Technical Paper from Mahle explaining their Motorsport bearings for S65/85

[PaulGros]

Originally Posted by Green-Eggs

Those measurements were in the process of being updated with his new air gauge when Bert got side tracked and building a new house. All measurements up to the BE shells were probably done on a ball-anvil micrometer, which may not be the most accurate way to measure. That's why he was in the process of redoing them. The air gauge rig is supposedly accurate to 0.000001 inch. I know he measured the Mahle Motorsport shells with it, and sent the results to either you or Alan, maybe both. So, it would be interesting to see how your measurements compare with those, and at least the BE Shells on the wiki. Would also like to know your measurement methodology and equipment if possible. More data, is better.

While we're at it, maybe M Power Motorsport will share his measurement accuracy, variance, resolution, stepper motor resolution, cosine error (which is probably 0, but nice to know). I'm really surprised nobody is asking these questions about that rig.

I didn't receive a copy of the MAHLE measurements, maybe Allen has them. For that level of accuracy you will also need to be measuring in a controlled environment - this point was made to Bert when we met.

Rather than introduce new variable measurements I've used all of the data from the BE Wiki, even for the OE & MAHLE parts. I did this to try and organise the data for my own understanding, but it ended up creating more questions. The most surprising was the variability of the wall thickness of the Clevite manufactured parts. The other is that average wall thickness on all parts are remarkably close.

Maybe get Bert to drop me a mail as there are other inconsistencies / inaccuracies in the information.

[Green-Eggs]

Originally Posted by PaulGros

S65 bearings is taking over my life

If you take the average wall thickness for each bearing in the WIKI, you will see either the data is wrong or if it is correct all bearings are surprisingly very close except the HX ACL & the VAC Clevite. This is no surprise as they are 0.025mm additional clearance.

The Wiki states the VAC parts are standard clearance, the picture states other wise - they are clearly marked HX and the measurements support this.

VAC buys the shells as HX without a coating. VAC sends the uncoated shells to Calico for coating, which, according to Calico, add between 0.0002 to 0.0004 inch on average. Our experience with Calico is that it adds minimum 0.0004 inch, and can go as thick as 0.0015 inch. When Clevite aftermarket provides coating, like BE V1 shells, they are marked HK instead. There is an entire set of facebook posts by VAC themselves where they discuss their bearing design target was OEM clearance. "Click the link on the wiki page at "History of VAC Bearings: In VAC's own words."

All of the measurements at the wiki will eventually be redone. The shell thicknesses will be redone with the air gauge, and the clearances will be redone with a custom LVDT bore gauge accurate to 0.00005 inch. Those custom gauges are already done and ready to use. Preliminary tests show that the measurements with the custom gauges are 99% repeatable. The clearance measurements are performed with 34-points around the circle of each connecting rod/bearing assembly. Time availability is the only limiting factor at this point.

[Green-Eggs]

Originally Posted by PaulGros

I didn't receive a copy of the MAHLE measurements, maybe Allen has them. For that level of accuracy you will also need to be measuring in a controlled environment - this point was made to Bert when we met.

I just realized that the Mahle Motorsport measurements are the ones posted at the wiki page.

All measurements, old and new, were performed within 2-degrees F of each other, and within about 5% relative humidity of each other. This data is logged using a Kestrel gauge. True, it's not a scientifically controlled environment, but it's consistent across all sets, old and new.

[PaulGros]

Originally Posted by Helmsman

Haven't looked into the BE page for long (and it has been gone for a while), but you mean wall thickness between the shells are pretty much the same incl BE? (again, except for the extended)

If you take an average of the min / max measured figure for the top & bottom shells combined then yes. The OE parts (if an accurate measurement) there is one out of tolerance. For ACL also based on the tolerance you were given by them.

The other anomaly is the assumption that the OE parts always use 1 of each colour code. Yes, this is the way it is presented in the BMW OEM catalogue, but it makes no sense to have graded parts and always to fit the same parts regardless of tolerance on the crank & rods. This could be achieved with a single bearing as both shells are the same design and material.

To me this suggests that a combination of red / red, red / blue or blue / blue were used depending on the crank & housing size inline with standard OEM grading schemes. Unless a number of used sets with the colour codes still visible can be found and these documented against crank & housing size this will have to remain a reasonable assumption. And yes, it is not unknown for BMW OEM catalogue to have an error. I know of one vehicle where it states only 7 main bearing bolts are required

[PaulGros]

Originally Posted by Green-Eggs

VAC buys the shells as HX without a coating. VAC sends the uncoated shells to Calico for coating, which, according to Calico, add between 0.0002 to 0.0004 inch on average. Our experience with Calico is that it adds minimum 0.0004 inch, and can go as thick as 0.0015 inch. When Clevite aftermarket provides coating, like BE V1 shells, they are marked HK instead. There is an entire set of facebook posts by VAC themselves where they discuss their bearing design target was OEM clearance. "Click the link on the wiki page at "History of VAC Bearings: In VAC's own words."

All of the measurements at the wiki will eventually be redone. The shell thicknesses will be redone with the air gauge, and the clearances will be redone with a custom LVDT bore gauge accurate to 0.00005 inch. Those custom gauges are already done and ready to use. Preliminary tests show that the measurements with the custom gauges are 99% repeatable. The clearance measurements are performed with 34-points around the circle of each connecting rod/bearing assembly. Time availability is the only limiting factor at this point.

Yes I am aware that the VAC shells are supplied uncoated, the Clevite part number tells you this. I've seen the debate in whether VAC are extra clearance or not. For me it is just that, a debate. My understanding of Clevite numbers is H is standard, HX is extra clearance. From the images in the WIKI, VAC are extra clearance. Now I haven't seen a set to know if they supply 1 H and 1 HX part, but the WIKI measurement data suggests both are HX and are wall thickness is within 7um (0.00028") of the BE V1 0.025mm part.

I look forward to seeing the new measurement data and the repeatability improving very close to 100%

[Green-Eggs]

Originally Posted by PaulGros

Yes I am aware that the VAC shells are supplied uncoated, the Clevite part number tells you this. I've seen the debate in whether VAC are extra clearance or not. For me it is just that, a debate. My understanding of Clevite numbers is H is standard, HX is extra clearance. From the images in the WIKI, VAC are extra clearance. Now I haven't seen a set to know if they supply 1 H and 1 HX part, but the WIKI measurement data suggests both are HX and are wall thickness is within 7um (0.00028") of the BE V1 0.025mm part.

I look forward to seeing the new measurement data and the repeatability improving very close to 100%

The VAC shell thicknesses were definitely measured with the ball-anvil micrometer. Those were sold to a guy who was in a huge bind for a set of bearings when BE was out of stock. So those very early VAC shells are no longer available, and the redo will be with the latest VAC shells which could be of another thickness revision from the originals.

Another problem with the wiki, and it should be an obvious one, only one set of shells were measured for everybody except BE. There's really no way to draw statistical meaningful conclusions from only one set of measurements, yet it's surprising how often people try to do exactly that.

On the plus side, Bert is probably the only person around who still has a virgin set of 088/089 shells.

[PaulGros]

Originally Posted by Green-Eggs

The VAC shell thicknesses were definitely measured with the ball-anvil micrometer. Those were sold to a guy who was in a huge bind for a set of bearings when BE was out of stock. So those very early VAC shells are no longer available, and the redo will be with the latest VAC shells which could be of another thickness revision from the originals.

So are the measurements for the VAC shells in the WIKI the same parts that were photographed and show as HX (extra clearance)? Are you saying that they aren't the current parts that VAC sell? I would expect any revision in design would also to be allocated a new part number from Clevite even if VAC retained their part number.

QUOTE=Green-Eggs;30948766]
Another problem with the wiki, and it should be an obvious one, only one set of shells were measured for everybody except BE. There's really no way to draw statistical meaningful conclusions from only one set of measurements, yet it's surprising how often people try to do exactly that.[/QUOTE]

OK, so is the data accurate or not? Forgive me if I'm misunderstanding reading the information, but the incorrect clearance theory was drawn from this data. Yet now we shouldn't draw meaningful conclusions?

If there are multiple sets of BE bearings measured is that statistically more accurate? Because if so, the BE parts (along with other Clevite parts from VAC) are showing the largest variation in wall thickness of between 12um (0. 000472") & 13um (0.00512"). This would suggest that of all the parts measured Clevite produced parts have the largest tolerance. MAHLE have published theirs at 9um (0.00354"), independant measuring (including BE) has shown they are hitting sub 6um (0.00236"). ACL are stating 8um (0.00315"), but the BE data shows a marginal difference of 9um (0.00354").

QUOTE=Green-Eggs;30948766]On the plus side, Bert is probably the only person around who still has a virgin set of 088/089 shells.[/QUOTE]

We have the OE data / print for the 088 / 089 parts so we no longer have to guess the specifications . What would be interesting is measurement data from removed 088 / 089 shells to compare to original drawing to establish the level of wear. Given that the OE colour codes tends to be removed by oil this would also involve a level of assumption. But given that indium copper intermetallics are often visible it would suggest the level of wear would be in the region of the overlay thickness of 13um (0.000512")

[Adam M]

Can I just ask, @Green-Eggs, do you own or work for BE bearings?

[charliev68]

This thread is getting super technical and over my head with all the specs and values! What I’m getting is that…OEM specs are good, staggering bearing sets “maybe” questionable. I think 🤔 that’s what Mahle is saying!

[PaulGros]

Originally Posted by charliev68

This thread is getting super technical and over my head with all the specs and values! What I’m getting is that…OEM specs are good, staggering bearing sets “maybe” questionable. I think 🤔 that’s what Mahle is saying!

Bearings is a technical subject, more so than many realise :-)

OEM specs are almost double what has been quoted ( 0.0254mm / .001" quoted vs 0.046mm / .0018")

In fact, if you run the data from the BE WIKI on the shells they have measured you will find ALL currently available bearings including OEM are remarkably close in clearance at around 2um (.00008") variation.

Now, I'm deliberately using the wiki data as this is often where people are pointed so there is consistency in comparisons.

Given the smallest part tolerance that is proven is the OEM parts at 6um (.0002") you begin to understand how small the difference is on all of these parts.

The only time we see a larger deviation on clearance is when you run theoretical min / max clearances between the parts. The problem is here that the largest measured deviation on some parts is over 12um (.00047") where others are deviating only 1um (.00004"). This then shows the parts with the largest measured size deviation will show a larger theoretical clearance. Theoretical clearances based on min / max is only relevant when absolute min / max sizes are known, including production tolerance. We don't have this data for all parts.

For reference here are the production tolerances that are known.

OEM 6um total - source original drawing
MAHLE 9um total - source original drawing
ACL 8um total - source email to ACL
King currently unknown, data requested
BE currently unknown

In short, the differences between bearings available are as follows:

1. Eccentricity (is this by design or wider tolerance? Without drawings it's unknown)
2. Material specification
3. Brand
4. production tolerances

So choose your bearings based on the above, local availability, price or preference of the person doing the job. All will have clearances within spec based on published data.

And staggering bearing sets or using extra clearance (ACL HX for eg) is perfectly OK. but if it were me, I would only use after shaft and housing measurements are checked to prove the extra clearance was required or desirable.

[Helmsman]

Originally Posted by PaulGros

Bearings is a technical subject, more so than many realise :-)

OEM specs are almost double what has been quoted ( 0.0254mm / .001" quoted vs 0.046mm / .0018")

In fact, if you run the data from the BE WIKI on the shells they have measured you will find ALL currently available bearings including OEM are remarkably close in clearance at around 2um (.00008") variation.

Now, I'm deliberately using the wiki data as this is often where people are pointed so there is consistency in comparisons.

Given the smallest part tolerance that is proven is the OEM parts at 6um (.0002") you begin to understand how small the difference is on all of these parts.

The only time we see a larger deviation on clearance is when you run theoretical min / max clearances between the parts. The problem is here that the largest measured deviation on some parts is over 12um (.00047") where others are deviating only 1um (.00004"). This then shows the parts with the largest measured size deviation will show a larger theoretical clearance. Theoretical clearances based on min / max is only relevant when absolute min / max sizes are known, including production tolerance. We don't have this data for all parts.

For reference here are the production tolerances that are known.

OEM 6um total - source original drawing
MAHLE 9um total - source original drawing
ACL 8um total - source email to ACL
King currently unknown, data requested
BE currently unknown

In short, the differences between bearings available are as follows:

1. Eccentricity (is this by design or wider tolerance? Without drawings it's unknown)
2. Material specification
3. Brand
4. production tolerances

So choose your bearings based on the above, local availability, price or preference of the person doing the job. All will have clearances within spec based on published data.

And staggering bearing sets or using extra clearance (ACL HX for eg) is perfectly OK. but if it were me, I would only use after shaft and housing measurements are checked to prove the extra clearance was required or desirable.

Hi again Paul!

Again I’m aware of that you are not employed by MAHLE and I don’t expect you to know all answers. Still curious of your (and possibly other – initiated - peoples) view here:

Quick question on the BMW OE shells
As the shaft and housing variation together is 0.030 on its own (according to the specs I’ve seen?), I still don’t quite get how BMW clearance variance can be as tight as 33um (0.062 – 0.029mm), with a shell variation of apparently 0.006 on top. There are a couple of different shell thicknesses to play around with but doesn’t help really if surrounding HW isn’t known. So each shafts and houses are measured during assembly of the engine where after matched with the shells?

MAHLE
MAHLE’s target extended clearance (54um vs OE 46um) from what I understand to avoid getting too tight/lose in the ends due to overall variation (rather than increasing the oil film)?

Others
The various “extended” shells seems to all target around 55-60um clearance, i.e. similar to MAHLE?

So...
should my understanding be that 1) Originally installed OE bearings should theoretically be fine “for ever”, obviously assuming the engine is maintained/treated correctly. 2) If one for whatever reason (tons of miles, 10th owner etc) want to replace the RBs it should be with extended clearance due to mentioned tolerance stack up?

Thanks,
Nik

[PaulGros]

Hi Nik,

Originally Posted by Helmsman

Hi again Paul!

Again I’m aware of that you are not employed by MAHLE and I don’t expect you to know all answers. Still curious of your (and possibly other – initiated - peoples) view here:

Quick question on the BMW OE shells
As the shaft and housing variation together is 0.030 on its own (according to the specs I’ve seen?), I still don’t quite get how BMW clearance variance can be as tight as 33um (0.062 – 0.029mm), with a shell variation of apparently 0.006 on top. There are a couple of different shell thicknesses to play around with but doesn’t help really if surrounding HW isn’t known. So each shafts and houses are measured during assembly of the engine where after matched with the shells?

The shell variation is per shell, not per pair. Unfortunately we do not know how BMW graded the engine assembly or what criteria they applied on the selection of bearing shells. It would be normal for the component manufacturer to grade the parts. But we honestly do not know if this was the case.

It seems odd to use one of each grade on assembly as per the parts catalogue. This could have been achieved with a single grade part. Again, I have seen no data to prove or disprove what colour codes were fitted from the factory.

Originally Posted by Helmsman

MAHLE
MAHLE’s target extended clearance (54um vs OE 46um) from what I understand to avoid getting too tight/lose in the ends due to overall variation (rather than increasing the oil film)?

Others
The various “extended” shells seems to all target around 55-60um clearance, i.e. similar to MAHLE?

Correct. The increase in clearance is mainly due to offering a single grade part to ensure it works at all extremes and the material MAHLE has selected to use. I think in the MAHLE document they used the phrase er on the side of caution :-)

What others are targeting is difficult to say. I see posts saying "we are targeting x clearance" for example, but no data to back this up. I know some are stating "real world" measurements, but for me the first stage has failed. How do we know the parts measured are the size the manufacturer intended i.e. are they in spec?

I know you kindly queried ACL on their tolerances so maybe I will use that data along with the MAHLE Motorsport data which will give us theoretical min / max as well as average clearance from drawing specs. Maybe BE will divulge their specs and we can add those too. I have asked king for the same information.

Maybe then we can start to build a clear picture of target Vs actual each bearing brand available.

Originally Posted by Helmsman

So...
should my understanding be that 1) Originally installed OE bearings should theoretically be fine “for ever”, obviously assuming the engine is maintained/treated correctly. 2) If one for whatever reason (tons of miles, 10th owner etc) want to replace the RBs it should be with extended clearance due to mentioned tolerance stack up?

Thanks,

Ignoring the debate on OE shells and only looking at data, yes the OE shells should be fine.

We really need to define the "extended clearance" It is not unusual for aftermarket bearings to run marginally larger clearances. This is due to aftermarket parts only having a single grade V's OE multiple grades. It ensures the part stays within spec at the extremes of tolerances.

For me "extended clearance" for means parts that have additional clearance over the standard part from the same manufacturer. These commonly use an X in the part number (ACL / Clevite HX, King STDX). These parts typically have 0.025mm (.001") extra clearance over the standard part from the same company.

Anything else would be a standard part (even though technically they have a little extra clearance to OEM).

[Helmsman]

Originally Posted by PaulGros

Hi Nik,

The shell variation is per shell, not per pair. Unfortunately we do not know how BMW graded the engine assembly or what criteria they applied on the selection of bearing shells. It would be normal for the component manufacturer to grade the parts. But we honestly do not know if this was the case.

It seems odd to use one of each grade on assembly as per the parts catalogue. This could have been achieved with a single grade part. Again, I have seen no data to prove or disprove what colour codes were fitted from the factory.

Hmm, I clearly missunderstand something here but eager to learn. How can MAHLE quote a (theoretical) Max/Min clearance if they don't have BMW's shaft/housing spec's..?

Originally Posted by PaulGros

Correct. The increase in clearance is mainly due to offering a single grade part to ensure it works at all extremes and the material MAHLE has selected to use. I think in the MAHLE document they used the phrase er on the side of caution :-)

What others are targeting is difficult to say. I see posts saying "we are targeting x clearance" for example, but no data to back this up. I know some are stating "real world" measurements, but for me the first stage has failed. How do we know the parts measured are the size the manufacturer intended i.e. are they in spec?

I know you kindly queried ACL on their tolerances so maybe I will use that data along with the MAHLE Motorsport data which will give us theoretical min / max as well as average clearance from drawing specs. Maybe BE will divulge their specs and we can add those too. I have asked king for the same information.

Maybe then we can start to build a clear picture of target Vs actual each bearing brand available.

Which again would be theoretical simply using housing-shaft-2 x shell wall...no?

Originally Posted by PaulGros

Ignoring the debate on OE shells and only looking at data, yes the OE shells should be fine.

We really need to define the "extended clearance" It is not unusual for aftermarket bearings to run marginally larger clearances. This is due to aftermarket parts only having a single grade V's OE multiple grades. It ensures the part stays within spec at the extremes of tolerances.

For me "extended clearance" for means parts that have additional clearance over the standard part from the same manufacturer. These commonly use an X in the part number (ACL / Clevite HX, King STDX). These parts typically have 0.025mm (.001") extra clearance over the standard part from the same company.

Anything else would be a standard part (even though technically they have a little extra clearance to OEM).

From what I see most fabs are offering various extended clearance (vs OE), which seems to make sense as they otherwise would introducing the risk to go too tight. My wondering really is; would replacement with OE shells be a viable option without the possibility to measure cranch and shaft?

[Green-Eggs]

Originally Posted by Helmsman

Hmm, I clearly missunderstand something here but eager to learn. How can MAHLE quote a (theoretical) Max/Min clearance if they don't have BMW's shaft/housing spec's..?


Which again would be theoretical simply using housing-shaft-2 x shell wall...no?



From what I see most fabs are offering various extended clearance (vs OE), which seems to make sense as they otherwise would introducing the risk to go too tight. My wondering really is; would replacement with OE shells be a viable option without the possibility to measure cranch and shaft?

Here's the formula for calculating theoretical clearances. This will work for different top/bottom shells too.

http://wiki.rcollins.org/core/index....ring_Clearance

[Helmsman]

Originally Posted by Green-Eggs

Here's the formula for calculating theoretical clearances. This will work for different top/bottom shells too.

http://wiki.rcollins.org/core/index....ring_Clearance

Thanks mate, seems straight forward.