Another of Mary’s glasses snapped at the temple joint:
This one has a spring inside the joint that latches the temple on either side of that square inner corner. Obviously, there’s no way to reconnect the broken stub with the spring retracted inside the brazed temple box, so:
File off the corner
Fill the socket with epoxy
Ease the stub in place
Wipe off the excess epoxy
Align on the workbench
Let it cure overnight
At least the hinge folds again, even if the spring doesn’t work:
She promises to scrap out her oldest glasses after the next eye exam…
With the Sony HDR-AS30V in its skeleton frame atop my bike helmet, the audio track for all my rides consists entirely of horrendous wind noise. You can get an idea of the baseline quality from the sound track of a recent Walkway Over The Hudson crossing.
The camera has two mics, although I’m not sure 15 mm of separation really produces meaningful stereo sound:
Note that two of the five pores on each side are closed flat-bottom pits. As with earbud vents , it must be a stylin’ thing.
I added a rounded pad of the same acoustic foam that forms an effective wind noise buffer for the boom mic:
That reduced the overall noise load by buffering direct wind impact, but non-radio conversations remained unintelligible; there’s just too much low-frequency energy.
Surprisingly, closing the mic pores with ordinary adhesive tape didn’t impair the audio in a quiet room:
Out on the road that’s even better than foam over open mic pores; I think it reduces the peak volume enough that the internal compression can regain control. Sticking the foam pad over the tape slightly reduced the noise during high-speed (for me, anyhow) parts of the ride, but didn’t make much difference overall.
The wind noise remains too high for comfort, even if I can now hear cleats clicking into pedals, shifters snapping, and even the horrible background music when I’m stopped next to the Mobil gas station on the corner.
While pondering the dead ET227 transistors, I dug an inrush current limiter (a.k.a. NTC power thermistor) out of the heap and made some measurements:
That’s from a bench power supply attached to a meter and the limiter with clip leads, which was entirely too messy for a picture.
Turning those numbers into a spreadsheet to calculate the resistances:
SCK 055 NTC Power Thermistor
5 Ω @ 25 °C
Imax = 5 A
Time constant on the order of 90 seconds
The data sheet recommends a minimum current above 30% of the maximum, which would be 1.5 A. That’s above the motor’s 1 A operating current, let alone the low-speed current limited conditions, but in this situation that just means the resistance will remain around 1 to 2 Ω with the motor chugging along.
If I had more of ’em, I could put them in series to build up the resistance, but it’s not clear why that would be better than, say, a 6 Ω aluminum-heatsink resistor dissipating a few watts.
Of course, I told the kids that Santa was on their side for getting a 3D printer under the tree…
Rumors from usually reliable sources indicate the two other 3D printers at the Faire had, shall we say, reliability issues and generally weren’t running. The M2 ran continuously from 10 am through 4 pm, cranking out eight Tuxes at a time, with no trouble at all; perhaps that’s why it got its picture in the paper.
The ET227 transistor (labeled A from the DC gain tests) I’d been using, ever since the very beginning, failed with a collector-to-emitter short when I started it for a data taking run. In most circuits, that would be a catastrophic failure accompanied by arcs & sparks, but the Kenmore 158 simply started running at full speed and ignored my increasingly desperate attempts to regain control.
OK, those transistors date back to the 1980s (or maybe even earlier), so maybe It Was Time.
I swapped in ET227-B, buttoned everything up, and continued taking data.
Two days later, ET227-B failed with a collector-to-emitter short when it turned on.
Although the ET227 can switch 1 kV and 100 A, the Safe Operating Area plot shows that the DC limit passes through 1 A at 200 V:
Bearing in mind that peak line voltage hits 170 – 180 V, 200 V looks like a convenient upper limit. Also, those limits apply at 25 °C case temperature and drop as the junctions warm up, although the datasheet remains mute as to the difference.
The circuit puts the following elements in series across the AC line:
5 A fast-blow fuse
Normally open relay
Full-wave rectifier block
120 VAC / 100 W universal motor
ET227 NPN transistor
25 T x 2 parallel 24 AWG winding
After screwing around with Spice for a while, I can’t convince myself that the simulation means anything, but the general idea is that closing the relay at maximum line voltage (about 180 V) produces a staggeringly high current pulse through the series capacitances. A small amount of stray capacitance across the motor passes line voltage to the collector, the collector-base capacitance feeds it to the base, the transistor’s gain slams essentially unlimited current against line voltage, and the operating point squirts through the top of the SOA graph.
I made up a snubber from a 220 nF X capacitor and a 5.6 Ω resistor. That won’t have any effect on the spike, because the various stray / parasitic capacitors remain directly in series across the line, so the snubber looks like an open circuit. The snubber does damp the ringing after the spike vanishes, but that’s not the problem.
Some scope shots from ET227-C show the magnitude of the problem; it hasn’t blown yet, but obviously this can’t go on. Note the varying horizontal time scales and vertical current scales (all are at 10 mV/div, with the Tek probe providing the scaling).
At 50 mA/div, the two humps come from the (damped) ringing. This one doesn’t have much of a spike:
At 100 mA/div, I must have caught it at a higher point in the voltage waveform:
At 200 mA/div, this one looks seriously worse:
Now, agreed, a 1.6 A spike in a transistor rated for 200 A pulses doesn’t sound like much, but catching the spikes depends on random chance. If the collector voltage starts at 100 V, then that spike comes pretty close to the DC SOA limit; that’s not enough to kill the transistor, but it’s certainly suggestive.
Putting an NTC power thermistor in series would add some resistance to the circuit and reduce the magnitude of the spike, but they’re really intended for power supplies that draw a constant load, not a sewing machine that starts and stops all the time. If the motor runs for a while, then the thermistor will be hot for the next startup and the relay will close with relatively little resistance in the circuit.
Plotting the motor RPM every 500 ms while increasing the nominal current by 50 mA per step from 550 mA:
And then downward from 950 mA:
No, the steps aren’t the same size going down as they are going up. The nominal current setting is open-loop = constant DAC values: the actual current for a given DAC value varies as the transistors heat up.
The motor starts running at 3700 RPM with 550 mA and stops well under 1000 RPM with 400 mA. Obviously, starting slowly and smoothly will require finesse: a swift kick of maybe 600 mA to get it turning, then immediately drop to 400-ish mA for slow stitching. Those currents must be the actual motor current, not the nominal DAC values, so the motor sees the proper current regardless of the transistor temperature.
The sewing machine requires four samples = two seconds to stabilize at each new speed on the way up, so the mechanical time constant is 2/3 second. Trying to stabilize the speed with a loop running much faster than that will certainly cause oscillation.
There is absolutely no deceleration control authority: reducing the current allows freewheeling as the machinery slows down to match the current. The undershoot after each step on the way down lasts 2.5 s, then there’s a persistent oscillation with a period of 3 s.
Forcing the firmware to run slowly enough to match the hardware should pose an interesting challenge… you don’t want to lock up the UI while the motor stabilizes!