Nugget

A few months ago after a few days of hot and bumpy driving out at Harris Hill Road we started getting a faint clattering noise coming from the engine bay. It was a subtle racket -- quiet enough to be drowned out by the lightweight flywheel chatter. But with the clutch pedal pressed in it was audible. Some described it as similar to the sound of an exhaust leak. To my ears it sounded sort of like that hollow rattle noise that a diesel engine makes. It was definitely louder coming from the left/driver rear wheel well, and less apparent from the right side of the car.

We got the car up on a lift and Warren at Rennsport Porsche WorkShop snooped around with a mechanic's stethoscope. He pinpointed the source of the noise to the vacuum pump. On our cars this is mounted on the trailing edge of the 1-3 cylinder bank, slaving off the cam of an oil pump. It's right where the shop light is pointing here, behind the exhaust valve:



Warren didn't seem too concerned about either the noise or the potential failure of the pump. He just sort of shrugged and said to call him when it finally gave up. The pump is used as a supplemental brake booster and isn't directly exposed to the oiling system or engine internals.

A few weeks later at a Driver's Edge event at MSR Houston the pump finally gave out. In the afternoon of the second day while I was driving I started getting really unusual ABS behavior under hard braking into a few of the turns. Not exactly like that ice mode ABS confusion you may be familiar with, but I found myself unable to threshold brake like I'm accustomed to and the brakes felt really wrong under heavy braking. About a lap later, the pedal got really mushy and pedal travel got much longer than it should be, but the car was still drivable. For the next session, equiraptor took the car out and experienced the same thing -- pedal feel was compromised but the car was still drivable at speed. After the track went cold as I was putting the car back onto the trailer I could hear what sounded like a vacuum leak hissing whenever I applied the brakes. Too quiet to hear over the din and fury of the track day, but quite distinct in the calm of the evening. Clearly the pump had finally failed.

The label on the pump itself was pretty worn down:



It looks like 997.105.016.71 but I was not confident with that last digit. I called up our friends at Suncoast and after they dug through the parts database they suggested that what I actually wanted was a 997.105.016.90 which, of course, had to ship from Germany. We figured the 71 and 90 just reflected an updated or revised part. We ordered that part (about $800) and waited for it to arrive. I got the car up on my lift and pried off the old pump, only to discover that the replacement part was not correct. The mating point for the pump cam was entirely different:



It turns out that the .71 pump is what goes on a 997.2 GT3. That engine has no scavenge pump driven by the cam and has variable exhaust cams hence the different drive system. Thwarted, I called Suncoast the following Monday and we all did more research.



The bad news is that you can't buy the vacuum pump for a 997.1. Instead you have to buy both the vacuum pump and the mated oil scavenge pump together as a single part number. What I had to buy was 997.107.022.91 (Tandem Pump Cyl 1-3) for about $1500. I received it yesterday and today equiraptor and I were able to successfully replace the failing unit.





The actual swap is quite simple if you can get the car up on a lift. Accessibility on ours is made a bit easier because we've removed the center exhaust (Sharkwerks bypass ftw) but even with the factory exhaust I doubt you'd have too much trouble. The vacuum pump is bolted onto the oil pump with three 8mm bolts, and there's a single press-fit vacuum line that attaches to the top of the pump.

The bottom bolt is bare and direct. Just an 8mm socket and you're all set. There are two 8mm bolts on the top of the pump (you'll sort of have to feel blind to reach them, but there's plenty of room to get your hands up there. Those bolts, in addition to holding the vacuum pump down, also secure a bracket around the vacuum line to prevent it from coming loose. Wriggle your 8mm socket up top and remove both of those bolts (there's a long one and a short one) and then you can detach the vacuum line just by twisting and tugging on it. Finally, there's one more bolt to remove. This will take a 10mm socket, and the bolt head is threaded. The shorter of the two 8mm bolts was actually screwed into the top of the 10mm bolt which is now the only thing holding the vacuum pump on to the car.




Once the fourth bolt is removed you can just pull the vacuum pump off. You may find you need to pry a bit with an appropriate tool or that big flat head screwdriver you always use as you tell yourself that you really shouldn't be using a screwdriver as a pry bar. Expect some oil seepage once you've pulled the pump off. I'd guess we lost about a 1/4 of a quart doing this swap, but if we had to do it again we'd lose a lot less since we'd know what we were doing.

Installing the new pump was the same process in reverse. I don't have a really good sense of how difficult it is to line up the cam shaft. I just sort of did my best to align the new pump as close to the rotation of the old pump by eye and it mated perfectly. I don't know if that means it's easy or if I just got really lucky. The new pump went on in just a few minutes without any drama at all.

We wiped down all the oil drippings, fired up the car, and went for a test drive. There was a lot of disconcerting oil smoke for the first minute or so of running, but it cleared up after a drive around the block. The brakes immediately felt normal again. It's frustrating that now I have an $800 oil scavenge pump sitting on the shelf in my garage that I'll probably never need, but it's super gratifying to have done the work myself for this repair.

I googled around and didn't find any other threads on this repair. There's a guy on sixspeedonline who retrofitted the tandem pump to his 996 Turbo as an upgrade, but his shop did the work and the thread wasn't very helpful for DIY. Hopefully this is of use to someone else someday. We spent about three hours total on the work. If your pump blows and you're near houston, bring it by. I think I could knock out the next one start to finish in about a beer and a half.

All the pictures online at https://www.flickr.com/photos/macnug...57676175854355

Special thanks to Ryan at Suncoast, Warren at Rennsport in Sealy, and James at Sharkwerks for their advice and internet-based hand-holding.

adi_d

When I saw a vacuum pump failure thread, I could not get myself not to add/respond!

I have a 997.2 GT3 and my failure occurred while on track. One of the bolts that holds the pump somehow came loose and overtime elongated the hole where the pump attaches to the camshaft.



I found the last 997.2 pump in stock through the dealer and ordered it. Here is a pic of the old pump in its entirety.



Since the bolt just flew off, the pump basically fell off during hot lapping and a good amount of oil started spilling on my exhaust. I suppose I was lucky that the oil did not catch on fire, or that the pump parts did not interfere with the camshaft. What was scary was the sudden loss of brakes. Once in paddock, I had to use my emergency brake to fully stop.



New pump is back on and working as designed. All good now. Don't mean to distract from your thread, but how many vacuum pump stories are really out there

Nugget

I wonder what the odds are that you just bought the pump that I was sold by mistake and then sent back to Suncoast a few weeks ago. When I bought it in late September there were none in the dealer network stateside at all. Thanks for the pics and the story.

rnh204

Interesting, guess this is sort of common. My 997.2 had the same issue at the track last year. Spewed oil on part of the track and pitted in immediately. Luckily my mechanics were there supporting another customer and got me patched up and I was able to drive home.

CosmosMpower

Good write up!

Nugget

I don't have any new insight, but I do have some new information and some photos to add. Tonight we learned that three pints of Guinness is the threshold between farting around on the internet and realizing that seeing the inside of a busted vacuum pump is truly the most fascinating thing in the house. I poured a fourth Guinness, grabbed my torx socket set, and we disassembled the old, broken vacuum pump to see if there was anything to be learned from its guts.

The first thing I noticed -- even before we took anything apart -- was the pump makes a distinct farting/belching noise if I spin it by hand. Neither of the new pumps I had in hand while we were doing the replacement made any kind of noise. They were just silent when spun. It seems likely that these chuffing bursts are the weird noise we heard coming from the car, just a lot faster when spun by the camshaft instead of my hand.

Going in, I had absolutely no idea what the inside of a vacuum pump would look like. Now we know. Four Torx bolts and we're in...



You might have never noticed, but the rotating shaft in the pump is not centered, it's offset. This, it turns out, is key to the operation of the pump. There's a flat steel plate that bisects the drive assembly, and pivoting end caps on the outer edges of the center plate. The plate slides freely through a channel on the main shaft, so it adjusts as it spins. This divides the interior cavity of the pump into two chambers, and as you spin the pump the chambers get alternately larger and smaller due to the offset rotation. It's a little bit like the inside of a rotary engine, I guess. Here's the innards in three places of the cycle:





I didn't notice anything obviously broken with the parts. It seems plausible that the channels where the center steel plate meets the end caps might get loose over time as the metal grinds/deforms, and I can imagine that the tolerances here are crucial to the operation. Sufficient oiling is almost certainly necessary, and if any of the oil channeling is occluded that might also impair operation. There was also some weird scoring on the top plate of the pump that is in a shape and direction I can't explain from the interior parts. I don't really understand how these marks would get made:



All told, the inside of a GT3 vacuum pump is a lot less complicated and more solid-state than I think I expected to see. This one makes a weird noise I can't explain. I'm not bold enough to disassemble a good one to compare, so I'm happy to leave it at that.

runet

A couple of comments, since I checked my own vacuum pump just a couple of weeks before your original write-up:

I think the pump is exposed to the oiling system as it is fed from the oil pump through a channel, and any metal grinded from a failing vacuum pump may end up in the oil and/or oilfilter.

When I rotated my vacuum pump by hand I had the belching noise as well, and everytime it made that noise some oil came out so I just took it for air pushing out the fluid. I find it very interesting that the new pump was dead silent, maybe mine is actually failing as well based on the belching noise. My car drives fine with the pump installed.

Can you put your old one back togehter easily, i.e. is it easy to check the internals and then use it again afterwards if it seems ok?
And does it still make the belching noise even when all of the oil is cleaned away from the pump? I might have to re-check it based on your findings.

Thank you.

modoz61

Nothing to add to the diagnostic but I had a similar failure earlier this year at Rennsport Australia.

The worst feeling when you see the white plume of smoke in the rear view mirror

We had just installed a rebuilt engine after a cam actuator failure (right bank) and this time as the vacuum pump sits immediately behind the cam actuator (left bank) I thought it was the same sort of failure

Car was towed from the track so I didnt get to do a close inspection but a nervous couple of days until it was diagnosed.

in the end quite relieved it was the the vacuum pump.

adi_d

Man totally true, except I did not even realize it could have been a catastrophic failure until the car was in the shop. At that stage it was a few days of "will it turn back on". Not fun!

Larry Cable

you sure that was the pump and not the Guinness?

flyingpenguin

Hate to revive threads, but doing so anyway :-)

Took the GT3 out of hibernation a few weeks ago (I am in Holland now) and noticed a slow warm-up (coolant and oil temps creeping up at same (slow) rate). This points to a defective coolant shut off valve on the tranny cooler. I had already replaced this one a few years ago, so before replacing, I checked its operation and as it turns out the lever does not fully extend. It is vacuum operated via a solenoid. Solenoid is OK, shut off valve also OK and there is a vacuum on the line. The vacuum is not strong enough though to fully extend the lever.

When I saw Nugget's great write-up I wondered if a refurb is possible instead of dumping $1,500 on a new tandem pump.

These pumps are made by Pierburg (OEM to Porsche) and are used on all sorts of cars from Jetta's to BMW's and Merc's. Charles the Humble Mechanic on youtube did a section on the issue:

So it appears to me that a simple o-ring is the culprit.

Nugget: Any chance you can ballpark the dimensions on that o-ring between the housing and the cover? That way I can just source a generic one before digging in.

Nugget

Heya! Sorry, I did get your email but I've been traveling and never got around to replying. Ik ben een bezig koe.

I opened my old pump back up and pried out the internal o-ring/gasket. It looks like it's 10cm, 2-3cm thick, and squared off (not toroidal).

Stig_997

My car is a 997.2 GT3. As I was changing the oil, I noticed some tiny shiny particles in the oil that looked like glitter. When I bought the car, I inspected the cam actuator bolts on passenger's side. Since they were all good and the driver's side was more involving, I decided to leave them alone for the time being. When I saw the shiny particles, I immediately thought it was the bolts at the driver's side. I had not noticed any noise or abnormal behavior with the car.

As I went inspect the driver's side, the vacuum pump needs to come out. Luckily, I mistakenly removed the vacuum pump cover first, instead of removing the whole pump and noticed the following:



I then removed the whole pump and inspected the bolts. They were all intact.

Looks like I found an issue that could become a much bigger problem just by chance, with no symptoms.

My pump is the 997.105.016.90.

Steven

RAudi Driver

Were there any other symptoms that you can remember. Motor running, did it make any unusual sounds? I can only imaging if you didn't need to change the oil what could have happened. Glad ya caught it.

Keep us posted.

Robocop305

So that rubbing is what caused the glitter in the oil? What was the fix? Thanks