An engine problem that “fixes itself” is a lie you shouldn’t buy into
I was driving the eight miles home along local roads from a doctor’s appointment in my 2003 BMW E39 530i, the car I’ve routinely described as the best daily-driver I’ve ever owned. It was beastly hot outside, by Boston standards—95 degrees and very humid. After a few miles, I noticed that the car was behaving sluggishly, barely responding when I tried to accelerate. Then it died at an intersection.
The car restarted and drove, but it still felt sluggish. I continued home but found that the car would only go a short distance before refusing to rev, losing all power, and forcing me to stop. Sitting for a few minutes seemed to revive it; the car started instantly and revved freely, but the power loss rapidly reappeared. This cycle repeated itself four times, with the power loss coming sooner each time.
There’s an old adage that if a car feels like it’s running out of gas, it probably is, at least at some level. That is, behavior like I describe above is more likely to be a fuel delivery problem than an ignition problem and could be due to a dying fuel pump, bad pressure regulator, clogged fuel filter, or some other related malady. On the vintage fuel-injected BMWs with which I’m most familiar, I’ve had instances where the fuel filter has gotten clogged from rust in the tank. Both the inlet screen of the fuel pump and the fuel filter can get blocked by rust as the fuel is pushed at high pressure through the system. Eventually the injectors get starved for fuel and the car won’t rev. But let the car sit for a bit and enough of the rust falls back off the filter and screen that the car will drive again for a short while before the problem repeats. A modern car like an E39 5 Series BMW has a plastic tank, so it’s not going to self-contaminate with rust, but you can certainly get a load of gas with sediment in it.
I really like this E39 530i, but it’s a woefully under-maintained car. I bought it five years ago for a song, and I’ve always been hesitant to drop money into it for the kind of prophylactic maintenance on the cooling system and fuel delivery system that I recommend on cars you need to make road-trip dependable. I rationalized this because a) when my wife and I go out of town, we drive her 40,000-mile 2014 Honda Fit, b) I work from home and have no commute, and c) if the 530i died while I was driving it locally, I could always just call AAA and have it towed back to my house. So, really, the only things that were surprising about the car’s death were that it took so long, and that it was a fuel delivery problem and not the kind of coolant loss from a cracked plastic expansion tank that modern BMWs are notorious for.
So, yes, I’d left myself wide-open for this, but in fact, any car can die and require a service call to AAA or Hagerty, even if it’s just a flat tire, and it’s never convenient. In this case, even though it was hot and humid, at least I was somewhere safe—on a local road with a shoulder, not in the breakdown lane of an interstate. Plus, the car could still be restarted and driven, although each time I tried to do so, it lost power sooner, and I still had a pretty congested part of Newton, Massachusetts, to drive through to get home. If it was evening or at night, I would’ve made a run for it, but it was three in the afternoon with a lot of traffic. I weighed it all and decided that the smart move was to just call AAA for the tow home before the car totally crapped out someplace unsafe or really inconvenient.
And yes, tow home, not to a repair shop, and certainly not to a dealer. Fuel delivery issues usually aren’t complicated. I’m not going to pay someone a thousand bucks when I can install a fuel pump and a filter and probably be done with it for a small fraction of that. And, on this car, the fuel pump drops into the top of the fuel tank from under the back seat. It doesn’t even require laying beneath the car and having gas run down your arm like fuel pump replacements on the vintage cars I’m used to.
Unfortunately, when I called AAA (Hagerty provides roadside service for my other cars), I was told my membership had expired, which was odd because I was looking right at my card and it showed it good through October 2021. Who knows; maybe AAA sent me the card for the coming year and then sent a bill that I’d forgotten to pay. I was told I could re-up on the phone, but there was a $75 surcharge for same-day road service. Further, the customer service rep warned that I’d need alternate transportation home, as due to COVID regulations still being in place, I was prevented from riding in the cab with the tow driver. Thus, I needed to call either my wife or an Uber. I was only five miles from home, so I called Maire Anne, explained the situation, and asked her to come with her AAA card (which, unlike mine, had not expired) so I could avoid the surcharge. She came with a cooler of cold drinks and said, “Me rescuing you! Just like old times!” (True.)
We made the service call on Maire Anne’s card, but a combination of high AAA call volume, an incorrectly-recorded address, and persnickety AAA policy on zone response (the car was technically in Boston, 1/4 mile from the Newton line) resulted in us still being there three hours later. By that time, traffic had abated enough that I felt comfortable trying to nurse the car back home. I fired it up, drove off, got about a mile and a half, and was not in the least surprised when the problem reappeared. The car’s power loss forced me to stop four times along the way and to drive up a hill at 5 mph, but at about 7 p.m., about four hours after the car first died, I pulled the hobbled 530i back into my driveway.
I thanked my wonderful wife, then posted the problem to my Facebook page and sought advice. A few friends thought it could be due to clogged catalytic converters, but most agreed that it seemed like a fuel delivery issue. Whether it was caused by sediment clogging the fuel pump inlet and filter, or by a combination of the pump being old (the car has 200K and the fuel pump is likely original), the low fuel level, and the high temperatures being insufficient to keep the pump cool, was unknown.
It’s always best to catch a fuel delivery problem in the act with a direct measurement of fuel pressure, but unfortunately, even though I have a standard fuel pressure gauge for cars with hose clamps on the fuel lines, and a C.I.S. injection test kit left over from my Porsche 911SC, I didn’t have a fuel pressure gauge with the Schrader valve on the end that’s needed to screw onto the test port on the E39’s fuel rail. So I ordered this $27 Actron gauge on Amazon. It’s something that I should have in the garage anyway.
I also looked online at the cost of fuel pumps and filters. Sheesh! Dealer list on the OE fuel pump assembly (which includes the plastic cage it sits in, as well as hoses), is $480, with discounted list around $280. There is a large variety, probably dozens, of aftermarket options for both the whole pump assembly as well as for just the a la carte electric pump motor, with prices as low as $35. However, not surprisingly, looking at the ratings of the inexpensive fuel pumps sold on Amazon, you can click on the tail of the curve (the one-star ratings) and see comments like, “My original pump lasted 18 years and 150K miles, and this one died after six months.”
[insert pic: genuine fuel pump. Caption: The OE-style fuel pump assembly in its plastic cage and with the associated hose and seal. (photo FCP Euro)]
The fuel filters are also pricey. As they include a pressure regular, they’re not like the simple paper-in-clear-plastic filters that vintage cars use. The dealer list is $160, discounted list is about $100, name-brand OEM units are as low as $50, generic Chinese-made ones as low as $22.
At some point during the car’s hissy fit, the Check Engine Light (CEL) had gone on, so the following morning, I hooked up a code reader and pulled the codes. I first used my old Actron generic OBD-II code reader, but all it read was a misfire code on cylinder #1. I could believe that the car misfired during its episode, but it was vanishingly unlikely that a misfire on one cylinder was the root cause.
So, I connected an old Windows XP laptop that I have configured with some aftermarket BMW scanner software. It pulled up a hail of codes, mostly old and uncorrelated with the symptoms. I didn’t see anything like a code from a throttle position sensor or air-mass sensor that might explain the car’s unwillingness to rev. In my mind, all of this reinforced that there was a fuel delivery issue, as something like low suction or a clogged filter doesn’t throw a code.
I then cleared the codes and drove the car a few miles. The symptoms didn’t reappear, and the CEL light remained out. However, the fuel warning light came on, so I really was low on gas.
But I was now faced with an interesting choice: I could wait until the air temperature heated up in the afternoon (it was forecast to be another hot day) and see if, with both low fuel and high temperatures, I could re-create the problem, but doing so risked actually running out of gas. I elected to split the difference. I put five gallons of gas in the 18.5-gallon tank. I drove the car several more miles completely without incident.
At about 6 p.m. that evening, when it was still over 90 degrees outside, I took the car out and hammered on it. I put 20 miles on it on a combination of local roads and highway, redlining it several times. Nothing. It was as if the incident had never happened.
The next morning, the Actron CP7818 fuel pressure gauge arrived from Amazon. If you look at the photo in the ad, the gauge doesn’t directly screw to the Schrader valve on the fuel rail; it screws to an adapter, which screws onto the Schrader valve on the fuel rail. Except it turns out that it doesn’t; the valve on the fuel rail is short and has a circular step at the bottom, and the bottom of the adapter hits the step before it depresses the valve. I had to grind a millimeter off the bottom of the adapter. But then it worked, and the gauge registered the correct reading of about 50 psi.
So, what do you do?
I’ve long pointed out the obvious fact that mechanical systems do not naturally heal themselves like biological systems. If a car exhibits a problem, it is almost certain to do it again, and then get worse. If you think “Maybe it’ll just go away,” you’re at best in denial, at worst an idiot, and in either case, you’ll get what you deserve.
The flip side of this is that I always bristle when someone describes a car dying and, even after it’s repaired says, “I never trusted it after that, so I sold it.” Feh. If you don’t like a car, then you sell it. If your needs change, then you sell it. If you want newer safety features, then you sell it. This vague “never trusted it” thing isn’t a real reason. Cars are electromechanical systems. With rare exceptions, problems can be diagnosed and solved. This wasn’t even a particularly difficult one.
If I was keeping the car for the long term, I’d swallow hard and drop $280 on an OE pump and $50 on an OEM filter, but I’m hesitant to do that (the car’s probably only worth about $1500) unless I’m certain it’ll solve the problem. If I was going to sell the car instead, I could do nothing, or spend $35 on an aftermarket pump assembly and $22 on an aftermarket filter and be able to say, “It’s got a new fuel pump and filter” and stifle the urge to blurt out, “Yeah, but they’re the cheapest garbage available,” but that’s really not my style.
When I road-trip vintage cars, I’ll often bring spare parts with me. So, yes, I could buy a cheap fuel pump to have with me just in case the one in the car croaks altogether. But if you suspect a part is going bad, it really behooves you to replace it in the comfort and safety of your garage rather than risk having it fail and require a roadside repair, so this approach seems silly.
I could do nothing and simply wait for it to malfunction again. But I just had the thing malfunction on me and be bad enough that I didn’t think I could safely make it home. Plus, this is ostensibly my daily driver, which, almost by definition, means it’s the car you don’t drive with a toolbox in the trunk.
Nevertheless, for now, since I don’t have a commute, I have other cars to drive, I don’t have to rely on the car, and my wife and I have a much newer and more reliable car for when we need to leave town, the engineer/mechanic in me really wants to catch the failure mode in the act. I think I’ll put the fuel pressure gauge, rubber gloves, and paper towels in the trunk, use the car, wait for it to re-exhibit the problem, and see if I can verify whether or not, when it occurs, the fuel pressure is low.
Or, maybe the problem will just go away …
(But seriously, I’ll let you know what happens.)
Rob Siegel’s new book, The Best Of The Hack MechanicTM: 35 years of hacks, kluges, and assorted automotive mayhem is available on Amazon here. His other seven books are also available on Amazon, or you can order personally-inscribed copies from Rob’s website, www.robsiegel.com.