. Point is, my roommate and fellow coffee lover one day recently brought home an automatic burr grinder (her boyfriend “had an extra one”). Little did we knew it was broken. Even littler did we know how it was broken. Of course, this is me we’re talking about, so I took it apart.

Oh shi–looks like someone’s got a few teeth knocked out. OR MORE LIKE GROUND OUT. Srsly, guys: it’s like a bad trip to the dentist in here!

Sure enough, the drive gear for the burr had a few teeth missing (it also happened to be made of shitty, presumably die-cast mystery metal). This explains why the grinder would work for a split second then sound like it was powering up for an attack. Well, what’s an engineer to do in a world of planned obsolescence? I guess order a new gearbox, actually. Turns out this was a common problem with their older models. Except I also have a 3D printer and the ability to CAD things, so, uh, I decided to print some new gears. Here’s what I ended up making (pinion gear and drive gear):

The red gears are mine. Herringbone for the win!

Of course, the little pinion that slips into the motor drive shaft splines has a pitch diameter wayyy larger than the original one. This thing went from having a drive ratio of 1:13 down to, like, 1:3. I thought about making some planetary gears (or borrowing from some creative folks who have made 3D printed harmonic drives), but ultimately I decided the geometries involved would be overly complicated and too much of a pain. I know, I know, boooo. Except now I’d get to play with the “speed control” of the grinder’s circuit board, so yayyy! I measured the resistance across the potentiometer that was the speed control as 22.6 kΩ, and the total resistance across the series of resistors with which it was connected as ~66 kΩ (or 88 kΩ when I adjusted the potentiometer to its most extreme position):

88 is great,
88 is great. Actually, it’s not good enough.

Well, if I were to assume that the motor’s speed was linearly dependent upon the total resistance of this speed control circuit, and if I had to reduce the motor’s speed by a quarter (3 x 4 = 12 ~ 13), I’d try to get the total resistance of said speed control circuit up to 66 kΩ x 4 = 264 kΩ. This involves replacing the speed control resistor (that involves some soldering). A little melting of lead and tin later, I was here:

This is incredibly hacky. Don't poke fun. Okay, poke fun.
This is incredibly hacky. Don’t poke fun. Okay, poke fun.

I got to 220 kΩ. Close enough. Especially since I’m straight guessing about the linear dependence of the motor speed on its resistor string that “controls” it. Well, let’s fire put it back together, fire it up, and grind some beans!

Actually, terrible idea. So many things are wrong. The motor’s no longer running in its optimal power zone, and it’s got a seriously reduced mechanical advantage. Jam city. I mean, after clearing the jam, fiddling with the thing, and dropping one bean in at a time, I was able to grind enough beans for my morning cuppa (it might have been an afternoon cuppa at that point, and I wasn’t happy). Another thing I figured out, as insult to injury: you know that assumption I made about the linearity of the total resistance value of those series of speed control resistors to the speed of the motor? Reasonable assumption, but, yeah, totally bogus. Know how I could tell (without a tachometer)? Because the angular velocity of the output shaft of an electric motor is about linearly correlated to its windings’ input voltage (this is actually true). And because I used an oscilloscope to figure out the root mean square AC voltage fed to the motor under various speed control resistor regimes:

Graphs aren't daft: they're INFORMATIVE, guys!
Graphs aren’t daft: they’re INFORMATIVE, guys!

Thyristors are often used in fun ways to make them control power by cutting off certain sections of the AC wave. By soldering in ~190 kΩ of resistors in place of the potentiometer, I was able to get the thyristor to cut things off at 50 V (and turn them back on at 100 V). This looks nice and all, but if you actually integrate under the curve, etc. you come to realize that the RMS voltage coming out of the thyristor is not even half that of the one coming out of the thyristor when it was hooked up to the potentiometer in its factory-set position (48 V vs 92 V). This means, again due to the well-established (read: by other, more reliable folks) roughly linearly dependent relationship of motors speed and voltage, that the burr was spinning over twice as fast as it was supposed to.

All this is to say that while this was engineering fun and games and all, I’m pretty stoked to have that manufactured-supplied upgrade gearbox come in tomorrow. Such is life.