No it doesn't. On this point I can categorically and respectfully state that you're wrong. Many people use different computers to view this forum (home, work, mobile, etc) and they can also view it as being signed in or not. Throw in a few who viewed and chose not to vote, and your calculation doesn't hold water.

You are absolutely correct and I knew it the second I pressed the submit button

I was just too lazy to go back and re-edit my post so no need to be respectful at all

I suppose that VAG does actually have the data to decide this question. Cars in for a service have their computer records read. VW must surely be able to draw some inference from all this information - if they put their mind to it.

That said, I don't say that just because VW recommends a service interval for this or that style of driving that that recomended service interval is likely to prolong the life of a turbo. Without saying all large corporations are evil, VW's recomendations on servicing will always involve more considerations than just turbo life.

Another point, could one save enough by keeping to variable servicing to ultimately pay for a turbo that fails "early"?

Even if it was 1300 separate users, out of the number of 2.0TDI sold worldwide, the survey size is statistically insignificant.

If the survey size gets less, then it's even less significant.

From this survey you can draw no serious conclusion and this goes even more so when you consider all the factors missed out.

If you assume that you can get maximum miles out of a variable service, then you're paying for 2 services for every 3 normal ones.

So £200+ every 3 years saved.

It's touch and go if you'd pay for a turbo by this, but there are many other factors and there is no way this survey can show that variable servicing kills turbo's or that non variable is any better.

Plenty of Fabia vRS killed their turbo's and every single on of these was on fixed servicing as variable was not an option.

Edited 21 March, 201016 yr by cheezemonkhai

Even if it was 1300 separate users, out of the number of 2.0TDI sold worldwide, the survey size is statistically insignificant.

Not correct. A sample size of 1300 would be more than twice what is required to make the survey "statisticall significant". And it doesn't matter whether there are 1000 or a billion 2.0Tdi s sold worldwide.

You're making a common mistake in assuming that a large sample size is required to give a statistically significant result. I will try to explain.

Governments/newspapers/market researchers/official bodies all use a sample size of much less than 1300 to assess whole populations of countries (counted in the millions), and these results are considered statistically significant.

For those who haven't had the unfortunate and dull experience of studying statistics, here are a couple of "sample size" calculators:

http://www.raosoft.com/samplesize.html

http://www.surveysystem.com/sscalc.htm

Basically you decide what "confidence level" you want (typically 95% or 99%), and what "margin of error" you want (typically 5%). Then input the population, or number of cars, or whatever you're assessing.

Worked example, assuming 99% confidence level with 5% margine of error:

Assuming 100 Octavias in the world, you would need to sample 88 people to achieve statistical significance

Assuming 1000 Octavias in the world, you would need to sample 400 people to achieve statistical significance

Assuming 10,000 Octavias in the world, you would need to sample 623 people to achieve statistical significance

Assuming 100,000 Octavias in the world, you would need to sample 660 people to achieve statistical significance

Assuming 1 million Octavias in the world, you would need to sample 664 people to achieve statistical significance

Assuming 10 million Octavias in the world, you would need to sample 666 people to achieve statistical significance

Assuming 100 million Octavias in the world, you would need to sample 666 people to achieve statistical significance

Assuming 1 billion! Octavias in the world, you would need to sample 666 people to achieve statistical significance

In summary, if there were 100,000 Octavias in the world, 660 people is enough to be statisically significant. Even increasing the number of Octavias to a billion only increases the required sample size by 6 people, to 666 people. Once you go past a certain size of population (be that people, cars, smokers, homeowners, left handers, men with one testicle, whatever), you don't need a larger sample size to make it stastically significant.

Apologies for the long winded post, but I didn't want to just say, "sorry you're wrong", without some explanation. Now I can have breakfast.

Try google scientific rigour while you're up then

I did reply properly twice, but the forum keeps losing posts.

Suffice to say there are many more cars on variable than not and that if 1 in 10'000 turbo's die for no reason, there would be more on variable than fixed for that reason alone.

Also since there are twice as many people showing variable and no failure, I'd say that any conclusions reached are not realistic.

That plus the original questions were heavily biased. I won't go into all the factors not taken into account, but as before the survey is still meaningless for the basis of any action or summation of things.

Will the long life oil play a part here? it takes longer to break down obviously, Im assuming long life is used for variables and and not fixed? or am I wrong?

Will the long life oil play a part here? it takes longer to break down obviously, Im assuming long life is used for variables and and not fixed? or am I wrong?

Good question. I remember reading on here that many dealers use exactly the same oil for both fixed and variable, but am sure someone will confirm this.

Good question. I remember reading on here that many dealers use exactly the same oil for both fixed and variable, but am sure someone will confirm this.

It depends if you have a DPF.

If you have a DPF the oil is VW507.0 regardless of if you are on fixed or variable.

If you have no DPF, then the oil is

- VW507 or VW506.1 for variable. The latter is a 0w30 and more expensive by a long way per L, so I can't see any dealer using that over VW507 which "replaced" it.

- VW505.1 or VW506.1/VW507 for fixed servicing.

I can see many dealers using VW507 across the range, because it's usually also a VW504 oil. That means one oil is suitable for almost all petrol and diesel cars at the dealers.

Like I said, I think it's more down to quality control of the turbo rather than oil lubricating it.

Problem is nobody is going to spend money doing a failure analysis on a dead turbo when it can cost a lot of money, so I doubt we'll really find out.

If the idea is to verify whether fixed versus variable servicing is responsible for more or less turbo failures then i guess that the failure mode

of the turbo surely needs to be ascertained.

There are failure modes I am sure which are not related to whichever service regime you are on so these would need to be discounted.

I would think it more relevant and interesting to see at what mileage turbo failures occurred.

My own turbo developed an oil leak and was condemned for that reason. I had it changed at 105,000 miles.

The turbo itself was functioning properly with no noise, smoke or whatever and it was only spotted at

service time.

I was on variable servicing and covered some 30,000 miles a year on a 70mile each way commute.

Would a fixed service regime have made it last much longer? Maybe/Maybe not. Who could know.

I would guess though that a different driver doing 105,000 miles at 10k a year round town on a fixed service regime over 10 years

may have experienced a failure at a lower mileage.

There are so many factors at play here I'm sure that it would be difficult to conclusively establish that the variable regime was

a poor idea. After all, if changing oil at 10k is better than 20k then why not change it at 5k or 2k?

Depends on your priorities I suppose, how long you plan to keep the car, how much you care, how much time and money you've got

and whether your dad owns an oil well!!

Try google scientific rigour while you're up then

I did reply properly twice, but the forum keeps losing posts.

Interesting ! It is probably a very good thing for the majority of you good people on Briskoda ,but just recently I have notice that when posting sometimes the screen goes back to the main front page of Briskoda ( which I do not usually use ) and the post if lost . Anyone else suffer this problem ?

Steve W

Mmmm worried now....had the turbo replaced at 14k due a horrible squeeling noise coming from it. Now at 21k and noticed the noise is back for about a minute or so when the engine's cold. Hope this one's not on it's way out too. Bill for the last one was over £1700 !!!