My Most Satisfying Diagnosis Ever

I’m under the weather with a surprise case of COVID, so I’m sequestered in the third-floor bedroom where I obviously can’t wrench on anything. It does, however, give me the chance to tell you about what is and always will be my favorite diagnosis story. What follows is a massively condensed version of what took over a month to unfold.

As I’ve written about before, every spring I attend a BMW event called “The Vintage.” It’s been held in Asheville, North Carolina—a two-day drive from Boston— for nearly a decade, but before that it was in Winston-Salem, which I used to knock off in a single 14-hour day. In 2015, as I was packing the car, I got a text from a friend asking me to look at a cry for help on Facebook from a guy and his wife who were stranded in a BMW 2002tii in northern Virginia on the way to their first Vintage. I called the guy and learned that they were safe—they were in a hotel parking lot, not by the side of the interstate—but that his car was exhibiting a strange buffeting and hesitation. I told him that I’d be passing by there tomorrow, and if they were still there, I’d be glad to help.

“I certainly hope we’re not still here tomorrow!” he joked.

But they were. So after driving nine hours, I and my road-trip companion pulled into the parking lot, and there they were, the agave (green) round-tail-light 2002tii with the hood up, the guy under the hood, and the spouse on the cell phone. And now here I was, The Hack Mechanic, riding in with a white hat on to save the day.

But no pressure.

 I met Brian and his wife Michelle and instantly liked both of them. He said that he wasn’t a mechanic and wasn’t intimate with the tii’s injection bits, but he was selling himself short; clearly he knew his way around cars and tools and had owned and maintained other interesting enthusiast cars, including a Mazda Miata, a V-8-swapped Porsche 914, and a BMW M3.

There are three fuel filters in a tii—the tomato paste-sized canister near the battery, a small cylindrical screen at the inlet of the Kugelfischer fuel injection pump, and another small screen at the inlet of the fuel pump beneath the car. If there’s rust or sediment in the gas tank, the fuel pressure shoves it up against the screens, eventually lessening fuel flow enough to starve the engine, making it feel like it’s running out of gas, which, really, it is. I asked if the problem was cyclical—did it get worse as he drove the car, but then lessen if he parked it for 10 minutes, which allows some of the crud to sluff off the screen, only to repeat the cycle when driven again? This is the textbook manifestation of the crud-on-the-screens fuel starvation problem. “Not exactly,” he said; “It’s just bad all the time. And worse at low rpm.”

That last piece was unusual. Fuel starvation problems are usually worse at high rpm, where engine demand for fuel is greatest. Perhaps it was an ignition problem, not a fuel problem. Plus, Brian explained, he’d had the gas tank boiled out not 700 miles ago, so that was known good; the root cause couldn’t be rust coming from the tank. And the fuel filter by the battery was new.

Brian had already removed the main fuel filter from next to the battery, and tapped the inlet side out on a paper towel. He reported that a lot of “black stuff” came out. He had a spare filter with him and installed it. How this jived with the tank supposedly having been boiled out was unclear. But, unfortunately, after replacing the fuel filter, the problem persisted.

He had a bunch of spare ignition parts with him. We tried replacing the coil, points, condenser, plugs, wires, but nothing affected the car’s odd stumble. I couldn’t even definitively diagnose it as fuel or spark.

I also should add that Brian’s wife Michelle was the picture of what any car person would want from a spouse. At one point I was on the phone with my professional friend Paul asking his advice. He said, “I’d try replacing the fuel pump. Any AutoZone will have one for, say, a 1985 E28 535i. You can make that work.” I overheard Michelle trying to locate one before I’d even gotten off the phone.

Finally, as 4 p.m. rolled around and I still had six hours of driving ahead of me, although it broke my heart to leave the problem unsolved, I had to call it quits. Again, they were not stranded—the car still ran and drove—but pressing on at highway speed didn’t seem prudent.

As I prepared to leave, Brian asked, “Could you do me a favor? Could you think about how you’d fix the car?” It took me a while to understand what he meant.

“First,” he said, “I have to figure out a game plan. I can either get the car home, or try and have it fixed locally. I can call AAA and have them recommend a vintage foreign car repair shop, but… if you couldn’t fix it, if you couldn’t even figure out conclusively if the problem is gas or spark, what are the odds that some non-tii-specific shop that AAA tows it to is going to diagnose it correctly and fix it once and for all, so I can actually drive it home?”

My ego notwithstanding, he did have a point.

“And,” he said, “Once I get it back to New York, I’m going to need to get it fixed. The guy who did all the recent work obviously missed something. Or something really weird has broken. So,” he asked, “I’m asking you to think… if it was your car, how would you diagnose and fix this problem?”

It was, in fact, a great question. As I continued down to The Vintage (and as Brian and Michelle rented a U-Haul truck and trailer to tow the car home to New York), I thought about it. And the answer became clear: I own a nearly identical 2002tii. If necessary, I’d swap every ignition and fuel injection part from my car to his until I isolated the problem.

By the time I got down to The Vintage, word of Brian’s troubles had spread. Folks kept asking me, “Did you fix that guy’s car?” I’m sure that all car communities experience something similar, but the vintage BMW world is a very welcoming one, where we have newbie’s backs. It broke my heart again each time I retold the story. A few folks said, “You should’ve had him limp it down here. We would’ve collectively figured it out.” Maybe, but that wasn’t my call to make.

There was, though, a literal consolation prize. Every year at The Vintage, an award of which you really don’t want to be the recipient of is given—the “A Funny Thing Happened on the Way to The Vintage” Award. It was given to Brian and Michelle. I accepted it in their absence.

Sometime after I got home, I swapped email with another Vintage comrade. He asked if I’d read the thread on Brian’s car on bmw2002faq.com, the main 2002 forum. “Even before the trip,” he said, “it’s absolutely heartbreaking what this guy’s been going through trying to get this car to run right.” I dove into it. A number of shops had taken their shot but hadn’t produced a smooth-running trouble-free car. As I wrote a few weeks ago, there are realistic limits to what can be done with 40- to 50-year-old cars, but a competent shop should certainly be able to diagnose and repair whatever the problem de jour is.

I decided that the confluence of events—the camaraderie I felt toward someone who owned the same car, and his supportive spouse, that I realized I was in a unique position to figure it out and fix it, and that, as a writer and a mechanic, I had the power to change the ending of the story. It made me make him an absolutely ridiculous offer: If he brought the car to me up in Boston, I’d try to fix it for him for free (well, parts only). Not surprisingly, he jumped at it.

A few weeks later, Brian’s lovely agave 2002tii was trailered to my door. I immediately dove into it.

I’d developed a pretty methodical plan of attack. I first checked bedrock stuff like compression and valve adjustment, then swapped in my entire stem-to-stern ignition system. It made no difference. This effectively isolated the problem to being a fuel delivery issue.

Despite what Brian had told me about the tank being cleaned and the fuel pump being replaced, I checked everything. The tank was pretty good, but the fuel pump was clogged with sediment. I was surprised the car ran as well as it did. There is a metal fuel pressure canister next to the pump that was also rusty.

I replaced the fuel pump and blew everything out with compressed air until no rust showed up on a paper towel. It made no difference.

The metal-mesh filter screen on the injection pump, though, was troubling—it had a long gash in it. Could sediment have gotten into the injection pump and into the injectors? I didn’t know.

The Kugelfischer injection in a 2002tii is a mechanical system similar to the Bosch injection on old Mercedes and some old diesels. It’s a belt-driven pump that delivers fuel at very high pressure to spring-loaded injectors that then open up and squirt. The injection pump itself is a marvel of engineering that, as long as it’s fed clean oil, has a remarkably long service lifetime.

There are only a handful of people in the country qualified to rebuild them. I don’t pretend that I am even on the same planet as them. However, I needed to know if the injection pump itself was contaminated with rust and sediment, and I thought that the only way to know was to look inside. By carefully removing the pump head, I learned that a) fuel, whether contaminated or otherwise, has no way to get down into the bottom of the pump, 2) the fuel passages in the head all looked clean, and 3) the little spring-loaded plungers in the pump that push fuel toward the injectors all moved freely. This led me to not suspect the injection pump, and so, for now, I did not swap it with the one in my car.

So I continued moving methodically down the fuel path from the injection pump out toward the injectors, which are attached via translucent plastic lines. When the engine is running, you should be able to pinch each injection line between two fingers and feel it pulsing or buzzing, but it’s subtle, as there’s a lot of other noise and vibration around. I wasn’t certain that I could feel any pulsation on the number 4 injection line, so I looked at it closely, and saw something odd—what appeared to be air bubbles lazily floating along. None of the other three lines looked like this.

Hmmmmn.

In between the injection pump and each of the injection lines is a suction valve and a delivery valve. The 2002tii factory manual says that, if rough running is isolated to one cylinder, you should swap the suction valve and delivery valve with those from another cylinder and see if the problem remains with the same cylinder or moves to a different one. I did, and the air bubbles remained in injection line #4.

The only remaining thing to check was the #4 injector. These aren’t easy to get to, requiring removal of at least some of the intake plenums, but I yanked the one out of my car and swapped it into Brian’s car. This, I thought, has to be it. I turned the key, fired up the car, drove it down the block, and was gut-punched to find that it made no difference. I literally turned white and felt clammy.

I was about to swap the entire injection pump when I thought “Wait a minute,” and re-checked my work. It turned out that fuel line #4 was weeping fuel. I snugged it down, tried again, and the hesitation was gone. I video’d me driving the car like I stole it and sent it to Brian, with the note “Was this how it drove before the problem appeared? :^)”

So what exactly was the source of the problem? I pulled apart the bad injector, and received a double surprise. The first was that there was a fair amount of rust between the housing and the injector itself. But the second was that there were small pieces that fell out, and I couldn’t figure out how to reassemble it.

It took me a while to wrap my head around the fact that the spring that holds the injector seat closed had broken off at the top, leaving the seat just flapping in the breeze, and fuel just dribbling in without ever building up pressure. No wonder the thing ran like it did. Going back to what Brian had postulated in the parking lot, something really weird had broken.

I pulled the other three injectors out. None were broken, but all were rust-contaminated. Brian sourced a good used set of injectors. I sent all seven off to be cleaned and tested for uniform opening pressures, and the best four were selected and installed.

A few weeks later, Brian took the train up to Boston, we shared a lovely surf, turf, and martini dinner, and he thanked me profusely. He said he’d been on the verge of selling the car, thinking that it was unfixable, or at least not worth the trouble, and that I’d given it another chance. In the morning he headed back home in his now-correctly-running 2002tii.

I had done what I wanted to do as a mechanic and a writer—I had changed the ending to the story.

But there was still one final chapter: Brian and Michelle still hadn’t made it to The Vintage. That happened the following year when the event moved to Asheville. I drove to their home and accompanied them down from there. This included a run-down part of the Blue Ridge Parkway (much of which is sadly still closed due to Hurricane Helene in 2024). Waze sent us from there down into Asheville on this insanely switchbacked narrow road that made the BRP look like an interstate. My arms were tired by the time we arrived at the hotel. I said to Brian “I’m sorry, I didn’t do that intentionally, that was a bit much.”

“What, are you crazy?” he responded. “That was awesome!”

It seemed that nearly everyone at The Vintage had heard the story the year before of the tii that didn’t make it. This year, Brian and Michelle were constantly met with “Oh, you’re those guys! And this is that car!” And, to top it off, they got to pass along the “A Funny Thing Happened” award to the new winner—an E28 M5 owner whose car succumbed to some expensive-sounding valve train noises.

It’s been 10 years, Brian and I are still good friends, he still owns and loves the car, and I still have never met anyone else, pro or DIY, who has ever encountered a broken fuel injector in a 2002tii, much less felt their way through to diagnosing one. But as I said last week, as cars age, things that never used to break begin breaking. I suspect that I was out on the left edge of the bell curve of something that, moving forward, won’t be unicorn-rare.

But yeah… Most. Satisfying. Diagnosis. Ever.