build diary

November 4, 2004:
Well, I've been innundated with helpful suggestions about the windshield!
Thank you everyone. There's a lot of information and a lot of interest in how to accomplish this. After talking to my local fabrication shop, I'll try annealing the aluminum several times before attempting to bend it and letting it air cool. I'll put a 1/16" shim in between the plywood and the channel to take up a little bit of slop that's letting it kink inwards. Once I've tried the bend and if it kinks again, I have been given a set of instructions on the best way to "tuck" the channel back flat again. Right now the channel is cooling down from the initial annealing attempt...
entry 671

November 5, 2004:
Closer.
I've made a number of improvements to the various tools involved. Lots of annealing with a propane torch (and a certain amount of melted aluminum due to an acetylene torch) was the start. I also tighened up the tolerances on my bending tool which seemed to help a little and added a clamp to hold things in place - but what helped most of all was putting a metal shim on the side of the form. See, my channel was 1/2" x 1/2" x 1/16" wall. If you do the math that means there should have been a 3/8" gap - like my form - but it was actually more along the lines of 1/16" greater. This gave the aluminum the opportunity to squish inwards. Removing this option seemed to make a big difference. It's still not perfect but it may be enough to accept for the time being. One tip from a successful builder was that his channel was 5/8" x 5/8" x 1/8". Perhaps the thicker wall is less prone to buckling. I might give that a try as well if I can find it at a hardware store this weekend.
entry 672

My original bending tool.
The tolerances are a little high. Could this be the problem?
entry 673
The New!
Improved! bending tool. Tighter tolerances. Is it good enough?
entry 674
Aha!
A clamp added to the form kept the channel in place and a metal shim ensured there was no play. It's a very good bend but not perfect.
entry 675
There's still a little crimping in the channel after being bent.
entry 676
The next challenge - how to bend this?
entry 677

November 7, 2004:
I've done it!
I started by remaking the tool once again with a larger wheel, as I'd seen a bit of distortion at the end of one of my failed bends that was due to the small radius of the wheel and to some overly-chamfered edges. A 2 1/8" hole saw and a piece of 1/2" aluminum plate and I had my new wheel. I was smarter about the order in which I measured, drilled and welded on this one so it ended up with tighter tolerances than before. It was so tight, in fact, that in one of my bends it started to tear up the metal. Some good oil solved that problem on the next bend. But the real secret was in the aluminum channel.

On a trip to Home Depot yesterday, I decided to look at what they might have to offer. I was looking for some 5/8" channel with thicker walls, as it appears that Ron Champion used 5/8" from a badly-labeled diagram. I didn't find that but I did notice that the 1/2" channel appeared to be softer. It was a slightly lighter colour than the one I'd been using and seemed to have a more satin finish. It was also about $3 cheaper for an 8' length, so I bought some home. I annealed it with the propane torch (the temperature of this is just about perfect - burn off a Sharpie line and you've got it) and it bent beautifully. So that was the problem all along. Whatever kind of alloy I was getting from the local fabrication shop was less excited about bending than the cheap stuff sold in every hardware store. The display was labelled "SteelWorks" and it's carried by Lowes, Home Depot and Tru-Valu. So, the tricks to a perfect bend are the correct aluminum, annealing, no slop in the form and some lubrication. I don't think many of the other things I tried matter. Now I just have to bend the bottom form but that will be easy enough to do by hand. I have the piece of plywood that I cut out of the bottom to use for shaping and some test bends indicate that this is pretty easy. Then it's just a matter of mounting it to the car and I already have ideas there.
entry 678

Yet another version of the bending tool, this time using a larger wheel.
Nice tight clearances and a decent bend - but I should have used some lubrication to prevent scarring like this.
entry 679
Success! A perfect bend.
entry 680

November 9, 2004:
The bottom bend was done by hand on the piece of plywood that had been cut from the bottom.
The shape was then tweaked slightly against the main template. A simple and quick job.
entry 681

November 11, 2004:
I fixed the clutch problem last night so that I could drive the Seven to work.
A Time-Sert thread insert and problem solved! I still need to put a limiter on the pedal. The windshield will go on the car this weekend as we're having a bit of a boy's night out tonight, playing Gran Turismo on a giant TV with my homemade "driving pod". Hey, we're all kids at heart.
entry 682

Staff parking at Flyin' Miata.
entry 683

November 13, 2004:
I spent the day working on installing the new lift at Flyin' Miata and building the windscreen for the Seven.
Jonathan Adams and I have been chatting about how to attach the windscreen to the car and I am shamelessly using his suggestions instead of the setup I had originally suggested. I'm having a bit of trouble getting everything built to the quality that I want so it's taking a long time. I might start a couple of sections over. But it's raining outside so it's not like I can drive the car...
entry 684

The windshield will be held to the side braces by this steel insert inside the channel.
entry 685
The first part of a side brace is ready.
I'm not sure if I'll expand this to use four screws instead of three.
entry 686
This is how I'll fasten the two sections of the windshield frame together.
The rivet heat will be hidden behind the bracket when it's installed. The "book" method wouldn't have worked on my 1/16" channel. Ignore the poor fit of my test item, please!
entry 687

November 14, 2004:
Well, the windshield frame is finished and the side braces are ready to go.
I started over with my setup for fastening the side braces to the frame because I just couldn't get the finish I wanted. A note - countersunk fasteners are not very tolerant of any sort of mistake! Your eye can easily pick out an off-center fastener. Some careful cutting and measuring has given me an acceptable fit now. The side braces are done and look good. They took a little bit of tweaking in terms of angles to allow them to sit flat on the sides of the car but it did go fairly quickly. I started to install the glass - and hit a problem. The glass must be very slightly larger than the plywood. Enough that my tight tolerances were too tight but not enough to notice when you lay the glass on the wood! Once the inner tube was added, it was obvious that there was no way to install the frame around it. So I have two options - make the glass slightly smaller or make the frame bigger. I'll ask the glass shop tomorrow if it's possible to grind this down, and how. I'm happy with how everything else turned out but I had really hoped to drive the car today with the new windshield. Oh well...
entry 688

Test fitting the completed frame.
Bill's suggestion? "Bigger holes in the wood so you can see where you're going."
entry 689
Starting to build the side braces.
entry 690
The side braces, ready for paint.
entry 691
To install the windshield into the frame, I think I'll use inner tube.
It's easy to remove and gives a good fit. Two 26" tubes are needed for the whole windshield.
entry 692
<< | show individual entries | >>