Posts Tagged Repairs
Within the space of four days, we had three rear-tire flats:
- A tire liner wear-through, after which I didn’t replace the liner
- Four miles later, a blowout through a tread gash previously covered by the tire liner
- A puncture flat directly through the tread
Basically, erosion from the (last remaining, I think) liner in the rear tire of Mary’s bike caused the first flat; I patched the tube and didn’t notice the gash. After the blowout, I patched the tube again, booted the gash (with a snippet from a roll of PET bottle plastic I carry around for exactly that purpose), stuck an ordinary patch atop the boot to cover its edges, and the whole mess has held air just fine for the last week. I’m reluctant to mess with success.
Not having a tire liner caused the third flat, this time on my bike. The wound looked like a nail or glass shard punched directly through the Kevlar armor behind the tread. Fortunately, it happened (or, more exactly, I realized I had a flat) half a mile from home, so I fired a CO2 cartridge into the tube and pedaled like crazy, which got me halfway to the goal and I rolled the rest of the way on a dead-flat tire.
Ya can’t win.
So I picked up a pair of Michelin Protek Max tubes, the weirdest things I’ve ever stuffed into a bike tire:
The bumps along the tread surface are much larger and uglier than shown in that picture:
The rubber forming the protrusions has the same thickness as the rest of the tube, so you’re looking at soft, flexible shapes, rather than thick bumps.
The “liquid” inside must be a thin film over the inner surface. I’ve never been a big fan of tire sealants, mostly because they’re reputed to ooze to the bottom of the tire into off-balance puddles.
For future reference, the Official Quasi-Instruction Manual / Blurb (clicky for more dots):
We’ll see how well these work…
We agreed that repairing the failed flag ferrule made the trailer much quieter, but it still seemed far more rattly than we remembered. It just had to be the fender, somehow, and eventually this appeared:
The obviously missing piece of the fender fell out in my hand; the similar chunk just beyond the wire arch fell out after I took the pictures. Yes, the wire has indented the fender.
The arch supports the aluminum fender, with a pair of (flat) steel plates clamping the wire to the fender:
The cardboard scraps show I fixed a rattle in the distant past.
Being aluminum, the fender can’t have a replacement piece brazed in place and, given the compound curves, I wasn’t up for the requisite fancy sheet metal work.
Instead, a bit of math produces a pair of shapes:
In this case, we know the curve radii, so the chord equation gives the depth of the curve across the (known) width & length of the plates; the maximum of those values sets the additional thickness required for the plates. The curves turn out to be rather steep, given the usual layer thickness and plate sizes, which gives them a weird angular look that absolutely doesn’t matter when pressed firmly against the fender:
The computations required to fit Hilbert Curve surface infill into those small exposed areas took basically forever; given that nobody will ever see them, I used the traditional linear infill pattern. A 15% 3D Honeycomb interior infill turned them into rigid parts.
The notch in the outer plate (top left, seen notch-side-down) accommodates the support wire:
The upper surface would look better with chamfered edges, but that’s in the nature of fine tuning. That part must print with its top surface downward: an unsupported (shallow) chamfer would produce horrible surface finish and life is too short for fussing with support. Given the surrounding rust & dings, worrying about aesthetics seems bootless.
The original screws weren’t quite long enough to reach through the plastic plates, so I dipped into my shiny-new assortment of stainless steel socket head cap screws. Although the (uncut) M5x16 screws seem to protrude dangerously far from the inner plate, there’s another inch of air between those screws and the tire tread:
Given the increase in bearing area, that part of the fender shouldn’t fracture for another decade or two.
I loves me my M2 3D printer …
The OpenSCAD source code as a GitHub Gist:
The original dimension measurement and design doodle:
At some point along a recent grocery ride, the top half of the flag mast on the BOB Yak trailer went missing.
We had a general idea of where it happened, but, fortunately, I Have The Technology:
The flag and pole ended up just off the road, only slightly the worse for wear. I hadn’t planned on riding two dozen miles on a rather hot and humid summer day, but so it goes.
The lower ferrule chafed away enough of the fiberglass pole that it could slip downward, eventually releasing the upper ferrule:
That split near the end enlarged the pole enough that the ferrule couldn’t slide off, so I contented myself with cross-drilling the whole affair for a 1-72 screw, packing epoxy into the hole, tucking more epoxy up inside the bottom end of the ferrule, then burying the screw and nut:
While I had it on the bench, I replaced the somewhat shredded fluorescent orange tape just under the flag and added a strip of diagonally striped red-and-white retroreflective tape for an attractive barber-pole appearance.
That should last for a little while longer…
Two bags of knurled brass M4 inserts arrived from halfway around the planet, so I could fix the offending hole behind the LMS mini-lathe’s electronics box:
Although you should remove the lathe from the chip pan and do it right, I gimmicked up a reducer for the long drill extension that, IIRC, came with the house:
I figured that would be close enough, given the starting situation. The cast iron frame is perhaps half an inch thick at that point, with steel brackets bolted to the far side, so use the hole as a guide and don’t drill with wild abandon.
A long M4 screw serves to align the insert eyeballometrically perpendicular to the surface while the JB Kwik epoxy cured:
It definitely doesn’t look like it grew there and, indeed, looks like the obvious repair job it is:
I thought about replacing all the screws, but decided it was so well hidden that, if I didn’t tell anybody, they’d never know:
The power switch on Mary’s “embroidery” Kenmore Model 158 sewing machine became exceedingly stiff, to the extent she said it was painful to push. Buying a shiny new switch seemed iffy, because a cursory search through the usual reputable electronic suppliers suggested there’s no way to specify how stiff the button might be, nor how that might feel in actual practice.
The switch harvested from the pulse-drive machine felt somewhat less stiff, so I decided to (try to) loosen it up and, if that worked, swap it for the stubborn one.
A pair of rivets hold the two halves of the switch together, obviously intended as a permanent solution. A carbide burr in the Dremel tool dealt with them easily enough:
Inside, the actuator drives a rotating brass contact:
Two stationary brass contacts are spot-welded to the wires:
The actuator under the button consists of a helix-twisted steel rod, a rather stiff spring, and a four-vaned phenolic blade that engages those two little flaps on the rotor. The rivet holes exactly fit plain old 1-72 screws:
Not seeing anything obviously fix-able inside, I wiped the excess oil off and reassembled it in reverse order:
Astonishingly, that bit of attention loosened it up: the button now presses easily!
I swapped it with the too-stiff switch and declared victory…
Mostly as an excuse to use the mini-lathe’s MT3 headstock collets, I made a cover for a tuning whistle (it’s an A, if that matters) case that’s been rolling around on the bench for far too long:
Yeah, it needs a bit more polishing and maybe a fancy 3D printed wrapper…
By some small miracle, one of the cutoffs in the brass tubing heap was exactly the right diameter and length, needing only a cap.
A cap looks a lot like a random piece of brass shimstock held in place with silver solder:
Fire the propane torch:
I trimmed the shimstock around the tube with scissors, grabbed it in a collet, and laid into it:
That’s just before the last few passes bringing the shimstock and solder fillet down to the tube OD, which sat nicely concentric in the collet. The carbide insert worked surprisingly well and produced shavings resembling stringy dust.
The collet drawbar, a.k.a. a hardened 3/8-15 bolt and washer, requires a distressing amount of effort to clamp the collet around the workpiece. I think it wants a Delrin / UHMW washer or some such to reduce the friction; a full-on thrust bearing seems uncalled for.
The little red shift indicator tab in the SRAM X.9 rear shifter on Mary’s bike snapped:
In a triumph of hope over experience, I tried gluing the pieces with a bit of fixturing and a dab of IPS #3 solvent:
Didn’t work any better than the last time, of course. Every gear shift snap must apply 1000 G to that poor little tab…
What’s new & different: one can now obtain Official Repair Kits consisting of the indicator tab, the plastic cover, and the two screws for $6.47 delivered from eBay.