Here is go again....I have a 2001 996TT (still under CPO warranty) eating 1qt ~450 mi and despite having all consumption documented through the dealer PCNA is giving me a hard time to fix the problem. They want me to do 3 more months of consumption tests (drive to the dealer every 500 miles to measure engine level) but apparently they believe that the car's oil gauge is not sensitive enough so before they pull a fast one on me, I want some ideas from people as to what is the BEST/MOST ACCURATE way to measure oil consumption. Below is what I have done so far:



1-waited for oil to drop to lowest level (1 bar left below arrow)

2-Dealer added 1 full qt (oil goes up 4 bars)

3-Return to dealer when level is at lowest level (1 bar left below arrow) and record mileage.


I guess a more accurate measurement may be to

1-add 12.5qts (correct full capacity? )

2- drive 500 miles

3-empty system and measure how much is left (of course some is stuck in the oil filter so dont know if this is perfect).


Other Ideas?

 

It's a dry sump.....difficult to get all of the oil out of the engine and turbos.....especially if you are looking for differences of less than 1qt.
AFAIK, you are stuck with the oil measurement via the OBC....unless PCNA has a dipstick.

 

it's not as good for the motor to drain it, and you still aren't sure
you drained it down to the same level each time.
The best way is to repeatedly record your mileage, fill, drive,
top off, record amount added and rerecord mileage. After 5 or 6
top-offs, you will have averaged away any inaccuracy in the
measuring system.
Joe

 

ltc - I think you are right in that I am stuck in relying on the OBC - how do you suggest I get rid off the positive bias of the system? I believe PCNA will fill it up and then ask me to come back after 500 miles to see how much is needed to be full again ( which equals consumption). The problem is that if say each bar is 200 ml. and I "consume" 5 bars in 500 miles, you can have 5 bars again by adding 850ml. The system will read 50ml as a full bar and then PCNA will conclude "car is within specs" 500/850ml = 588miles/ no problem. Thoughts?

 

I'm missing something here.
PCNA suspects that there is nothing wrong with your engine internally (ignoring the spark plugs for now), suspecting that the OBC oil level sender/electronics is not working properly/inaccurate....yet they want to use the same OBC oil level to monitor how much oil the engine uses?

Yikes!

 

You've got it!!! That is exactly what I told my Service Advisor which made him scratch his head and push back on PCNA (waiting to hear back on Monday). Point is that we need to isolate the variable so either the OBC system is very fina & accurate (and reliable for this type of test) or is not accurate/needs to be fixed OR a different measuring method is needed (drain & measure or something complex like that). And I thought Germans were rational beings (my guess is that their accountants arent german!)

Assuming I can convince PCNA that the OBC is working fine (as weird as that sounds) what would be your thoughts to ensure I dont get a false positive? That is if I run the car 500 miles and then come back to MEASURE OIL AMOUNT NEEDED to be full again (proposed method by PCNA), there is a risk of a false positive since it is a positive biased system (eg assuming 1 bar = 200 ml, 1000ml = 5 bars, 850ml = 5 bars).

 

As I said, no matter how inaccurate the computer measurement is, you can
completely make that insignificant over time. Yes, it will take more miles,
but that's all you have for a test mechanism. If you have 10 successive
top-offs, each after 250 miles, and you know what you added each time,
the consumption will be accurately calculable. If you are anywhere close
to 1 qt/450 miles, that would be a total of 5.5 qts, and there's no way any
inaccuracy of the car's measurement system will be able to mask the
problem.

 

Fair point Joe - I just have to believe that there is a better way to diagnose a problem than 6 months of measurements (but then again, maybe not!) In your opinion does it make any difference if you are adding oil "until it is FULL" and only then consider that a valid read of miles driven/oil added(suggested PCNA method) OR can I somehow validate the method used by the dealer (full oil, let it drop to the bottom, measure miles, add 1 qt, let it drop to the bottom, measure miles, add 1 quart). I just dont want to "waste" 3 months of reads!

 

Just as long as you know (or think you know) how much oil you have when
you start (the easiest is to add just enough that the car says it's full), and
how much you have when you're finished (also easiest to do what you did
at the start) you will have an accurate measurement. The more miles and
consumption you have between beginning and end, the more accurate you
will be. If you happen to start measuring when the oil is at some other level,
then just add oil at the end so it gets to the same place.
As long as you document what the system reads at the start, and how
much oil you added over a long enough time, and finish with the measurement
at the same place, you'll have an inarguable case. Otherwise PCA would have to
claim that the system gives random data at that fullness level. Distance and accurate
oil top-off records are the keys.

 

Thank you - Let the games begin....

 

Completely flawed logic. Please see me after class.

You are still using the "questionable" OBC oil measurement to define 'top off'.
IF the OBC is "questionable" for oil consumption/quantity per 'bar', then it is "questionable" for "topping off".
IF PCNA is stating the OBC is unreliable/inaccurate, then it can not be part of any experiment/validation.

This wouldn't stand up to any cross examination in a court of law, let alone any engineering discussion.

PCNA is playing a game......they are very good at it. Replacing the OBC, oil sender, engine costs $$....PCNA hates spending $$ unless 'convinced' to.

 

ltc - ANY suggestions? My plan is that I will ask them to put in writing that the method of testing (whatever that is but probably one of the options above) an which conditions will lead to which action (ie. OPEN engine for further diagnose). If they refuse to put it in writing I will have to bring in a lawyer to agree on a plan because I am loosing too much sleep over this.... I dont think it is too much to ask to have an agreed plan of action that both parties concur with -

 

Just tell them to fix the f***ing electronic dipstick then if they are so sure it's wrong. Then it will be right and measurement is possible.

 

first let me clarify language. The issue is precision not accuracy. you can not change the precision of the measurement device by increasing the sample size. What you can do with a larger sample size is get a better estimate of the mean (1/sq.rt. n). I believe this is what joe means. The second issue with the
The device is the discrimination (effective resolution). The device rounds
(as all devices do), the problem is we don't know
How it is rounding?
Multiple measures is your best option
Without some alternative.

 

Statman - agree. Problem is that if PCNA is either playing games (do this, try that) or plain stupid (do the measurement again with an instrument that we believe doesnt measure accurately), I wonder if the will even understand what is a mean....or rounding error. Here is another idea. I had considered the idea of going to another dealer but discarded the thought given they didnt do the original oil consumption test, dont know me etc. HOWEVER - given the large profit involved for them in fixing/replacing engine I wonder if the other dealer may be more aggressive in negotiating with PCNA. Long shot but I am running out of ideas and from all the responses (thanks to everyone) it seems that there is not ONE undisputable measurement method AND of course PCNA will set up the test in the most favorable & biased way for them.