X35d CBS-Indicated Oil Change at 9000 Miles

[FunfDreisig]

Originally Posted by Penguin

....One thing I can say is that changing the oil on my X5 had zero impact on the "miles to oil change" shown on the i-drive. After I changed the oil, the "miles to oil change" was unchanged, and after driving 1,000 miles it was 1,000 miles closer.....

Exactly!

IIRC none of us, that have done an intermediate oil change, have reported a significant increase/reduction in the "miles to oil change" reported by the iDrive afterward. And the eventual CBS Oil Services requested by 35Ds owned by 'compulsive oil/filter changers' are well within the range of those whose owners think their BMW knows best

FWIW speculating about how a black box works can be entertaining. But actually running enough physical experiments to deduce the full workings is hard work. Particularly if the black box includes multiple logic branches and overrides. So unless some one can turn up a document describing BMW's CBS oil change black box in some detail, we will be left with our impressions based on largely conjecture and anecdotal evidence.

Funf Dreisg

[JCL]

Originally Posted by FunfDreisig

Exactly!

IIRC none of us, that have done an intermediate oil change, have reported a significant increase/reduction in the "miles to oil change" reported by the iDrive afterward. And the eventual CBS Oil Services requested by 35Ds owned by 'compulsive oil/filter changers' are well within the range of those whose owners think their BMW knows best

FWIW speculating about how a black box works can be entertaining. But actually running enough physical experiments to deduce the full workings is hard work. Particularly if the black box includes multiple logic branches and overrides. So unless some one can turn up a document describing BMW's CBS oil change black box in some detail, we will be left with our impressions based on largely conjecture and anecdotal evidence.

Funf Dreisg

Well stated. Especially the part about it being entertaining to speculate.

But one comment on your own experiments to date: Those who are changing the oil early are not using the monitoring system as designed. (This is not a value judgement, simply an observation) That is hardly a realistic test of the system, it is rather a type of failure test. It could be that since the fluid changed properties substantially in a matter of moments when the oil was changed, that the algorithm decided to ignore further readings, assuming a sensor failure or out of range condition. A more realistic test might be to see if oil change intervals called for by the CBS vary significantly between different vehicles over the full cycle, with the original oil, and by more than any logged change in vehicle mileage (which BMW has stated in an SIB is the initial forecast variable before it starts counting down).

[ard]

Originally Posted by JCL

Why did you add the word "discrete" to the Continental description? (Bold added by me for emphasis) It may be, but they never said it was discrete, they just provided the operating range and expecting typical oil quality range.

Good point.

Their assignment of some "oil quality metric" to a range of "1 to 6" seems arbitrary and limited. Replace "discrete' with "arbitrary, limited, undefined and non-validated"

Originally Posted by JCL

The accuracy of the same sensor for fluid level measurement is noted to be around +/- 2.5%, so why would you expect the quality sensor to be so different?

In my professional experience when a spec is vague and imprecise it is because it is bad. This is a permittivity sensor- NOT really an 'oil quality sensor'. They are likely related, but I expect the operating curve for the permittivity really reflecting oil quality will be like measuring battery voltage to tell you your battery is bad. By the time the oil has enough water/coolant/metal in it to create a rise in the value, you are harming the engine.

Originally Posted by JCL

Purely anecdotal, but your thread doesn't appear to support your position that the condition based monitoring is completely ignoring the inputs from sensors such as these.

True. I assume if it is not mentioned it is not used to drive CBS calculations, you assume it is. We'll never know.

Originally Posted by JCL

Thanks for the Continental catalogue link, it is interesting reading. I would like to find out more about the different sensors used, as RealOEM list different part numbers for the diesel and the gasoline engines on the E70. It may be due to different oil pan depths, but it may also relate to the ability of the sensor (as claimed by Continental) to measure soot levels in the oil, something that would be quite relevant to the diesel model.

IMO all that stuff (soot+fuel+water+coolant+metal) is not resolved as discrete contaminants- all of it affects the single value from the sensor.

Originally Posted by JCL

I hope that we have now reached agreement that there is in fact an oil quality sensor (even though we don't know who makes it).

Well, a permittivity sensor.

Originally Posted by JCL

I also hope we can agree that it is able to provide inputs to the CBS, which is primarily based on fuel mileage/consumption, but which is likely to consider multiple other inputs.

No, I cannot see the oil quality sensor input as an input they use to decrement the mileage. My belief is that it is a catastrophic oil failure sensor, given the inherent imprecision of the sensor.... and the fact that interstitial oil changes have no effect on CBS.

Originally Posted by JCL

I actually think that BMW doesn't want us to know much about this technology, as the best information is coming from the sensor manufacturers, not the vehicle manufacturers. BMW won't likely have a lock on this device, and so there is little reason to promote it, if you can get the same sensor in BMW's competitor's vehicles. The proprietary part of the system will be the algorithms that the manufacturers use, and they will likely keep those very much to themselves.

Interesting topic.

For sure agree that BMW wants us to think things are a lot more advanced than they really are...

[FunfDreisig]

Originally Posted by JCL

.....It could be that since the fluid changed properties substantially in a matter of moments when the oil was changed, that the algorithm decided to ignore further readings, assuming a sensor failure or out of range condition.....

So either....
1 - The CBS Oil Service algorithm is lazy and hard headed. It decides pretty early on when it will request a CBS Oil service and doesn't worry its pretty little head with oil condition facts from then on
or
2 - The CBS Oil Service algorithm is so sophisticated that it takes into account that some wacky BMW owner's are compulsive oil/filter changers and requests a CBS Oil Service anyway based on unrelated maintenance items that are part of an 'oil service': e.g. air filters, DEF flush & fill, etc.

Funf Dreisig

[JCL]

Originally Posted by FunfDreisig

So either....
1 - The CBS Oil Service algorithm is lazy and hard headed. It decides pretty early on when it will request a CBS Oil service and doesn't worry its pretty little head with oil condition facts from then on
or
2 - The CBS Oil Service algorithm is so sophisticated that it takes into account that some wacky BMW owner's are compulsive oil/filter changers and requests a CBS Oil Service anyway based on unrelated maintenance items that are part of an 'oil service': e.g. air filters, DEF flush & fill, etc.

Funf Dreisig

Well, we know that (1) isn't true, since you and others are being asked to change the oil at intervals that aren't directionally consistent with your fuel usage over time, ie shorter when you expect longer or vice versa. Something is driving it other than fuel mileage and distance. They just don't want to tell you, for the same reasons they don't want you to have actual temperature readouts. Most drivers don't know what to do with them.

With respect to (2), I think they are hanging the intervals on oil changes and just picking up the other items when it is in (apart from the time-based elements like brake fluid).

All good fun.

[JCL]

Originally Posted by ard

Their assignment of some "oil quality metric" to a range of "1 to 6" seems arbitrary and limited. Replace "discrete' with "arbitrary, limited, undefined and non-validated"

In my professional experience when a spec is vague and imprecise it is because it is bad. This is a permittivity sensor- NOT really an 'oil quality sensor'. They are likely related, but I expect the operating curve for the permittivity really reflecting oil quality will be like measuring battery voltage to tell you your battery is bad. By the time the oil has enough water/coolant/metal in it to create a rise in the value, you are harming the engine.

Applying the same logic to the rest of the vehicle, perhaps we should stop using our BMW back up sensors and automatic rain sensor wipers, since the specs are not published and they seem to use arbitrary and non-validated steps in their algorithms. Or we could just notice that they seem to work and accept that though we didn't design them, and probably couldn't fix the internals of them, they function just fine.

In my professional experience, when multiple companies have been producing and supplying a similar design of sensor for up to 14 years, and those sensors have had wide global market acceptance, I tend to think they have moved beyond witchcraft. Maybe I'm just too trusting.

[Penguin]

Originally Posted by JCL

It could be that since the fluid changed properties substantially in a matter of moments when the oil was changed, that the algorithm decided to ignore further readings, assuming a sensor failure or out of range condition.

I think you're stretching a bit here.

It is hard for me the believe that the computer would be smart enough to recognize a "sensor failure" reading in computing the CBS and, at the same time, dumb enough not to report it.

[JCL]

Originally Posted by ard

I assume if it is not mentioned it is not used to drive CBS calculations, you assume it is. We'll never know.

...I cannot see the oil quality sensor input as an input they use to decrement the mileage. My belief is that it is a catastrophic oil failure sensor, given the inherent imprecision of the sensor.... and the fact that interstitial oil changes have no effect on CBS.

The SIB you reference says that the initial forecast of service interval remaining is primarily based on fuel consumed. They are discussing the forecast number, ie when you get an oil change and the system is reset it won't now be automatically showing 15,000 miles, but rather a distance related to your consumption. It says that this is only a forecast, and that it will drop more quickly once it reaches 50%. We can conclude from this that since the forecast is primarily based on fuel consumed, and not completely based on it, that there are other factors. We can also wonder what causes the oil change interval (the actual, not the initial forecast) to start moving more quickly once 50% is up. It isn't fuel consumption, it is something additional. Maybe even sensor data, either from the QLT or other sensors. Your link suggests that until one reaches the 50%, the forecast will not consider those other inputs. That sounds to me a little like a damped temperature gauge, which reads in the centre most of the time.

What makes this interesting is that several 35d owners have anecdotally noted that oil change requests from the CBS system don't seem to be directly related to fuel consumed or how hard the engine was working. It will be good to learn in the future what in fact does drive it.

[Penguin]

Well, I think we can be pretty sure that the sensor is Continental's QLT sensor, as this presentation says BMW is a customer for the sensor on Page 47:

http://www.google.com/url?sa=t&sourc...hJaYvN1XEw5pMA

So, to summarize:

(1) The BMW sensor looks like the Continental sensor,

(2) BMW used the same "QLT" nomenclature in technical documents that Continental uses for the sensor, and

(3) Continental says BMW buys the QLT sensors from them.

OK, it's not proof, but pretty darned convincing!

[JCL]

Originally Posted by Penguin

I think you're stretching a bit here.

It is hard for me the believe that the computer would be smart enough to recognize a "sensor failure" reading in computing the CBS and, at the same time, dumb enough not to report it.

Fair enough. I figure that just about everything after post #5 was a stretch, albeit an intriguing one.