Sunday, August 19, 2007

Big patch or little patches

I started off today thinking I would pick up where I left off last week. After cutting away some more rust I decided that I really ought to fabricate a big piece of the rear end, pretty much everything below the tail lights, side to side.

What got me started down this thought path was the difficulty in making so many small patches and having the result turn out smooth. I am not a welder. I plan to teach myself how to weld with this project to practice on. With so many rustly spots so close together I am afraid the result will look like a cotton shirt that didn't get ironed.

As I looked more closely at how this part of the car is put together I began to have doubts. The entire rear panel is one piece, with several shallow but tight horizontal folds running side to side. It is a compound curve, with a lip along the bottom. Easy to make with a factory stamping die, difficult even with proper sheet metal tools (a beader and an English wheel come to mind). Lacking such tools and the associated skills, such a piece is beyond me.

I recalled a conversation I had with one of the guys at the shop back when I was exploring the front. Tom's advice was not to try to make perfect welded in patches -- let's call that artisian style body work. His suggestion was to make a "pretty good" patch, then blend it in with bondo. I decided to forget the dream of a perfect replacement and go back to making as small as possible patches. I keep promising myself that when it comes to bondo I will show more restraint that the previous owner.

One reason this approach should work well for me is the complete lack of accident damage. There is no practical way to repair a crushed rear end without splicing in a new section, but rust holes can be cut away without destroying the overall shape.

The challenge that revealed itself today is that the rear end rusted along the joint with the trunk floor. The edge of the trunk floor is flanged upwards and the rear end is welded to it. Dirt and moisture collected along this joint. As I cut away the rust I found that a lot of the weld is still good, but it needs to be removed to create clearance for the new skin. I'm thining maybe I need to buy a corser grinding wheel.

BMW081907-01.JPG
More rust cut from rear end below red light and extending right. Ground away bondo on remaining metal below tail light, looks like old repair using lead.

BMW081907-02.JPG
Exterior view showing rust at joint with trunk floor. Rear section is one piece with several shallow horizontal bends. Repairs will be built up. Area visible through cut-out is curved wall of spare tire well.

BMW081907-03.JPG
Rusted area cut away from trunk floor. Need to grind away places still stuck together to provide clearance for patch.

BMW081907-04.JPG
Inside view of bumper attach point. Note doubler, might not be original. Curve of spare tire well just visible. Trunk floor flange bends up, rear panel skin welded to it.

BMW081907-05.JPG
Inside view, more visible. Angled toward left side of car.

BMW081907-06.JPG
Inside view angled toward right. Rusty slot is above today's cut out area, which was below the trunk floor.

BMW081907-07.JPG
Close-up of rusty slot above trunk floor flange.

Wednesday, August 15, 2007

Sticking together

When I decided to undertake this project -- mind you, I'm still in the feasibility phase -- I decided that the time had come for me to learn how to weld. The shop had a couple of welders, a huge old Miller tig welder and a new little Lincoln. It turns out the big unit, which had been donated, was so old it was hard to find the pieces needed to get it going. That left the Lincoln, a Wel-Pak HD flux-core rated at 35-88 amps.

After spending several hours trying to weld pieces of scrap sheet-metal together and getting nowhere I started to ask around. I was told that the Lincoln was not a good match for auto body sheet metal work. As soon as the work became hot enough to bond it burned through, and the flux-core wire was spitting gunk all over the weld area that interfered with the bond. One guy said it was designed to weld 1/8 in. steel, not thin sheet-metal.

One system recommended by GasPro that looks good to me is the Miller Millermatic 140 with Auto-Set. This unit can be used with inert gas, something the shop's Lincoln lacks, and is rated 30 - 140 amps. Miller claims it can weld mild steel from 3/16 in. all the way down to 24 gauge, thinner than the E9's body panels. Base price $789.00.

The slightly larger Miller 180 looks like a good machine. Same capabilities with more power but without the automatic setup, for about $200 more. Problem is, it requires 220 volts, and the only 220 outlet on the shop floor is nowhere near my bay.

Lincoln also sells a 120 volt MIG welder rated at 140 amps. The 140T uses a tapped transformer and is limited to a stepped output, while the 140C is continuously variable. The web page hints at why I am having so much trouble: "... for MIG welding on thin gauge steel, stainless and aluminum. Or, use gas-less flux-cored welding for deep penetration on thicker steel, even outdoors!" See that part about no gas and flux-cored wire? That gives better penetration on THICKER steel. We have to guess how thick thicker is, but it appears to be thicker than thin.

The July 2007 issue of Popular Mechanics, the engineer's bible, has a story on doing rust repairs using epoxy instead of welding ("No-Weld Rust Repair," John Decker, pg 119). The basic technique is the same. After cutting away the rusted area you install a backing plate, then install the patch. The difference is that in place of welds the pieces are stuck together using epoxy. Messy, but it saves buying a welder and eliminates panel warping caused by the heat of welding. Something to consider, especially since for years aircraft have relied more on glue than rivets. Here are some likely sounding products: Smooth-On and LORD Epoxy Adhesives.

Some old skool guys prefer to use gas welding for automotive body work. The attraction is that the weld rod can be so similar to the sheet-metal that after filing and sanding you cannot see the weld. The down side is the likelyhood of heat warping.

Whichever way I decide to go I need to decide soon.

Sunday, August 12, 2007

Cutting virgin steel

Today I reached a new milestone. I made my first patch. As a result, my thirty-five dollar sheet of galvanized steel is no longer virgin.

I started off trying my hand at cutting away a rusted spot using a shop die grinder with a cut-off wheel. I wanted something easy, so I picked a spot along the lower edge of the left tail light. Easy.

I decided to try another spot and moved a little to the right, with visions of having the rear of the car looking like Swiss cheese by the end of the day. I just started the second cut when I noticed that the lower right corner of the tail light opening was rusted away. Same for the lower left corner of the odd rectangular red light just inboard of the tail light. I decided that before I removed anything more I needed to plan out my patches.

I ended up making a simple, flat backup piece to reinforce the hole for that odd rectangular light. I thought I would use the shop's plasma torch but Carl (the manager) thought I should use a die grinder. After making this one I definitly want to try the plasma cutter, because this way took a long time. I can understand using the die grinder on the car body, where the torch would have melted the wiring harness still in the neighborhood.

One tricky aspect to this spot are the sharp, shallow bends around the tail light opening. The bottom one continues to the opposit tail light. I decided that reproducing these bends is impractical, so I will construct the step using two layers of sheetmetal.

It may not look like much, but with a project like this you have to expect forward progess in small steps. If I were a commercial shop working at this rate I'd be broke in no time.

The next big decision is how to attach these pieces. The shop's welder is not up to the task. I might buy my own, but not before I decide to go ahead. Another possibility is to use epoxy.

One more thing. As I was removing the rear trim strip I discovered that to get at the nut on the front bolt I had to remove much of the interior. I have included two photos of that. The rest illustrate today's cutting and patching.

BMW081207TrimHole1.JPG BMW081207TrimHole2.JPG Trim strip nut

BMW081207Patch1.JPG BMW081207Patch2.JPG First cut in body.

BMW081207Patch3.JPG Left - my grinder, right - shop die grinder with cut-off wheel.

BMW081207Patch4.JPG Patch after center rough cut with cut-off wheel, die grinder fitted with rotory file.

BMW081207Patch5.JPG Finished patch hanging next to location, will go on back side.

Sunday, August 5, 2007

Solex locks

I didn't work on the car last week because my wife was in the Tin Man Triathlon and I went along to cheer. It was her best ever, so that was good.

The Friday night before was her training group's carbo loading party. It took place in one of Honolulu's better neighborhoods, yet while we were inside digging through fourteen types of pasta (even though I wasn't in the race I did my part) someone screwdrivered the passenger side lock on our 87 Camry station wagon. The driver side door had gotten the same treatment a month ago, but we had put up with getting in on the passenger side. Now we could not lock the car. Well, we could, but then we would have to enter through the tail gate.

Locks are a specialized business. I know I could remove and install one, but to straighten out this mess I decided it was time to call in an expert. Not knowing any I looked through the yellow pages for a shop near my office and a sensible looking ad that included the word "repair." I chose Access Lock and Key. The service was great, hard to find these days. I asked for a quote on Toyota locks so we could use the ignition key, and a quote on Solex locks. I expected the Solex locks to cost more, and so did the locksmith, but it ended up that the Solex were a bit cheaper. Better yet, Ron Pang came to my office and installed them. I was expecting an overnight stay at his place. Sure they cost a bit, but with such first class service I feel like I got a bargin.

Today's work on the E9 was uneventful. I expected to see some surface rust on the bare patches I had painted with Rust-Prep, because the instructions state the surface should be painted within 24 hours. No rust, after two weeks! Maybe because the car is inside and well protected? Just to be sure I ended today's work with a light coat of Scotch Acid Etch primer. I applied it over a coat of Rust-Prep, with a 30 minute drying period in between, per the instructions. The guy at Redline Automotive thought this would work. I will start off my next session by testing the primer for adhesion.