After the trip to and from The Vintage and its multiple rest-stop repairs helping my new friend Mike keep his ’73 2002 Automatic running, any at-home automotive work seemed mundane, but I had a backlog of work on multiple cars to slog through. High up on the list was some strangeness with my 2003 E30 530i stick sport, the car that’s been my dutiful daily driver for seven years.
A few months ago the car developed an intermittent momentary hesitation. It wasn’t that it lost power; it was more that it felt like it momentarily wasn’t paying attention to the input to the throttle. Over the next few weeks, although the problem remained intermittent, it appeared to get worse, to the point where one night, when returning from a gig, the cruise control wouldn’t stay engaged. I wondered if it might be an issue with the drive-by-wire electronic throttle, either at the throttle itself or at the accelerator pedal.
I put the question to my Facebook friends, and the answers were all over the map. Some folks agreed that it sounded like a throttle-control issue, though others thought that such a thing would certainly throw a code that illuminated the Check Engine Light (CEL), which it didn’t. Some said it sounded like a fuel delivery issue, which I thought unlikely as I replaced the fuel pump two years ago. Others suspected the Mass Air Flow (MAF) sensor.
When in doubt, scan for codes, right? Like many other 20-year-old BMWs, the E39 did have an intermittent lighting of the CEL that, when scanned, revealed the common P0491 and 0492 secondary air injection codes. Truth be told, right or wrong, I regard this as The World’s Most Trivial Code (and no, it’s almost never due to the gas cap being loose), and unless there’s an actual symptom, I’ll clear it and go on with my life. If it’s weeks between reappearances, I’ll do this indefinitely. If it comes back quickly, I’ll eventually do something about it, as it may represent a sizeable vacuum leak.
I did both a standard OBDII scan and a BMW-specific scan on the E39 with my Foxwell 510NT and also with my ancient copy of BMW PA Soft, and in addition to the air injection codes and a few old misfires, they revealed a DME plausibility error in the accelerator and brake pedals. A-ha, I thought, so it is a drive-by-wire issue. I cleared the code and it didn’t come back. However, both the air injection codes and the the hesitation itself did.
Interesting.
Okay, then, air injection it is. The secondary air injection pump lives hanging upside down like a bat in front of the right front tire, . In this type of troubleshooting, I like to get to the heart of the matter and know whether or not the hardware itself is bad, so I pulled off the connector, exposed the power and ground pins, and ran jumper wires to a battery. The pump immediately whirred contentedly, indicating that the issue was likely in one of the vacuum lines connecting it to the car.
Two terminals, no waiting.
This was, as they say, déjà vu all over again, as I had this exact issue when I first bought the E39—the CEL was on, the code was for the air injection system, the pump itself was fine, and I needed to find and fix the vacuum leak to get the car inspected. At the time, I bought one of the $80 homemade smoke testers on eBay. It’s basically a paint can with a little 12V heater in it and a regulated pressure fitting to which you connect a compressor. Pour mineral oil in the can, close it up, connect it to the car’s battery to fire up the heater, let it warm up, and it blows a stream of smoke out a thin rubber hose with a plastic nipple on the end that you connect to one of the car’s vacuum lines. To find the leak, just follow the smoke. If necessary, stretch a latex glove over the mouth of the throttle body and over the exhaust to force all the smoke to come out the leak. Nowadays, $80 will buy you something that looks like an actual machine instead of something BillyBob was cobbling together as a side-hustle, but it was what I had, so I pressed it back into service.
Not kidding about the paint can thing.
Seven years ago, the process immediately pinpointed the leak as a rotted vacuum hose between the back of the head and the firewall—a place I never would’ve known to look. So you can imagine how pleased I was when, this time, I hauled out the smoker, connected it to the E39, and smoke wafted up from the same place. Seriously. I’m not being facetious. I don’t need every repair to pose a new challenge. I was perfectly happy to file this one under “Did it before, know how to do it now.”
Well, THIS sure looked familiar.
I looked between the back of the head and the firewall and immediately found that one of the hoses on the check valve, likely the one I hadn’t replaced seven years ago, had rotted away.
Gotcha.
I cut and installed a new length of vacuum hose and nearly raised the Mission Accomplished banner and began buttoning things up when I realized that the smart thing to do would be to smoke-test it again to make sure there wasn’t anything else.
And…
Oops.
This time, it took me much longer to pinpoint the source. I followed the hoses behind the head and under the intake manifold and found they went to a small vacuum box nestled under the back of the manifold. I replaced both hoses, one of which was very difficult to access, but the smoking leak persisted. There was no way to look directly at the box, but by jamming my phone behind and underneath the manifold, I was able to verify that assembly itself was the source of the smoke. The video can be seen here.
The smoking valve.
With some web-searching, I learned that the box is the vacuum control valve. The part (BMW part #11747810831, Pierburg part number 7.023180.10) is used in many different cars and is readily available, but that doesn’t mean it’s stocked at your local AutoZone. While I wanted to rip the leaky valve out of there and call that piece of the repair done, this was all happening a few days before I left for The Vintage, and there was the need to keep the car drivable so my son could use it for local errands while I was gone. I put a vacuum control valve on order so it would be in the mound of mail greeting me when I returned.
So this week, I had at it. The valve has an integrated plastic clip that attaches it to an oval-shaped slot under the intake manifold, so in theory, to remove it, you just need to bend back the clip and slide the valve toward you. However, as I said, you can’t see the valve, so you need to slide a screwdriver back there and bend the clip back by feel. Plus, with 21 years and over 200,000 miles on the car, even though the valve didn’t have metal-on-metal rust, things like this can still certainly stick. I eventually grabbed the valve’s side-facing electrical connector with needle-nose Vise Grips and pushed back on the plastic clip with the screwdriver so hard that, if I broke it off, good. It gave way.
Old and new vacuum control valves.
As I said, the vacuum hose connection to the front of the valve is difficult to make with the valve in place, so I was certain to connect the hose before installing it. With the new valve in place, the smoke test showed the leak was gone. Neither the code, the Check Engine Light, nor the hesitation have returned.
It wasn’t quite as satisfying as diagnosing that Mike’s points had closed up and helping him get down to The Vintage, but hey, daily drivers aren’t supposed to be exciting, right?
—Rob Siegel
____________________________________
Rob’s newest book, The Best of The Hack Mechanic, is available here on Amazon, as are his seven other books. Signed copies can be ordered directly from Rob here.
