[Rob in VA]

Originally Posted by BlackTalon

He's secretly installing a bunch of movement sensors and a wireless communication system. So if the BMW comes close to starting a rockin', the next sound heard will be Jazz's shotgun a cockin'

[Jazzbass]

No updates in a while because I've been busy working on the car. Some progress - last post I pulled the rear subframe out, disassembled it, pulled all the bushings and repainted it. Easy-peasy.

Well, time to tackle the real meat of the job - checking out the sheet metal and looking for the dreaded "E46 subframe cracks". For the non BMW guys out there, the subframe I posted pictures of in my last post is not what actually cracks - it's the sheet metal of the unibody where the subframe bolts up. Story is that the bushings wear out, the twisting of the drivetrain under acceleration deflects the differential clockwise, which deflects the subframe, which stresses the mounting points in the unibody. Early cars (1999-2000) like mine apparently weren't reinforced very well so they tend to crack. M3s do as well thanks to more power. Lots of internet arguments about whether BMW ever really fixed this problem, especially on the M3. Some say pre-2003 M3s are the only ones affected, others say that's crap and that all E46 M3s are susceptible. Whatever the case, being an early production car, I know that my car IS in the danger zone.

So, start with removing more stuff. First the fuel tank:



Then all the evap stuff in the right rear - expansion tank, charcoal canister, and lots of old rubber hoses that are dry rotted and brackets that are rusted away.



Ugh - not as dirty as the subframe and differential, but still nasty. Man, doesn't anyone clean the underside of their $1300 car on a regular basis anymore?



Man, if only I had a lift and a pressure washer... hey, wait a sec...



Yes, clean all the things. Look at that little sway bar. It's adorable. 13mm. I can actually bend it by hand.



Ah... that's better. Not I can work in here without getting filthy. I can also see what problems are lurking under the crud.



The first thing that jumped out at me is the brake line - see the red arrow above. Here's a closer look:



Yeah, that simply won't do. Another thing to replace. Anyway, inspection of the subframe mounting points:

1. Right front: Good, no cracks
2. Left front: good, no cracks
3. Right rear: good, no cracks
4. Left rear: Uh oh:



Here you can see that most of the crack is right under the bushing area, with some hints of the cracking outside of the bushing under the undercoating. Well, a little wire wheel action will clean off the undercoating and we can see just how bad the crack is:



So - all in all, not great, but not as bad as many of the E46 horror stories I've seen. The red arrows in this last picture show the extent of the spidering of the crack. For of the spot welds in this area are affected as well. One of four mounting points cracked, but no tear outs. Could be worse I guess.

BTW, for anyone in the E46 market, doing the research on this stuff I found out a little trick. Cracks like the one I have are pretty hard to see - it's not an obvious floor-falling-out crack like you see on the internet and if you looked at the car on a lift there's a good chance you'd miss the cracks as it's almost completely obscured by the subframe. However, there is evidence of cracking that is easier to see. When the unibody starts to flex, it's the left rear subframe that seems to fail first (according to the internet and my own sample size of one). This will lead to cracking at the subframe mounting AND a failure of the spot welds at the seam behind the rear left shock absorber. When these spot welds fail the floor panel will flex away from the side enough to crack the seam sealer, like so:



Were I to go look at another E46, this would be one of the first things I'd check for.

[Jazzbass]

OK, so there are some cracks. Well, many years ago I bought a MIG welder just for things like this, and since then have literally welded hundreds of things so this should be a cinch. Well, OK maybe 10s of things. OK, so I've welded about 10 things before. Yeah, I'm a little nervous. But I went to the Choas Off Road welding course, so I should be good, right?

Actually, the course did help a lot. It was worth it to spend time with a pro and getting some tips. And I have been welding some in the past couple months, but (a) never on thin material and (b) never upside down. One of the things they stressed in the Chaos course was practice, practice, practice - a lot - before doing something for real. Good advice. So step one was to order the E46 chassis reinforcement kit from Turner Motorsport:



This is made out of 14ga mild steel and is a great deal if you consider paying $130 + shipping a good deal for $20 worth of steel that fits mostly but not exactly so you need to spend time doing additional fitment. If I were to do this on another car, I'd just make my own - the Turner kit is OK but I like the ones from Europe a lot better, like this kit from Redish Motorsport in the UK:



BTW, the Redish guys both (a) know their shit when it comes to E46s and this problem in particular and (b) have very good taste in music. This video proves both, and was a great source of info when I was inspecting my car: https://youtu.be/ubP7ustvRXs

So I'm going to be welding 14ga plates to 20ga body steel. After a quick trip to Lowes to get some 22ga and 16ga welding sheet (what was available and a decent approximation), out comes the plasma and I make up many, many batches of these:



22ga on the bottom, 16ga on the top. Now I need to weld upside down, so I clamp these things to my lift arms, like so:



My first attempt went.. well... not well, as alluded to in this thread. Vicegrip nailed my first problem - too much heat. At first I was just welding on one plate at a time and overheating the crap out of it. Next day I made up more test pieces but this time clamped three up at once and alternated between them - two welds on one, switch, two more welds, etc. I also grabbed my IR thermometer to monitor temps and took breaks when things started getting too hot. This helped a lot. Finally after three straight days of practice and a LOT of welding (and about 40 cu ft of C25) I was starting to get it. Here is my test piece progression over the three days - left is the first day overheated piece, second is day two with better heat management but still haven't found a good machine setting or rhythm, and the right was one of the last pieces from day three when things finally started falling into place.



So, time to prep the car and weld. The three good mounts just got a wire wheel and sanding disk cleaning down to bare metal. Here you can see me prepping the right rear plate - the mount has been cleaned and the plate outlined on the metal. What I did on all the mounts was tape around the weld lines shown here and then hit everything else with weld-though primer. This was done to cover as much bare metal as possible.



The left rear mount also got cleaned, but I also stop drilled the cracks and used a spot weld cutter to cut out the four broken spot welds.



Here you can see me fitting the left rear plate after it had been modified with extra holes for the spot welds. My plan here was to weld up all four broken spot welds to the new plate. The cracks on the body and stop drill holes were welded up before this. In this picture you can also see weld-through primer and the tape around the periphery of the plate.



Once all the plates were fitted (which took a little pounding and additional cutting), it was time to weld. All in all, they came out OK:



Not the greatest welds ever, but functional I think. What's funny is that my biggest worry was overhead welding, but in the end that turned out to be not so bad. The 14ga plates really helped make this easy - for whatever reason I had a lot of trouble getting an arc started on the body of the car itself. So things like the crack welding under the left rear plate were a big PITA. I didn't have any problems welding on the 22ga practice steel, so I thought the 20ga of the body would be a piece of cake. Not so. I'd either get burn back on the wire or blow holes in the metal. But when welding on the plates themselves, I could get an arc started on the edge of the 14ga pretty easy. I could them move the puddle down into the sheet metal of the body and got pretty good welds that way.

What WAS a bigger problem for me was basic out-of-position welding. This is typical n00b welder stuff, but I really, really like to weld one specific way - torch in right hand, supported in my left, face ahead of torch pushing the puddle, welding counterclockwise. As long as I could do that things were hunky-dory. However, there were some places (like the back side of the right rear plate) where I couldn't get into my preferred position due to obstacles. My welds there looked like crap - shaky, too fast, and about 50% of the time I missed drawing the puddle into the body and hit just the 14 ga plate. If you look at the welds you'll see many places I went back and tied the plates into the body with a second (or even third) pass.

Once done I ground the welds down in places where the subframe bushings would sit, primed, painted, undercoated, and top coated.



And done.

[roundel]

Outstanding!

[Trak Ratt]

Nice Jazz, great write up too! Makes me want to get some steel and start practicing!

[N0tt0N]

Big Bawls Well Handled! Good job dude!

Hrmmm..... That's going to raise the car 1/16". Now you will HAVE to go with the lowering springs and coil-overs. Too bad....

[HughA44s]

Again Outstanding. I had heard the the cracking started at the spot welds - this appears to be clear evidence of that. Also, seeing the gas tank out of the car makes it obvious why these cars (non-Ms) have fuel starvation issues in hard right hand turns (especially with sticky tires). Again, Outstanding work.

[BMAN]

Show off!!

[Lupin..the..3rd]

All these close-up shots of nude and exposed BMW underside. This is downright porny!

[Vicegrip]

I don't tend to stop in on the BMW threads but glad I did today. Good to see it came out well and proves that proper prior planning prevents piss poor performance.