When last we saw Otto, his head was literally dangling from the ceiling of my garage. Pretty grisly. Once the head was off, I slid the head gasket off the top of the block and held it in front of me like a prize bass, fully expecting to see clear evidence of failure between the #3 compression ring and one of the coolant passages. That would explain both the low compression and the fact that the cooling system sounded like a fish tank when I did a leakdown test.
Instead, the head gasket looked fine. There was no smoking—or gurgling—gun.
Unfortunately, this meant that the cause of the loss of compression was probably a warped or a cracked head. However, that wouldn’t be too bad, although the scope of the project had just expanded quite a bit. I could no longer simply clean the head and the block, slap a new gasket in there, and torque the head back down.
I rolled up my sleeves, took a deep breath, and began to disassemble the bolt-on components of the head. Much of this work is trivial, and an air-ratchet wrench makes short work of removing all the nuts holding on the intake-manifold sections.
But then I came to the exhaust manifold.
It looked like it had never been off. The eight studs and nuts formed eight small rusted units, like eight toddlers sitting in high chairs daring me to serve them pureed beets, just waiting to spit them in my face.
I blasted each one red-hot with a MAPP-gas torch, then applied a wax stick to the threads. (I’m so enamored of the use of heat and wax that I added the technique to the “Removing Stuck Nuts” chapter in the new printing of my first book. The section head is “The Brazilian.” I’m stunned that Bentley let me get away with it.)
When nuts rust like this, they can actually drop one or two sizes, so disassembly often requires trial-and-error fitting of smaller-size metric or English sockets or box-end wrenches, tapping them on with a hammer. If there’s nothing hexagonal left, it’s time for the Vise Grips.
To my profound delight, on all eight, either the nuts came off or the studs backed out of the head, which is just as good. I didn’t have to cut or drill or tap anything.
Next came disassembly of the head itself. I could’ve just taken it in to the machine shop with the cam, shafts, and rockers still installed, but they charge for disassembly. Plus, disassembling it lets you see the state the rockers and shafts are in, and judge for yourself whether you should reuse them.
On these old single-overhead-cam BMW engines, the cam is held in under tension by the rockers. The way the factory manual says to disassemble the head is to use a BMW special tool, a jig that screws down on top of the head, presses down on the rocker arms, and forces the valves open, releasing the grip of the rocker pads on the cam lobes and allowing you to withdraw the cam. Then you remove the jig, undo the rocker clips, and use a soft metal (e.g., aluminum) drift punch to drive the rocker shafts out, unthreading the rocker arms, springs, and spacers along the way.
If you don’t have the special cam-removal jig—and almost no one does—you drive out the rocker shafts the same way; but because everything is under tension, it’s more of a battle. I thought that maybe I could build, buy, or borrow the cam-removal jig. A web search revealed none for sale, but a thorough read on bmw2002faq revealed a short list of folks I knew who actually had one. A few e-mails, a few promises of bar tabs paid at the Vintage, and a jig was on the way to my house.
When it arrived, I gleefully ran into the garage to try it out.
I immediately had two problems. First, seven of the eight arms of the jig fit perfectly into their slots in the rocker arm, but one was noticeably off. However, I found that I was simply able to slide that rocker arm a bit to one side to get it to line up.
Second, I couldn’t screw the jig down far enough to release the rockers from the cam lobes before the intake and exhaust valves were perilously close to touching—and if you crank it down when they touch, you’ll bend valves. Fortunately, I found a PDF of the original directions on bmw2002faq that listed the trick: Crank the exhaust side of the jig down first, making sure that the exhaust valves open before the intakes. Then they don’t hit.
In ten minutes I had the cam out. But before I removed the cam completely, I slid it back in place. I’ve seen heads warped so badly that the cam wouldn’t fit back into the journals without a lot of pressure. This one felt fine—so if the head was warped, it couldn’t be by much.
Even with the tension removed from the rocker shafts, it’s a bit of a pain to drive them out. They’re supposed to be snug—and age, heat, and varnish tighten things up considerably. If the head is noticeably warped, they can be murder to get out. As I said, you smack them at one end with a sacrificial drift that’s softer than the rockers. I’ve tried drift punches of wood and Delrin, but those just crack. I now use aluminum rods bought from McMaster-Carr; I use a Sawzall to cut them into a few lengths, the longest being just slightly longer than the rocker shafts so you can drive them all the way out.
Both the hammer-beating and end-of-the-shaft-beating ends of the aluminum rods will mushroom from impact. You need to watch them and pull them out of the shaft holes through the head before they mushroom so big that they won’t get through, and file off the mushroomed part. The main thing is that you really, really, REALLY don’t want to mushroom the end of the rocker shaft itself. If you do, you may be hosed; it may not want to go forward or backward. So, when doing this job, you really don’t want to get your mechanic’s red mist up, lean into it with the hammer, and say, “This baby is coming out RIGHT BLOODY NOW.”
It took maybe fifteen minutes per side, but I got both shafts out.
Some folks buy an inexpensive valve-spring compressor and use it to pull the valves and seals out, but I’ve always taken the head into the machine shop as-is, because I’ve never felt competent to judge wear on the seats and guides. So in to the machine shop it went.
I told Hal, the machinist, what the symptoms were on #3, including the low compression, the air in the coolant, and the fact that the two head bolts on the exhaust side of #3 were very difficult to remove, as the threads were covered in carbon mung.
A day later, Hal called me. He said that the head was warped 0.007"—enough to explain all the symptoms, but not so much that it couldn’t just be milled flat. Hal found no visible evidence of cracking, and advised that since the symptoms were all explained by the warp, and since I smelled no antifreeze in the exhaust, it didn’t need to be pressure-tested.
Later that day, I realized that if Hal was going to mill the head, I should take him the upper timing cover so that he could mill it by the same amount. I called the shop and Hal answered. “Rob,” he said, “I’m so glad you called. I forgot to tell you that you need to bring me the upper timing cover so I can mill it at the same time as the head.”
Nice that after doing this for nearly 35 years, I actually know a few things.
And so I wait. I’d like to say I’ll start to clean up the threaded holes in the block, but really, I’ll be working on preparing Kugel the ’72 tii for the trip to the Vintage. The Recapitation of Otto will commence when I return.—Rob Siegel
