Quadcammer

I don't care if it has 6 cylinders, or 16 cylinders, a properly built ICE should not burn a quart of oil every 600rpm.

The boxer engine tends to result in a little extra oil consumption compared to other engine styles, but the difference is no where near that big, especially since some 911s burn hardly any oil.

Aerokitted

i would sell my 911 if it burned anywhere close to the max amounts allowed. it's a joke.

tbrom

-

Wow, can't believe it - well I believe it but I can't fathom that kind of oil usage in a new car or broken in car. Just goes to show you and I just learned something about Mercedes.

Still that is rediculous oil consumption - tolerances that vary that much would make me think twice about build quality but they are still great cars. Very interesting. Thanks for sharing that info.

simsgw

Japanese parts vary in size. German parts vary in size. Spacecraft parts vary in size. It's too hot and muggy for an engineering seminar even if I felt motivated, but I'll try the quick drive-through version of an explanation. The bottom line, for those who don't care to read all this, is that oil consumption is a problem only if it indicates something bad is going on. The design engineers at Porsche have told us that no problem is indicated at consumption levels of a spit per lap... uh, a quart per 600 miles. More on that after we have some background.

Machinists, whether human or computer controlled, operate to very precise goals but they have to have goals. We engineers provide those goals. For example, we might say "32.795 centimeters". But if that's all we say, the machinist (actually a quality control review of our docs -- but in spirit, the machinist) will say "Yeah, but what are your specs?" We have to provide tolerances and bounds. So we might say "32.795 centimeters plus .005 cm minus .010 cm." That's nominal plus a high bound and a low bound. Manufacturing engineers translate that to production processes that achieve those goals in an acceptable percentage of the cases. Sometimes in space work, we throw out three parts to find the fourth one that fits our need. (You don't want a car built to engineering goals that push the manufacturing processes that hard.)

Typically for cars, I suppose the manufacturing engineers expect their processes to achieve a rejection rate no higher than one in 10,000. Mind, I'm just guessing because I was a design engineer and only had to 'appreciate' their work, not perform it myself; and honestly I never designed anything for a production volume higher than... I think it was 500 units and that was rare. Most of my designs contemplated a full life cycle quantity of less than a hundred. So treat the numbers in these examples as notional, not values to remember.

Let's consider why I can't just say 32.795 and take an early lunch. Well, at least a dozen factors give rise to variations in the properties of a part. By properties, I mean not only dimensions but alloy mix, surface hardening, toughness and so on. Tensile strength, dielectric constant, thermal coefficients, yada yada yada. I will have specified some of those and left others to Manufacturing to choose so they have the flexibility to meet my other goals. The trick is to know which specs are vital to my design goals and which I can leave to the engineers downstream so they have enough freedom to make their part of the job feasible. At the bleeding edge, we never hand off the design. We designers stay with it and have the other specialties in the room while we collectively create a widget that does the job. We design and build and redesign and rebuild until it works. You don't want a "bleeding edge" car either, but from what I hear that's what Formula 1 is like.

Let's get back to 'why' the parts vary and ignore how I'm trained to know which variation matters and which ones can be left to downstream work. Just think about one aspect. How big is the part? Well, that depends on its temperature and how much it expands and contracts with temp changes. Okay, tell the machinist a standard temp for his piece while he's working. Not yet. What are we going to use to machine the piece? Oh? What's that? You didn't want a million-dollar car built of parts machined from billet stock? Ouch. Then you want another whole set of questions answered. But we'll skip those and just acknowledge that casting and forging and autoclaving and stamping et al each have their own questions that must be answered. So back to machining examples. How hot is the piece of stock when the machinist begins? How much heat does it pick up as waste energy from the cutting/polishing/drilling steps? Do we heat treat this part or just let it work harden during manufacturing steps with other purposes? Pick one, now specify a cooling period and method to reach a 'nominal' temperature at which dimensions are to be compared to the goals I've set.

Whew. Now do all that with top quality production processes and measure ten parts. They will vary. (In all their properties, but we're just discussing dimensions.) More to the point, any two randomly joined parts will vary and they both may be on the high end of that dimension or the low end, or they may vary toward or away from each other. I said "specify a cooling period" but what if it's a particularly cold Winter and the enviro guys are keeping the plant five degrees cooler to save on heating bills? Or it's a hot Summer and they're keeping the plant hotter than usual? You want machining quality control to send the parts back to be reheated because they are cooling ten degrees more than I contemplated before they can get around to the measurements? Or you want a temperature-controlled bath before inspection? Not for products you can sell for consumer prices, like say less than a quarter million. What they do is ask a manufacturing engineer to estimate a temperature correction. So if Sims said "32.795 centimeters plus .005 cm minus .010 cm." at nominal plant temp of 80 F, then Manufacturing engineers have to interpret that and adjust for plant temp being 90 F. Or 60 F or whatever. Can they do that successfully? Well, usually. Sometimes they have to call the design team because the 'critical' dimension of the part might be responding oddly to temp changes. "Does that indicate a problem in operation? If we let the measurement reach 32.795 plus 0.20 cm because of the elevated plant temp, will the effect be linear in operational use? Or will the part expand too much and interfere with other components that might have come out of a cold plant in Finland last Winter? "What if we...?" "Can we..." "Should we..."

That is the process of getting a design from a drawing board (even a computer one using Autocad) to the end user. Then he goes out and buys an aftermarket paint because he likes the color and it's ten millimeters thicker and we get.... Well, never mind that part.

My point is that parts vary. Not only in dimensions, but in properties like thermal coefficients that affect their behavior in operation as much as the 'cold' dimensions. All that has to be controlled to whatever extent the design requires, but it can't be over-controlled or you end up trying to build an Apollo spacecraft for consumer prices. That ain't gonna happen.

When you design a car for performance goals well within the state of the art for manufacturing, the variations won't matter. Their effect won't be visible to the consumer, except that his Corolla may get an extra tenth mile per gallon and he won't notice that unless he's a complete obsessive and then he'll be sure its his own driving skill. When you design a Porsche, you have to be more cautious. Porsche uses more advanced manufacturing techniques (in the aspects that matter for this question) but even so, the original design team has to leave the manufacturing team room to maneuver if they are going to get the cars out the door in working order. You must have tolerances, not absolute goals. You know Manufacturing can't afford to do that measure and match process, "hand fitting" in other words. They need to know that this piston built to my specs will work with that cylinder block, machined to my specs. Okay, maybe a few critical parts can be measured and matched in a low-production model. Maybe the GT3. But you can't afford to build a Cayman that way or even a production Carrera engine that has to ship several thousand units each year.

So what I do instead -- that is, what the Porsche design engineers must do -- is limit how far we push the manufacturing processes. They won't have me hovering over the engine assembly saying "yeah, that piston's fine" or "nope, send that back to reman." I cost too much to hire 500 engineers to do that all day long at my pay scale. Lacking a designer to evaluate each part choice means my original specs have to be within manufacturing routine measures. They can be expensive measures, but they have to be routine, not requiring creative analysis at each step. Then when you get done thinking about all that, you ask yourself: How much oil will get by the rings if my smallest permissible piston is mated to my largest permissible cylinder? What will the dynamic sealing pressure look like at average urban rpm cycles? Track cycles? How about the ring gaps? Will they successfully rotate randomly if my largest piston is placed in my smallest cylinder? What happens if the gaps for ring two and ring one overlap more than X percent and it's a small piston in a large cylinder bore so the gap is at its maximum? How much oil gets by then?

Does that sound like I care about you excessively? No. So far none of those questions take into account your opinion of my design effort. Marketing does feed back crap that somebody has to listen to, but usually not me. Lead design engineers worry about whether the design works. The field engineers have to deal with service managers who have to deal with customers. They care about customer opinions so I don't have to. When the opinions get raucous enough, then design engineers hear about them, but they rarely care unless the reports indicate a design problem.

Does that sound the opposite? Do I seem insensitive to your concern about oil consumption? Well, possibly modern Germans take sensitivity training that I missed, so don't necessarily blame them. But for myself, the issue is whether it works, not whether you think my design works. What I really mean is this, just to take that one simple question: What is too much oil consumption?

Well, oil is a consumable. As spec'ed by the factory, a car whose oil level never dropped would still be consuming... what is it? Eight quarts every ten thousand miles? I don't feel like looking it up, but let's say we put in eight quarts at an oil change. Now suppose we ship a 'loose' engine that needs an additional quart every thousand miles? Well, the consumption just climbed to 17 quarts every ten thousand miles. (It will be one quart low when you start the oil change.) In that distance, the car is going through 500 gallons of premium gasoline and 17 quarts of oil. Does the extra nine quarts matter, except to give people something to discuss on forums? Of course not. Could I build an engine to minimize oil consumption? Certainly. But I have other goals for a Carrera engine. As long as the consumption level that results isn't a problem, I have bigger fish to fry. I'm not going to use up any of that flexibility that Manufacturing needs solving a problem that isn't a problem.

What does matter is oil consumption as an indication of a problem. As end-users we traditionally have used oil consumption as a naive measure of engine health. (I mean 'naive' in a non-insulting way here. It's just that an oil analysis is the only way to really learn anything from a car's oil habits. Consumption is too little information.) It is a traditional thing to speak of how little oil an engine uses as one sign of a good break-in and continued robust health. We say "the rings are good," or "the rings are shot" because we like or don't like the amount of oil that disappears between changes.

When the Porsche factory says "oil consumption of up to one quart per 600 miles is normal" they really mean that routine consumption up to that level is not a sign of an engine problem. Now if a car that routinely consumes only one quart between oil changes suddenly starts consuming a quart every thousand miles, then something has changed. Operational habits might be the answer, but you'd certainly have to inquire diligently. A sudden marked increase in oil consumption is an indication of a problem unless something has happened to explain it benignly.

I've muttered long enough. Let me get back to your second question now that we have the background. A German engine can probably be built to the same engineering goals and tolerances as a Japanese one. (I say probably because they haven't had as long to recover from World War II you see. A big chunk of their population spent half a century not recovering and not learning how to be productive workers. It was a laudable goal of West Germans to embrace their reunited East German cousins when the Wall fell. They went to great lengths to absorb them into their work force and it showed for a few years in the quality of German cars. For the sake of discussion, we can assume they have inculcated that new work force with an attitude and work habits that match the Japanese work force. Maybe not, but we can assume it.)

Given that you can, remember to compare equal designs. Pilots speak of "pushing the envelope" when they explore the corners of a new airplane's design, the combinations of airspeed and g-load and temperature and ... and as many different parameters as an engineer thinking about a piston design. In our mundane world of Autocad, we have a similar challenge. We can only "push the envelope" so far when designing something. The 'envelope' is the range of skills and processes and material suppliers that go into the manufacturing art.

If we push moderately, we get an economic, possibly fun, but definitely cheap car like the Civic or the... (Does Germany make a cheap car?) well, the Renault Ego or something. If we push a little harder, we get an Acura RSX or a Volkswagen Golf. If we want to push really hard to get a Carrera, then we have to leave wiggle room for the manufacturing engineers. We specify exotic materials (by their standards) and demand room for enough coolant to manage a small office, and in compensation... you know what's coming, don't you? To make up for pushing the envelope everywhere, we consciously do not sweat public relations questions that loom large in an economy model. Questions like why is this one using an extra nine quarts of oil between changes? Big deal.

An engine that produces low levels of specific power can be set up with pistons, rings and cylinder bore liners that are so mundane they rarely let a whimper of oil sneak past. An engine that produces 385 horsepower and is the size of a piece of luggage is pushing that envelope. Pushing hard. Built with routine measures, built without hand-fitting, it may well be quite frugal in its oil consumption, or it just as well may consume a tablespoon of oil (half an ounce) in a few miles at the track. Call it a teaspoon per lap.

That last value is what the consumption of a quart per 600 miles amounts to. One tablespoon in ten miles. As a design engineer, I can see a Civic buyer caring about the price of a teaspoon of oil. A Porsche buyer will not. The only issue for a Porsche buyer is whether that spoonful indicates a problem with the engine.

It does not. They already said.

Gary

mdrums

Gary...your the king of long posts...LOL...

simsgw

Yup.

Quadcammer

a great post, and then you go and ruin it at the end.

the price of a quart of oil is irrelevant. Obviously the cost of the excessive oil burning is not the issue.

Furthermore, you went through your whole manufacturing tolerance dissertation, and then say "well its fine cause they say its fine"

That line might be great if you are a service advisor.

But, a 3.6L making 385 bhp is nothing crazy. Whether its the size of a piece of luggage or not is irrelevant. The car uses hypereutectic pistons which don't change size as much as forged pistons. It also doesn't use boost or ridiculously high compression, so ring gaps don't have to be massive.

this engine is basically the porsche equivalent of a camry motor. Its run of the mill, plain jane, etc.

Finally, the 1.6quarts (not one quart) per 600 miles is a number reached by actuaries and lawyers. they calculated what percentage of motors would burn what, and determined what the warranty cost was, and arrived at this number.

Ask any decent engine builder or porsche engineer (outside the factory) how much oil a moderately stressed 3.6L flat six should consume, and I'll guarantee the number is much lower than porsches stated level.

I know porsche thinks its fine, and I think its absolutely pathetic that they do.

RED HORSE

Now take all the information in Gary's post and apply those concepts to manufacture of prescription medicines. They all apply. Some bottles of meds you receive fall into the "quart every 600 miles" and some are better. The only difference is you can't measure potency at home. Variability is a fact of life.

Quadcammer

thats all fine and well, but to go with your medical guidelines:

Thats like saying johnson & johnson has tolerances for tylenol of acetaminophen of 10% per pill, and the chinese knockoff has tolerances of 5% for the generic acetaminophen.

In other words, why do ford, gm, chrysler, etc specify one quart per 1000 miles, and Porsche, a brand renowned for its engineering has standards of one quart per 375 miles?

bmuhonen

From my 2010 Toyota 4Runner owners manual:

"Oil consumption: Max. 1.1 qt./600 miles (0.9 Imp. qt./600 miles, 1.0 L/ 1000 km)"

I get the impression that this may be somewhat of a ballpark industry standard.
BTW, the Toyota burns no oil, while my 2009 C2S is keeping Mobil oil profits at record highs.

RED HORSE

No idea how Ford, GM, etc. go about setting their oil specs. My point is many are obsessing over the amount of oil used when I suspect there are many other things going on with the vehicle you can't see, just like you can't see the potency of medications, that are just as important or maybe more important. What about the specs of the lug nuts/bolts for example, or better yet aftermarket wheels. You can't measure the differences with a dandy little gauge on the dash and you don't have to replenish anything, so you don't worry so much about it.

And the amount of oil consumed isn't a specification set by Porsche anyway, it is an indicator of performance.

Quadcammer

yeah, and when your car consumes 1 quart every 375 miles, the performance of your car suffers.

there will always be something else to worry about, but that doesn't mean we should ignore what is a shortfall.

simsgw

That's what we needed. An expert opinion. Wish I'd known. Could have saved my time.

Gary

simsgw

Works for me. Nobody is paying me, by reader or otherwise. But if you notice, I already compressed it: If that's still too long: This is not a 54 Mercury where oil consumption indicated a broken ring. Neither is it the "equivalent of a Camry engine." Oil consumption is just a characteristic, like the color of the engine block. It isn't a symptom of anything except hyperchondriacal owners.

Gary

Alan C.

Gary, having spent my career working with the automotive manufacturers as a supplier I can tell you that type of failure rate would get you thrown out of any plant. A 10 ppm failure rate will net a talk with QC.

On the main topic of oil consumption my 09 CS appears to be using a bit more than it had in the past. I've started to measure oil consumption beginning at 22000 miles. I filled it to the full mark and set a fresh quart of Mobil 1 on the shelf. At 22334, today, I added 12 oz. Not knowing how accurate the gauge is I wouldn't put a lot of faith in this first addition. Interestingly the engine doesn't use a quart on a 3 day track weekend.