Ed Nisley's Blog: Shop notes, electronics, firmware, machinery, 3D printing, laser cuttery, and curiosities. Contents: 100% human thinking, 0% AI slop.
Category: Science
If you measure something often enough, it becomes science
Mary signed up for the National Bike Challenge and is currently ranked 4201 out of 32 k riders, by simply getting on the damn bike and riding. About 3/4 of her miles count as “transport”: grocery / gardening / shopping / suchlike. We’re no longer biking to work, but when we did, riding ten miles a day, every day, added up pretty quickly; we chose houses in locations that made bicycle commuting possible.
Her father, at age 84, also signed up and ranked neck-and-neck with her until cataract surgery cut into his riding schedule; their standings flip-flopped depending on who updated most recently. He’s our role model for getting old without slowing down.
I’m not participating, being far more quantified than anyone really should be.
Makes you wonder what the bottom 28 k (*) riders are doing, doesn’t it? I mean, sheesh, my esteemed wife spots most participants an entire lifetime or two; her father spots them three or four. They’re not star athletes, that’s for sure, but they’re doing just fine.
(*) The Challenge had over 40 k riders at one point. We think they’ve tossed folks who haven’t done any riding at all, which might serve to improve the overall averages.
At least until I blew out the MOSFET, which is about what I expected. It’s screwed to that randomly selected heatsink, with a dab of thermal compound underneath.
Incoming AC from an isolated variable transformer (basically, an isolated Variac) goes to a bridge rectifier. Rectified output: positive to the motor, motor to MOSFET drain, MOSFET source to negative.
MOSFET gate from bench supply positive and supply negative to source.
Hall effect current probe clamped around the motor current path.
The MOSFET was an IRF610: 200 V / 3.3 A. That’s under-rated for what I was doing, but I had a bunch of ’em.
I actually worked up to that mess, starting with the bare motor on the bench running from the 50 VDC supply. That sufficed to show that you can, in fact, control the motor speed by twiddling the gate voltage to regulate the current going into the motor. It also showed that a universal-wound motor’s square-law positive feedback loop will definitely require careful tuning; think of an unstable fly-by-wire airplane and you’ve got the general idea.
In any event, flushed with success, I ignored the safe operating area graph (from the Vishay datasheet):
IRF610 – Safe Operating Area
Drain current over half an amp at 160-ish peak volts (from rectified 120 VAC) will kill the MOSFET unless you apply it as short single pulses, not repetitive 120 Hz hammerblows.
I also ignored the transfer characteristics graph:
IRF610 – Typical Transfer Characteristics
The curve starting at the lower left should be labeled 25 °C and the other should be 150 °C. The key point is that they cross around VGS = 6.5 V, where IDS = 2 A. Below that point, the MOSFET conducts more current as it heats up… which means that if a small part of the die heats up, it will conduct more current, heat up even more, and eventually burn through.
Yes, MOSFETs can suffer thermal runaway, too.
The motor draws about half an amp while driving the sewing machine, which suggests the gate voltage will be around 5 V. In round numbers, it was 5.5 to 6 V as I twiddled the knob to maintain a constant speed.
At half an amp, the MOSFET dissipated anywhere from a bit under 1 W (from RDS(on) = 1.5 Ω to well over 25 W (while trying to maintain headway with friction on the handwheel). I ran out of fingers to record the numbers, but dropping 10 to 20 V across the MOSFET seemed typical and that turns into 5 to 10 W.
It eventually failed shorted and the sewing machine revved up to full speed. Sic transit gloria mundi.
In any event, I think the only way to have a transistor survive that sort of abuse is to start with one so grossly over-rated that it can handle a few amps at 200 V without sweating. It might actually be easier to get an ordinary NPN transistor with such ratings; using a hockey puck IGBT or some such seems like overkill.
This doesn’t happen very often, but, after a few road trips and some jostling around, the M2’s platform was definitely out of alignment: the first layer came out generally too thin, with the X-Y+ quadrant very much too thin.
I tried a quick and dirty adjustment that didn’t produce meaningful results, then broke out the Starrett Taper Gauge and did it right.
Jog around measuring the height of the nozzle above the platform
Adjust screws to reduce variation
Change Z offset in startup G-Code
Run off a few test patterns to get the platform heated
Measure actual thickness
Change Z offset to get the right answer
Done!
This progression of cold measurements, read top-to-bottom, left column first, shows the observed nozzle height above the platform around the edges and at the center:
M2 Platform Leveling Progression – 2014-06-30
The final measurements seem to indicate the glass plate is 0.2 mm convex in the center, but I wouldn’t trust the measurements to that level of accuracy. It’s probably bowed upward, but it’s certainly close enough.
The cold measurements suggest that the Z offset should be -4.80 mm, but the measurements on the hot platform with actual extrusion threads showed that -4.50 mm produced the correct thicknesses.
It’s not clear automating the movements would produce better or faster results than just manually jogging the nozzle around the platform, particularly since it happens only every few months.
This would be easier with the Z offset stored in the EEPROM and some modified startup G-Code to retrieve it.
For lack of anything smarter, I marked the Kenmore 158 pedal’s range of motion in 2 mm increments, starting at the top:
Kenmore 158 foot pedal – motion calibration
With the Hall effect sensor connected to a +5 V supply, the output looks like this:
Hall sensor output vs pedal depression
The point at 10 mm looks a bit out of place; other than that, the curve is about what you’d expect. The sensor saturates at about 0.84 V and 4.4 V, more or less, so you’re seeing the bias magnet on the low end and the main magnet on the high end.
Obviously, you shouldn’t take these measurements too seriously, but they’re in the right ballpark.
The pivot pin is 75 mm from the base of that line, so the subtended angle is more-or-less 16° = arctan(22/75), which is small enough that plotting the results as a function of the pedal angle doesn’t look any different.
Although you could linearize that, I think the curve has the right shape for a foot pedal speed control: it starts slowly and tapers off smoothly at the high end.
I think I could add a few more millimeters of magnet travel, but this will certainly suffice to get the crash test dummy running.
Stuffing the AC motor back into the Kenmore Model 158 crash test dummy sewing machine, tightening the belts, powering it from the bench supply, and recording speed vs. voltage produces this interesting graph:
Kenmore Model 158 AC Motor on DC – Loaded and Unloaded RPM vs Voltage
The blue curve comes from the unloaded motor sitting bare on the bench. The red curve represents a more useful situation, with the motor driving the sewing machine’s main shaft, moving the needle carrier, spinning the bobbin housing, rotating a bunch of cams, and shoving the cranks. I expect the load would be higher while it’s actually punching thread into fabric / zigzagging / whatever, but probably less than a factor of two.
The sewing machine’s top speed is around 8500 rpm, useful only for bobbin loading. Feeding that speed into the linear fit equation and turning the crank backwards says the motor would run from (wait for it) 99.5 V. The motor’s rating is 110 to 120 VAC, so it’s within 10%; that’s ignoring the whole AC vs. DC discussion and my relatively imprecise measurements.
The motor draws about 300 mA unloaded and 500 mA loaded; those values remain essentially constant at all speeds. The loaded current increases by about 10% over the speed range, likely due to increasing mechanical load / windage losses inside the sewing machine.
The locked rotor current is 880 mA at 40 and 45 V, rising to 1 A at 50 V.
The bench supply has an adjustable current limit that steps in 30 mA increments. Starting with the supply in constant voltage mode, reducing the current by 30 mA from the free running value brings the motor to a gradual stop. As with all motors, the output torque comes from the winding current, but in a (series-wound) universal motor the same current energizes both the rotor and the stator windings: there’s a square-law positive feedback loop ending in a high current stall or a low current runaway.
The usual triac speed control will not be useful in this situation, because it will generate an unacceptable level of audible noise.
Closing the feedback loop through the operator’s foot on the pedal works surprisingly well, due to the relatively slow motor response. Duplicating that with, oh, say, an Arduino might require a bit more than just a PID loop.
Having a NEMA 23 stepper fit almost exactly into the spot vacated by the sewing machine’s AC motor was too good to pass up:
Kenmore 158 – NEMA 23 stepper – on adapter
So I wired a power supply to an M542 stepper driver brick, connected the pulse output of a function generator to the brick’s STEP inputs, swapped motor leads until it turned the proper direction (CCW as seen from the shaft end), and turned the function generator knob:
Kenmore 158 – NEMA 23 stepper test
The object was to find the step frequency where the motor stalls, for various winding currents and supply voltages. The motor won’t have enough torque to actually stitch anything near the dropout speed, but this will give an indication of what’s possible.
With a 24 V DC supply and 1/8 microstepping (40 k step/s = 1470 RPM):
1.00 A = 11 k step/s
1.91 A = 44 k/s
2.37 A = 66 k/s
3.31 A = 15 k/s
With a 36 V DC supply and 1/8 microstepping:
1.91 A = 70 k/s
3.31 A = 90 k/s
With a 36 V DC supply and 1/4 microstepping (40 k step/s = 2900 RPM):
1.91 A = 34 k/s
2.37 A = 47 k/s
2.84 A = 47 k/s
3.31 A = 48 k/s
The motor runs faster with a higher voltage supply, which is no surprise: V = L di/dt. A higher voltage across the winding drives a faster current change, so each step can be faster.
The top speed is about 3500 RPM; just under that speed, the motor stalls at the slightest touch. That’s less than half the AC motor’s top speed under a similarly light load and the AC motor still has plenty of torque to spare.
90 k step/s at 1/8 microstepping = 11 k full step/s = crazy fast. Crosscheck: 48 k step/s at 1/4 microstepping = 12 k full step/s. The usual dropout speed for NEMA 23 steppers seems to be well under 10 k full step/s, but I don’t have a datasheet for these motors and, in any event, the sewing machine shaft provides enough momentum to keep the motor cruising along.
One thing I didn’t expect: the stepper excites howling mechanical resonances throughout its entire speed range, because the adapter plate mounts firmly to the cast aluminum frame with absolutely no damping anywhere. Mary ventured into the Basement Laboratory to find out what I was doing, having heard the howls upstairs across the house.
She can also hear near-ultrasonic stepper current chopper subharmonics that lie far above my audible range, so even if the stepper could handle the speed and I could damp the mechanics, it’s a non-starter for this task.
Given that the AC motor runs on DC, perhaps a brute-force MOSFET “resistive” control would suffice as a replacement for the carbon disk rheostat in the foot pedal. It’d take some serious heatsinking, but 100 V (or less?) at something under 1 A and intermittent duty doesn’t pose much of a problem for even cheap surplus MOSFETs these days.
That would avoid all the electrical and acoustic noise associated with PWM speed control, which counts as a major win in this situation. Wrapping a speed control feedback loop around the motor should stiffen up its low end torque.
The “Garden” trace comes from a waterproof Hobo datalogger buried a few inches underground, beneath a thick layer of chipped leaf mulch. The “Patio” trace comes from the center of the cramped space below the concrete patio, buried flush with the bare dirt floor. The “Water” trace is the temperature at the incoming water pipe from the town water main, which passes 150 feet under the front yard.
Calculated eyeballometrically, the temperature rose 7 °F in about a month.
The datalogger in the garden came from the “cold cellar” veggie storage buckets, so I don’t have a year-long record. On the other paw, it looks like the patio temperature will be a pretty good proxy for the minimum garden temperature.
I hand-cleaned the Hobo CSV files and fed the results into a Gnuplot script that’s replete with the cruft of ages:
#!/bin/sh
#-- overhead
export GDFONTPATH="/usr/share/fonts/truetype/"
ofile=Temperatures.png
echo Output file: ${ofile}
#-- do it
gnuplot << EOF
#set term x11
set term png font "arialbd.ttf" 18 size 950,600
set output "${ofile}"
set title "Ground Temperatures"
set key noautotitles right center
unset mouse
set bmargin 4
set grid xtics ytics
set timefmt "%m/%d/%Y %H:%M:%S"
set xdata time
set xlabel "Date"
set format x "%Y-%m-%d"
set xrange [:"07/15/2014"]
set xtics font "arial,12"
#set mxtics 2
#set logscale y
#set ytics nomirror autofreq
set ylabel "Temperature - F"
#set format y "%4.0f"
#set yrange [30:90]
#set mytics 2
#set y2label "right side variable"
#set y2tics nomirror autofreq 2
#set format y2 "%3.0f"
#set y2range [0:200]
#set y2tics 32
#set rmargin 9
set datafile separator ","
#set label 1 "Garden" at "05/31/2014",25 left font "arialbd,10" tc lt 3
#set arrow from 2.100,110 to 2.105,103 lt 1 lw 2 lc 0
plot \
"Garden.csv" using 2:3 with lines lt 3 lw 1 title "Garden",\
"Patio.csv" using 2:3 with lines lt 2 lw 1 title "Patio",\
"Water.csv" using 2:5 with lines lt 4 lw 1 title "Water",\
EOF
The Smell of Molten Projects in the Morning 