The Smell of Molten Projects in the Morning

Ed Nisley's Blog: Shop notes, electronics, firmware, machinery, 3D printing, laser cuttery, and curiosities. Contents: 100% human thinking, 0% AI slop.

Category: Recumbent Bicycling

Cruisin’ the streets

Makergear M2: Nozzle Z Offset Recalibration

After a few days of downtime, an Official Makergear Thermistor arrived and is now installed amid a dab of heatsink compound:

M2 – Thermistor with heatsink compound

With the hot end set a bit higher than usual, position the platform at Z=0, lower the nozzle to be flat on the platform, tighten the lock screw, then run off a set of large calibration squares:

M2 – Nozzle Z Offset Recal – first test

The scrambled square in the front left says the Z=0 nozzle position came out just a bit too far above the platform and, indeed, the measurements (upper left numbers) say it’s off by 0.15-ish mm:

M2 Nozzle and Platform Re-Cal Measurements

Probably a little PETG stuck to the nozzle; I hate adjusting things when they’re burning hot.

The walls are also thin by a smidge, but the first order of business is to reset the Z offset with M206 Z=-2.15. With that in hand, the second set of squares came out at 3.00 to 3.08 mm (lower left numbers), which I defined to be Close Enough.

The 0.08 mm variation across the platform isn’t enough to worry about.

The first skirt threads were too thick and not solidly bonded together, but the second skirt came out normally, with a thickness from 0.21 through 0.30, which is also Good Enough.

The three-thread walls were still 1.15 mm, rather than 1.20 mm, so the EM should go from 0.95 to 0.95*1.20/1.15 = 1.05.

Next, a set of single-thread thinwall boxes to verify the Z offset and recheck the Extrusion Multiplier:

M2 – Nozzle Z Offset Recal – thinwall test

They’re dead on 3.00 mm tall, varying by not enough to worry about.

Their single-thread walls are 0.38 mm, not the intended 0.40, which suggests the EM should become 0.95*0.40/0.35 = 1.00.

It turns out the filament diameter at this part of the roll is scant of 1.75 mm, maybe 1.73 mm, so I decided to not fiddle with the EM.

The first production part came out fine:

M2 – Nozzle Z Offset Recal – DRL Clamp

The flange around the bottom of the arch support grid (in the middle) is intentional; it’s not an overstuffed first layer. The clamp sections rise from the platform just like they grew there.

So the M2 is back in operation and I have a spare thermistor on the shelf!

2018-08-24
Burnett Blvd at Rt 55: More Speed On Red
We’re waiting at the end of Burnett Blvd, with the signal red and the clock at T = -0.17 seconds (photo numbers in 1/60 second frames):

RedRunner-0194

You can’t hear the car (barely visible) approaching on the far left, but we can.

T = 0.00 – We get a green light and the (more visible) car is accelerating hard:

RedRunner-0204

T = 1.00 – The car reaches the crosswalk:

RedRunner-0264

Note that the driver of the car to our right isn’t moving, either.

T = 2.03 – Car passes through intersection:

RedRunner-0326

The view from above, showing the distance between those two positions is 100 feet:

Burnett at Rt 55 – Distance along Rt 55

Do the math: 100 ft / 1.03 s = 97 ft/s = 66 mph.

There’s a reason we don’t start moving instantly when a traffic signal turns green.

T = 3.17 – We start moving, as does the car to our right, with our signal still green:

RedRunner-0394

T = 4.88 – Whoops, our signal turns yellow:

RedRunner-0497

T = 9.28 – Our signal turns red:

RedRunner-0761

The signal timing hasn’t changed over the years:
- Green = 4.88 s
- Yellow = 4.40 s
Elapsed time from green to red: 9.28 seconds. No problem if you’re a car, death if you’re a bike.

T = 10.42 – We’re pedaling hard in the intersection:

RedRunner-0829

The white car to our far right started moving into the intersection about the time we did. If you’re going to say we shouldn’t run the light, you gotta deal with cars first, OK?

Note the car approaching from the right on the far side of Rt 55. That’s a 40 mph zone, the driver sees a green light, and we’re still in the intersection.

T = 12.50 – We’ve been moving for 9.33 s, which puts Mary directly in the path of the oncoming car:

RedRunner-0954

T = 14.83 – The oncoming driver having spotted us and slowed down, we’re asymptotically approaching the right-hand lane of Rt 55, passing beyond the steel manhole cover:

RedRunner-1067

If you plunk “burnett signal” into the search box at the upper right, you’ll find plenty of previous incidents along these lines.

Despite bringing this hazard to their attention many times (“We appreciate and share your interest in making our highway systems safe and functional for all users.“), NYS DOT obviously doesn’t care.

If any of their employees commuted to their office building (which overlooks this very intersection), perhaps they would care, but they don’t: we have yet to see a bicycle in the DOT’s token bike rack.

DOT says they’re in favor of Complete Streets, but, seven years on, it’s just another day on the only route between Arlington and the Overocker Trailhead of the Dutchess County Rail Trail.
2018-08-17
Monthly Image: Belmar Bridge

About five miles south of Franklin PA along the Allegheny River Trail, the Belmar Bridge carries the Sandy Creek Trail over the Allegheny River:

Belmar Bridge – Rivets

The gap in the rivets along the main truss show where someone pried off the bronze plaque surely commemorating the bridge. The scarred surface suggests a bronze-steel battery was in effect for quite some time.

I’m a sucker for big ironwork:

Belmar Bridge – Truss Joint

It’s a look at engineering done in the days of slide rules and limited data, when overengineering wasn’t nearly as bad as ensuring the thing never, ever fell down.

The bolts holding the beams and struts together show considerable confidence:

Belmar Bridge – Ironwork – bolt detail

Each bolt counts as single point of failure, but this one can rust for a long, long time before the risk becomes important.

Each of those gazillion rivets required a crew to heat white hot, shove into the hole, and hammer tight.

They don’t make ’em like that any more and I suppose it’s a good thing …

2018-08-15
Baofeng UV-5: Squelch Tail Elimination

Baofeng UV-5 radios can (mostly) eliminate the loud hiss heard at the end of a transmission before the squelch kicks in after the received carrier drops: Menu → 34 STE → ON. A detailed description of the option suggests it’s a 55 Hz subaudible tone sent for 250 milliseconds after the sender releases the PTT and before the transmitter stops sending, with the receiver muting its audio during the tone. Obviously, this requires a Baofend radio at each end of the conversation, which applies to our bikes.

Saying “laaaa” while kerchunking (into a smaller dummy load than the hulk) with STE OFF:

Baofeng – STE OFF – laaaa

Compared to the received audio, the squelch tail hiss is really really loud.

Then with STE ON:

Baofeng – STE ON – laaaa

You can see the STE tone reception start about 250 ms before the audio cuts off, although it’s not at all clear the audio is muted on either end. In any event, there’s no squelch tail worth mentioning, even if there’s an audible tick when the STE tone starts.

Saying nothing with STE ON:

Baofeng – STE ON – silent

It’s unlikely the audio output would include the subaudible tone, but you might convince yourself something happens in the 250 ms between the STE blip near midscreen and the final pop (now clipped) as the audio drops.

All in all, a definite improvement!

2018-08-13
M110A2 203 mm Self-Propelled Howitzer

It could be an M107 155 mm gun, but the double muzzle brake (the front vent is shadowed) identifies it as a mighty M110A2 203 mm self-propelled howitzer:

M110A2 203mm Self-Propelled Howitzer – York PA

Back in the day, being 30 km away from a kiloton or ten of nuclear blast was deemed Far Enough, although nobody actually pulled the string to find out. Apparently, sections of surplus barrels make hella-good bunker buster bombs, at least when you’re in a hurry.

Obsolete, of course, explaining why it’s parked behind the York Agricultural and Industrial Museum, seen from the wonderful Heritage Rail Trail. We rode south from York almost to the the Maryland line, then back again; a good time was had by all.

2018-08-11
Baofeng UV-5: Audio Attenuation and Knob Pointer

Perhaps because we’re using better quality earbuds, the Baofeng UV-5 radios on our bikes produce extremely loud audio, even with the volume knob just above its power-on click. Reducing the volume requires a series resistor downstream of the diodes clipping the pops:

Baofeng Headset Audio Attenuation

The color codes come from previous work.

Because we have different earbuds and different hearing, my radio has a 140 Ω resistor and Mary’s has a 430 Ω resistor. Getting the right value requires a few iterations of on-road testing, but it’s not particularly critical; the volume knob should end up roughly in the middle of its range.

For now, all the “circuitry” lives among layers of Kapton tape:

Baofeng headset wire plate – detail

Speaking of volume knobs, Baofeng radios have large flat-top cylindrical knobs (unlike Wouxun’s fluted knobs), so I added a pointed snippet of reflective tape to make the position visible:

Baofeng volume knob – reflective pointer

The flash lights it up, but there’s enough backlighting behind your (well, my) head to make it easily visible under normal conditions. Once you figure out the proper volume, it’s easy to set the pointer in that direction before every ride.

To the road!

2018-08-10
Baofeng UV-5: Squelch Pop Suppression

Our first ride with the Baofeng UV-5 radios subjected us to loud pops around each transmission. Back on the bench, this is the signal applied to the earbud during a no-audio simplex kerchunk:

Baofeng – squelch pops

The small noise burst to the right of the center, just before the downward pulse, happens after the carrier drops and before the squelch closes; it’s familiar to all HT users.

The huge pulses, upward at the start and downward at the end, cause the pops. They’re nearly 3 V tall, compared with the 300-ish mV squelch noise, and absolutely deafening through an earbud jammed in my ear. Mary refused to listen, so we finished the first ride in companionable silence.

I think the radio switches the audio amp power supply on and off to reduce battery drain. It’s obviously a single-supply design, so we’re looking at a hefty DC blocking capacitor charging and discharging through the earbud resistance. I suppose that’s to be expected in a $25 radio.

The obvious solution: clamp the audio signal to something reasonable, perhaps with a pair of nose-to-tail Schottky diodes across the earbud. Rather than using axial diodes, along the lines of the 1N5819 diodes in the WWVB preamp, I used a BAT54S dual SMD diode as a tiny clamp:

BAT54S dual-Shottky diode – SMD package

No pix of the final result, but it’s basically two wires soldered alongside the SMD package, surrounded by a snippet of heatstink tubing to stabilize the wires and protect the SMD leads. It might actually survive for a while, even without the obligatory epoxy blob.

The BAT54S clamps the pops to 200-ish mV, as you’d expect:

Baofeng – squelch pops – clamped – 500mV-div

That’s a kerchunk at twice the vertical scale. The very thin spike at the start of each pop isn’t audible, as nearly as we can tell, and I’ve cranked up the audio gain to make the squelch noise more prominent. Your ears will determine your knob setting.

With the audio amp applying 3 V to the diodes at the start of each pop, you’re looking at an absurdly high pulse current. I’m sure the radio exceeds the BAT54 datasheet’s 600 mA surge current limit by a considerable margin, but I’m hoping the short duration compensates for some serious silicon abuse.

Tamping those pops down made the radios listenable.

I’ve often observed that Baofeng radios are the worst HTs you’d be willing to use.

2018-08-09