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.

Tag: Repairs

If it used to work, it can work again

Cold & Fractured Solder Joints

Joint 1 – solder not bonded to lead

A friend brought over a broken toy (well, an Argent GPS tracker) with a peculiar problem: everything worked, but after a few minutes the front-panel LEDs would get intermittent. The LEDs are hand-soldered to the board with leads that extend maybe 7 mm from the surface.

After a bit of poking around, I stuck the gadget under the microscope, at which point the problems became obvious.

See that distinct line where the solder meniscus ends at the lead? Yup, that’s the teltale sign of a cold solder joint. The lead never got hot enough to bond properly with the solder, so the failure extends all the way down through the board. The only electrical contact is at a random point where the flux layer is thin enough to pass current; as the joint heats up, that point Goes Away.

Worse, do you see (click on the pix for bigger images) the small discontinuity about 1/3 of the way down the solder cone? My buddy Eks alerted me to that failure: that’s where the solder joint fractures from repeated heat stress.

Solder Thermal Stress

Here’s the quick sketch he drew on the canonical back-of-the-envelope. I added the red oval as a replacement for his emphatic gestures; with any luck, you’ll never forget it, either.

In this case the LED is anchored in a front-panel hole and the lead is mechanically locked to the board. As the lead heats & cools, it expands & contracts (duh) at a slightly different rate than the solder. After a while, the solder cracks; it’s much less ductile than copper.

Joint 2 – clear fracture

I’m not convinced that’s what happened here, as the LED leads have a bend in the middle that should relieve the stress, but it’s at exactly the spot where he sketched the failure he’s found in many, many gadgets. Power transistors standing above boards with their backs screwed to heatsinks seem particularly prone to this failure, as they have short leads stressed by the differential expansion between copper and aluminum.

Here’s another LED lead from the same gadget. A random out-of-focus fiber enters from the right and exits around to the left rear, but you can clearly see the bad joint at the top of the solder cone and the fracture line just below the fiber.

A touch of the soldering iron generally solves the problem, although you might want to suck the old solder out so the new solder can re-flux the joint.

Arduino Pro USB (cold) Solder Pads

This doesn’t happen only to hand-soldered joints. The USB header fell right off an Arduino Pro board while I was debugging something else. I had to re-heat the joints and the header separately, add flux, and then solder ’em together. Notice the bubbles in the solder layer? That header just never got up to the proper temperature. The current version of that board uses a through-hole header, which is more rugged than this surface-mount equivalent.

TinyTrak3 cold-solder joints

And a TinyTrak3+ board had few cold joints, too, where the leads just didn’t bond at all.

In both of those cases, the vendors did a quick check and didn’t find similar problems with their stock, so the boards I got seem like random failures on the soldering line.

Now, if I’d never made a cold solder joint in my life, I’d be in a position to get all snooty. That’s just not the case: it happens to everybody, once in a while, and you just learn to live with it.

2009-07-03
Lathe Chuck: Unstuck!

Witness marks

I managed to jam the 3-jaw chuck on my lathe by turning the lathe on without the formality of snugging the chuck against the spindle first; IIRC, there was maybe 1/8″ clearance. The resounding thunk when the irresistible force hit the immovable object was the prelude to about a year of increasingly desperate attempts to remove the chuck, punctuated by long periods of despair.

The absurd derring-do with clamping the 4-jaw Sherline chuck in the 3-jaw lathe chuck described there finally prompted me to ask my buddy Eks for advice, which is what I should have done in the first place. He suggested removing the chuck body from its backplate, building a lever that bolted to the backplate with the same six bolts as the chuck, blocking the spindle with wedges under the belt pulley, and wailing on the lever with a lead hammer.

We wondered if a hard hammer would be better than the lead hammer, perhaps because the impact would be less squishy, but that was in the nature of fine tuning.

The key idea is that removing the chuck body also removes a tremendous amount of rotational inertia, so that wailing on the lever arm actually transfers force / impact directly to the backplate, rather than trying to spin the body. The somewhat risky part is that there’s a pin connecting the spindle with the drive pulley (it’s disengaged when using the back gears), so that it’s entirely possible to break the pin rather than unstick the chuck. But it was still a better idea than any I’d had so far.

Stuck backplate

Making note of the witness marks on the backplate and chuck body, I removed the body. Fortunately, there was just enough clearance between the front bearing journal and the backplate that I could get the bolts out without dismantling anything else.

That left me with the rather grody and still firmly stuck backplate. The bolt disk was brazed onto the threaded cylinder with a keyway. Although the chuck body had a key slot, it looks like the matching key had been machined off of the cylinder, so the bolts were taking all the cutting torque. Worked OK for both the previous owner and for me, so I suspect it’ll continue to work just fine for the next owner, too.

Bicycle handlebar stem in spindle

Peering through the spindle reminded me of some recent bike repairs and it occurred to me that maybe, just maybe, a old-skool split-wedge handlebar stem would get enough traction inside the spindle to hold it in place. Some rummaging in the Bike Junk box produced just such a stem and it fit exactly into the spindle bore. Now the spindle is fixed in place by its ID and there’s no risk of breaking the locking pin or (shudder) the back gear.

Even better, the lumber pile emitted a chunk of 1×4 (actual dimensions!) wood that was precisely the correct length to reach from the floor to the stem. I like it when projects work like that; finding exactly the right stuff in the pile is sort of an omen that things are going well.

Fundamental rule: always start with a hunk of something that looks a lot like what you want to end up with.

Corollary: ya gotta have stuff!

Coordinate-drilling the lever arm

Some rummaging in the parts heap turned up several feet of nice “angle iron”, so I bandsawed off a hunk. I should have realized something was wrong when a foot of teeth stripped right off the saw blade, but I ascribed that to, oh, maybe weakening a few teeth when I soldered up the blade.

I had our daughter run the trig to generate coordinates for the six holes, then lay out the center bore and bolt holes on the plate for practice. Drilling the first hole prompted me to resharpen the drill, but poking the remaining five holes into that plate produced vast clouds of wood smoke from the sacrificial plate underneath, despite boiling copious quantities of cutting fluid off the top.

I finally admitted defeat when the “angle iron” rubbed the teeth right off a 2-inch hole saw.

As nearly as I can tell, that plate is un-machinable stainless steel, hand-forged by the Devil himself specifically to taunt me, and is good for nothing. Obviously, I hadn’t used it for anything in the years it had been in my pile and, perhaps as an omen, it didn’t have any other holes in it from anybody else’s efforts. I’ll keep the pieces around just to sneer at them; won’t get fooled again.

So, at this point, I am out a bandsaw blade, a drill bit, and a hole saw. We won’t discuss the circle cutter or my abortive attempt to lash the damn thing down to the Sherline and perform helix-milling upon it.

Unstuck backplate with beating bolt

While licking my wounds, I wondered if the bolt circle on the backplate would provide enough lever arm to make any difference. I tightened a sacrificial bolt & nut with one face of the bolt aligned along a radius from the spindle center, then deployed a big drift punch and the two-pound ball-peen hammer (a.k.a., The BFH).

A half-dozen good shots later, the backplate spun free. Notice the very small gap between spindle and backplate… that’s all it takes!

I added a closed-cell foam washer to fill the gap between the backplate nose and the butt end of the chuck; there was a remarkable amount of crud built up in there.

I am so happy that it even makes up for the death toll among the tools…

It’s worth noting that the headstock has two honkin’ big bronze spindle bearings, no delicate balls, and a few mighty thwacks didn’t do them a bit of harm.

2009-07-01
Fixing a Too-small Derailleur Cable Ferrule

Brass tubing to enlarge ferrule

While replacing the rear derailleur on Mary’s Tour Easy, I rediscovered that I have two different cable sizes in my stash: large brake cables and small derailleur cables.

The large ferrules are 0.235 inches in diameter, the smaller 0.187 inches. The brazed-on cable-stop sockets are obviously sized for the larger ferrules, which makes perfect sense.

If you put a small ferrule in a large stop, it tends to cant whichever way the cable pulls it. That results in the cable sawing into the edge of the ferrule… and that results in excess friction and sometimes a broken cable.

In the past I’ve snipped out little brass shimstock rectangles, wrapped them around mandrels, and generally spent a lot of time fiddling around. This time I remembered to rummage in my collection of brass tubing cutoffs, which yielded a pair of very-nearly-perfect slip fit pieces that neatly adapted the small ferrules to the large stops.

Rear shift cable with modified ferrules

Life is good…

I have no idea what the cables look like on weight-weenie exotic-frame bikes. For sure, this isn’t a trick for hydraulic disk brakes.

Incidentally, the cable housing length worked out to 130 mm. Neither of the charts in the SRAM X.7 instructions matched the TE’s butt end; the seat stay angle is halfway between what’s normal for diamond-frame bikes. So we picked a reasonable length and it seems to be OK.

2009-06-29
Bicycle Mobile Mic Amp Debug: It’s the Connector

The radio on Mary’s bike has been misbehaving over the last few months: the PTT button on the handlebars occasionally had no effect. Debugging this sort of intermittent problem is quite difficult, as it would sometimes fail and repair itself before we could get stopped in a safe place where I could poke around in the wiring.

After months of this nonsense, I narrowed the failure down to the short cable from the HT’s mic jack to the interface board: by positioning the cable just so, the radio would work fine for days or weeks at a time. I taped the thing in position and all was well, at least for a few days or weeks at a time.

HT audio interface – 2001

HT audio interface – 2009

These two pictures show what the interface looked like back in 2001 when I put it together (modified from another version I did in 1997!) and what it looks like today. The most significant change is in the plugs connecting the whole affair to the HT: a CNC-machined plate holds them perfectly parallel at the proper spacing and an epoxy-putty turd fuses them into a rigid mass. More on that sub-project tomorrow…

Loose plugs, it turns out, vibrate the HT’s jacks right off the circuit board in short order and those jacks are a major pain to replace. Ask me how I know…

The wire break seemed to be precisely where the mic cable exits the epoxy turd. You’d expect a fatigue fracture to occur at that spot, so I wasn’t particularly surprised, although I was amazed that the thing hadn’t failed completely over the months I spend fiddling with it. I finally resolved to fix this once and for all, which meant either flaying the cable and patching the wire in situ or rebuilding the whole connector assembly. Either choice requires enough fiddly work to discourage even me.

Sooo, disconnect everything & haul it to the Basement Laboratory, Electronics Workbench Division…

Before cutting into the cable, I measured the mic voltage on the PCB and tried to make the thing fail on the bench. The HT (an ancient ICOM IC-Z1A) normally presents 3.5 V DC on the mic wire and the external PTT switch pulls it to ground through a 22 kΩ (or 33 kΩ or thereabouts) resistor. The mic audio is a small AC signal riding a volt or so of DC bias with the PTT active.

The wire measured maybe 0.25 volts and the PTT dragged it flat dead to ground. Yup, through that honkin’ big resistor. Well, maybe the last conductor in that mic wire had finally broken, right there on the bench?

Measured from the 2.5 mm plug tip conductor (tip = mic, ring = 3.5 V DC, sleeve = mic common) to the PCB pad on the PC, the mic wire stubbornly read 0.0 Ω, regardless of any wiggling & jiggling I applied to the cable. But no voltage got through from the radio to the board…

Sticking a bare 2.5 mm plug into the HT mic jack produced a steady 3.5 V on the tip lug. Reinstalling my epoxy-turd plug assembly produced either 0.25 or 3.5 V, depending on whether I twisted the thing this way or that way.

Ah-ha! Gotcha!

Pulled out my lifetime supply of Caig DeoxIT Red, applied a minute drop to the end of the mic plug, rammed it home & yanked it out several times, wiped off the residue, and the PTT now works perfectly. Did the same thing to the adjacent speaker plug, just on general principles, and I suspect that’ll be all good, too.

Diagnosis: oxidation or accumulated crud on the mic jack inside the radio.

Now, to try it out on the bike and see how long this fix lasts. Anything will work fine on the bench, but very few things survive for long on a bicycle.

Memo to Self: It’s always the connectors. Unless it’s the wires.

Here’s the schematic, just in case you’re wondering. I wouldn’t do it this way today, but that’s because I’ve learned a bit over the last decade or so…

[Update: A more recent attempt is there.]

IC-Z1A Mic Amp Schematic

2009-06-26
SRAM Twist Grip Shifter Pointer Repair

The little red pointer inside my Tour Easy’s rear SRAM Grip-Shift broke. Back in the old days, this wouldn’t be a problem, as we used friction shifters on the downtube (as we rode to school, uphill, in the snow, both ways) and knew by feel which gear was engaged. But that was then, this is now, and fixing things is what I do anyway.

The pointer turned out to be a thin plastic strip, molded into an L with a domed arch over the pointy end. It simply rests in a slot in the shifter mechanism, held in place by the transparent cover (which, mercifully, came off without dismantling the bike or even removing the cable).

I made a similar replacement from thin red-anodized aluminum, but that didn’t work out at all. The mechanism snaps from one gear to the next at roughly the speed of heat, accelerating the pointer so rapidly that the aluminum deformed. Score one for plastic!

Actually, I made two aluminum pointers. Prototype One vanished into the Sargasso heap in front of the Solvents & Lubricants Shelves at the first upshift; that’s when I discovered just how much snap that shifter applies to the pointer. Made another one, installed the cover, and then discovered it wasn’t going to work.

So I applied some Plastruct solvent adhesive to the broken plastic bits, lined the parts up on my crusty surface plate, applied a bit of gentle pressure overnight, and in the morning had a like-new pointer. It installed just fine and works like the original.

Solvent-bonded plastic is supposed to be just about as strong as the original material. We’ll see just how long this repair lasts…

Pop Quiz: Do you know the first four derivatives of position w.r.t. time?

Answer: Velocity, acceleration, jerk, and snap. You could look it up…

Update: Alas, the repair lasted only about two weeks before failing at the same spot. Some deep rummaging produced a similar (but more thoroughly dead) SRAM shifter. Turns out the pointers are similar, so I salvaged the older one. Ya gotta have stuff… and remember it, too, which is becoming something of a challenge.

2009-06-24
Avid Rollamajig Repair
Avid Rollamajig with new ball socket

Mary’s shifter cable broke at the rear derailleur, causing the Avid Rollamajig to undergo spontaneous auto-disassembly. The only part we couldn’t find was the socket between the ball and the derailleur’s adjusting thimble.

Good news: my parts heap had the Rollamajig from my bike, which I’d replaced because the most recent derailleur has an integrated gadget that serves much the same purpose.

Bad news: the socket had a chunk broken out of it and I didn’t want to put a broken part on Mary’s bike.

Good news: at least I could measure the dimensions to build a new socket.

Bad news: it needs a spherical socket for what measures out to be a 6.8 mm (0.268 inch) plastic ball and that’s not one of the three ball-end mills I have in the tooling cabinet.

Good news: this isn’t a really critical high-speed / high-stress rotating joint. Pretty good will be close enough.

Sherline chuck in lathe chuck

Turning the part was a quick lathe job on a random hunk of what’s probably nylon.

Bad news: the nylon was a rectangular cutoff from a slab and the three-jaw chuck on my lathe has been firmly stuck for the last year. It’s resisted all the non-Armageddeon-scale techniques; I fear I must machine the damn thing off.

So I…
- mounted the nylon in the Sherline 4-jaw chuck
- grabbed that teeny little chuck in the lathe’s much bigger 3-jaw
- converted one end of the square hunk into a cylinder
- removed the small chuck
- mounted the cylinder end in the 3-jaw
- completed the mission
Offset roughing mill

Lacking the appropriate ball-end mill, I offset a ball-end roughing mill in the tailstock chuck so the near side was at the right radius from the lathe axis, then poked it into the end of the socket-to-be.

Which, of course, produced a not-quite-spherical dent that was a bit too shallow, so I chucked up a too-small ball mill (on the centerline) and carved out the bottom of the socket. The result was a more-or-less spherical socket of about the right depth, pretty much.

The right way to do this, and what I was going to do before I came to my senses, was turn the part on the lathe, drill the axial cable hole, then chuck it up on the Sherline CNC mill. Getting a spherical socket of exactly the right radius and depth using a too-small ball-end mill is then a simple matter of G-Code. Maybe I should write that up for my Digital Machinist column…

Yeah, you could use a ball-turning attachment, if you should happen to have one. Sue me.

Broken and new sockets

Anyhow, it all worked out OK. The new socket is slightly longer than the old one, as it’s made to fit the derailleur thimble at hand. The end around the socket is slightly thicker, too, as it seemed more meat would add more durability where it was most needed.

The Rollamajig seems to be discontinued, although some of the smaller online sources still offer it. Building one looks like a straightforward shop project to me.

Ball socket dimensions

The sketch has dimensions in inches, because I was doing this on the lathe. Our daughter measured it in metric and came out with much the same answers, so it’s all good.
2009-06-21
Terracycle Idler Shaft Cleanup

Crusty Sliding Shaft

I installed a Terracycle Idler on Mary’s Tour Easy when the old chain tensioner wore out. It’s significantly quieter than a double-idler tensioner, but the rear derailleur can barely handle the 11-34 sprocket / 30-42-52 chainring combination.

She likes it, that’s what counts.

Anyhow, while poking around under the bike, I noticed that the idler no longer slid left-to-right on the shaft through the bearing. The bearing itself spun fine, but the shaft… ugh, they should have used stainless steel.

The sliding motion is important, as the idler should self-adjust to the chainline during shifting. I don’t know how long this one has been jammed, but it could contribute to the noises she’s been mentioning of late and that have prompted me to embark on a major maintenance project.

Cleaned Shaft Installed

It shined up nicely with a Scotchbrite wheel in the drill press and now looks merely horrible; you can see the copper plating (wrong: see Update below) showing through. I had to hit one end of it with a medium diamond file to knock off an invisible high spot.

I added a bit of lube and reinstalled it; the bearing slides back & forth like it used to, but I have my doubts as to how long this will last. Fairly obviously, the plating is shot.

The next time it fails, I’m sure I’ll wind up trying to turn an exact 0.3125-inch diameter stainless-steel shaft with a polished surface…

Oh, and the three orange retro-reflective strips? The idler turns backwards because it’s on the return side of the chain: it’s rather disconcerting and I figured it’d be fun to highlight it.

Update: The folks at Terracycle say it’s plated zinc over a brass bushing… which (Ah-ha!) explains the corrosion.

The zinc forms an anode against everything else on the bike; nothing is more anodic than zinc. Because the plating has no volume, it turns into a Fizzy at the merest sight of the usual road salt around here.

Unplated brass would be better: more volume, cathodic against steel, anodic but pretty close to stainless, just as slippery. Might tend to wear against the inner bearing race, but I’d expect it to be at least as durable as the plating.

Worn Terracycle Idler shaft

Here’s a pic of the shaft from another Terracycle Idler I had on my TE for a while. While it’s not corroded, it’s worn through to the brass underneath. So maybe the plating isn’t buying much, anyway.

I spent some quality one-on-one shop time with a random hunk o’ stainless hex rod, came up with a good-looking 0.304-inch OD (a nasty bit of overshoot, but I haven’t done any lathe filing in recent memory and forgot how fast it removes metal), and verified that the race will cock-and-jam rather than sliding nicely.

The Terracycle folks will send a replacement shaft; they’re good folks who build quality stuff and stand by their products. I’m obviously abusing the poor thing…

Update: The stainless shaft arrived and is sized for the 6 mm bolt they’re using in new production. When we discussed this, I said it’d be no big deal for me to adapt it to the existing 5 mm bolt. A length of heat-shrink tubing does the deed, as it’s rigidly held on both ends. A dab of Loctite, a dot of oil, and it’s back in service. We’ll see what happens after a few months of riding under my regime of benign neglect.

Old brass shaft, new stainless steel shaft, 5 mm bolt with heatshrink

A tip o’ the cycling helmet to Terracycle!

2009-06-17