Scion xB '06, in-dash Atom 330, Lilliput 889GL; details, pictures, links. Index: pg 1

**rdholtz** · 05-08-2009, 11:04 PM

Looks Like I Need to Go Bigger or Lighter

The results of last night's test bends are not what I expected, but they make sense. It looks like it's important to get the clamps as close to the bend as possible. My big C-clamps are somewhere in the storage area, but I haven't found them yet, so I'm using Josh's smaller ones. His C-clamps are shorter in throat depth, so they don't reach as close to the bend. The resulting bends come out with a larger radius, especially on the aluminum, and I suspect that's because the clamped surface tends to lift a little in bending because the clamping pressure is too far back from it. I've either got to find those big clamps, or go buy some.

I also see that I'm near the maximum thickness of material that I can bend consistently in this bender. If I go thicker, I need a heavier-built bender, or the same weight bender with a shorter length, because that appears to be another source of variance in the bend radius.

I guess it's no surprise that home equipment doesn't do as nice a job as industrial-strength tools. The industrial benders have a lot more beef to them, and are braced much better.

There's another approach to solving this, too: choose lighter material. This chassis is going to wind up so strong that it's almost ridiculous. I could easily go a couple of gauges lighter and still have a very strong chassis. It's just going to sit in the dash and hold some very light electronics gear, and I'm building it like it has to be explosion-proof.

**rdholtz** · 05-09-2009, 06:36 PM

Bull in a China Shop

Because I designed these parts, I think I know their dimensions by heart, even without the drawing. I accidentally proved that wasn't true the other night.

I built the relatively complex end panel, and was absolutely delighted with its final dimensions -- they were right on spec -- and, for a hand-built first part, that's decent. In the morning, I read the print. The dimensions I had pulled out of memory were wrong by the thickness of the material on on both the top and bottom flanges.

That means the end panel flanges and the top and bottom panels are trying to be in the same place. That's a conflict I avoided in the design, but added in the shop. Now I have to do something fancy with the top and bottom panels, or re-make the end panel.

Oooooops.

I don't want to remake the end panel, so I get to try something new: a double bend that produces a jog in the panel, so the top can come across to the end panel, and then jog enough to sit under it; the bottom panel would do th esame thing, but flipped. Since I've never done jogs, and don't really have the equipment for it, it means I have to be very ingenious. I think I have a technique in mind, but I'll put that off 'til I'm back in the shop next week and can try it out.

**rdholtz** · 05-10-2009, 08:15 AM

Making Holes

One difference between sample parts and production parts comes up in making holes. In production parts, the holes would be in the blank parts before bending. Because sample-making processes are less precise, I'll mark and punch (or drill) the holes after bending to be sure they match the holes in the mating parts.

I'll be using 6-32 (#6 screw, 32 threads/inch) coarse thread fasteners. These are the same as the screws used to hold cases on PCs, and the same thread used on hard drive screws. Since there are 28 fasteners, I'll have to drill or punch 56 holes.

The 6-32 rivet nuts are 3/16-inch (about 5mm) diameter, and the holes for them are 7/32 inch (about 5.5mm). The proper clearance holes for the 6-32 fasteners are #27 (.144 inch, or 3.7mm), but I plan to use a larger 5/32-inch (4mm) hole to allow some additional tolerance because the parts are handmade, and thus not as precise as a CNC machine would produce.

In a production environment, the screw clearance holes are normally larger than the base specification to allow some tolerance for manufacturing variation. Producing by hand may require even larger tolerances because it's so imprecise. If I need to go for looser tolerances, I could just use the 7/32-inch hole everywhere, and let the fastener heads make up the difference. That will still provide a nice, tight chassis; it isn't going to be in a dangerous environment, so I guess I don't need to maintain nuclear vessel construction standards.

Still, I'd like to keep this as close to production standards as possible. I'll use hole-matching, which means I'll lay the holes out on a piece of light sheet metal and drill a tiny hole just large enough for the center punch, then use that pattern "plate" to lay out the centers of all the holes I'll make; it's a sort of homemade drill guide. If I do the pattern plate well, the 5/32-inch holes will provide just enough tolerance.

**Maheriano** · 05-10-2009, 01:03 PM

Where can I buy these white connector pieces that came off the wires? I've looked everywhere locally and nobody has them. Is there an online store for it? I need them to reconnect some broken wires to the motherboard of a laptop.

**colin** · 05-10-2009, 01:16 PM

Originally Posted by Maheriano

Where can I buy these white connector pieces that came off the wires? I've looked everywhere locally and nobody has them. Is there an online store for it? I need them to reconnect some broken wires to the motherboard of a laptop.

Try futurlec. They have $4 shipping. I bought a few hundred of those for a production run of a USB I/O controller similar to the fusion brain. The also have 28awg multicolor cable and everything so that you can make your own connectors.

EDIT: Or eBay. I bought a bag of 100 male connectors, 100 female connectors, and the matching pins for a few dollars. It's hard to search though.

**Maheriano** · 05-10-2009, 07:38 PM

Originally Posted by colin

Try futurlec. They have $4 shipping. I bought a few hundred of those for a production run of a USB I/O controller similar to the fusion brain. The also have 28awg multicolor cable and everything so that you can make your own connectors.

EDIT: Or eBay. I bought a bag of 100 male connectors, 100 female connectors, and the matching pins for a few dollars. It's hard to search though.

I couldn't find them on futurlec.com nor Ebay. Do you know what they're called?

**colin** · 05-10-2009, 08:01 PM

Futurlec sells simple headers and header connectors with pins. If you want the specific white ones, they're from eBay. It took me about a week to finally find a bag of them, and that was ages ago. I got really lucky, and then I had to design my own PCB diagram of the connectors because no library had them!

All I can say is search like crazy, "header" "connector" "pin" "bag", that kind of stuff.

**rdholtz** · 05-11-2009, 06:47 AM

Colin, thanks for providing that guidance for Maheriano. I didn't have that resource, and I appreciate the help.

Seeing that pic reminded me that someday I should go back and see if I can revive that failed 10.4-inch Lilliput. Some day . . . after I finish the chassis, and the wiring, and the bezel, and then a backup battery, and . . .

**rdholtz** · 05-11-2009, 06:29 PM

Drilling vs. Punching

I thought I was out of luck using the punch for two reasons: (1) my heavy duty punch requires 1/2 inch (13mm) between the center of the hole and any adjacent structure, so I can't use that punch to make holes in flanges, and (2) it doesn't come with a 7/32 punch and die set. That meant I'd have no choice but to use a 7/32-inch drill -- in a drill press -- for the rivet nut holes. I'd be out of luck on the matching screw holes, too: the punch kit doesn't include a 5/32 punch and die set, so I'd have to use a 5/32-inch drill for them.

But I did a lunchtime run to Harbor Freight, and there was a smaller punch on sale, and it needs only 3/8 inch space. By widening the flanges a bit, I can use that. Even better, it comes with 5/32 and 7/32 punch and die sets. If you've ever worked with sheet metal and tried both drilling and punching, you know how much nicer punching is; I splurged the $20 to take that little pup away with me.

**rdholtz** · 05-12-2009, 09:47 PM

Testing, Testing . . .

It was a testing night. I pulled out the new punch and tried it out on scrap. Like its bigger brother, it makes nice, clean holes, right where I want them, and there's no deburring or cleanup necessary. It's a little less comfortable than the big one because it's hand-held, rather than mounted on a base, but it does what I need it to do, and it was only $20. Sweet.

I also tried out the new rivet nut installation tool. I threaded a rivet nut onto the tool, stuck it in the hole, crimped it, and unscrewed the tool. I noticed that the threads are slightly angled, apparently because I didn't hold the tool quite straight. That means I have to be careful to align the tool perpendicular to the work surface, or I'll get crooked threads. That's a good lesson to learn on scrap. I was more attentive on the second one, and it went in straight as an arrow.

I also noticed that these little rivet nuts are terrific. Man, they put threads right where I want 'em. I took a second piece of scrap -- one that already had a hole in it -- and fastened it to the piece with the rivet nut in it, using a 6-32 case screw. No need to fumble around trying to get a nut in there, just pop in the screw and tighten it up. Oh, that's very nice. I used a screw with serrated surfaces on the underside of the head -- as is standard on most case screws -- so I won't even need a lock washer. I like this fastening system.

That's all I got done. I was going to experiment with countersinking for the rivet nuts, but I was tired and headed off to the shower so I could get up in the morning.

Working for a living is interfering with my play time.