Official S65 Bearing Specification/Clearance Wiki

[BPMSport]

You can see more about people with the click noise here:

http://www.m3post.com/forums/showthread.php?t=412843

(Page 12 is where I posted the pictures of the valves that came out of my car).

My check valves were replaced 10-15k miles ago as customer pay because I wanted to see if it got rid of that noise. It didn't. A waste of $1100 but was worth knowing that they were not faulty.

Many 08's when left cold for awhile make a not so pleasant noise on start. I can record a few others if you like.

This thread has nothing to do with my motor and I'm simply sharing my experience, it's not intended to start an argument with you. I'm happy not to post anymore if sharing my experiences will have such a result. As an M3 owner myself of 5 years, I'm just as interested as everyone else in learning more about these issues.

I've spoken with three shop foremans and they advised that new motors are filled with 10w60.
I have been unable to find Bmw documentation that specifies that the break-in oil for the S65 as something different than 10w60.

Thanks

Edit: There are four check valves, two in each head - not two total. It also doesn't require 75% motor disassembly to replace them.

[regular guy]

Originally Posted by Mike Benvo

You can see more about people with the click noise here:

http://www.m3post.com/forums/showthread.php?t=412843

(Page 12 is where I posted the pictures of the valves that came out of my car).

My check valves were replaced 10-15k miles ago as customer pay because I wanted to see if it got rid of that noise. It didn't. A waste of $1100 but was worth knowing that they were not faulty.

Many 08's when left cold for awhile make a not so pleasant noise on start. I can record a few others if you like.

This thread has nothing to do with my motor and I'm simply sharing my experience, it's not intended to start an argument with you. I'm happy not to post anymore if sharing my experiences will have such a result. As an M3 owner myself of 5 years, I'm just as interested as everyone else in learning more about these issues.

I've spoken with three shop foremans and they advised that new motors are filled with 10w60.
I have been unable to find Bmw documentation that specifies that the break-in oil for the S65 as something different than 10w60.

Thanks

Thanks for the clarification of customer expense on the check valves. The whole notion of changing a perfectly running motor and then rebuilding it anyways before reusing it -- "just because" -- still doesn't make any sense. But we can leave it at that.

I've never found anything that specifies S65 break-in oil. Current TIS only tells you that no special break in oil is required. But the image posted on post-9, note-3 is a direct export from TIS circa 12/2007 - 01/2008. At that time, TIS unequivocally says the break-in oil as 5W30 for S85. I asked the question of sunsweet earlier whether S65 would have a different break in oil than S85, and if so...why? He never answered.

[regular guy]

Why do we re-measure?

When this thread was originally created, the data was based on measuring the thicknesses and clearances on used parts. Two OEM crankshafts were measured, one set of factory connecting rods, one set of Carrillo (aftermarket) rods, one single (virgin) 088/089 bearing pair, one single (virgin) 702/703 bearing pair, two Calico coated bearings (not installed), and two WPC treated bearings (installed). The measurements were performed on the OEM rods and Carrillo rods and clearance checked against the two OEM crankshafts. Where possible, additional data and comparisons were made against the supercharged stroker build at Van Dyne Engineering.

At the time of the original measurements, it was thought that the 702/703 bearing was simply a materials change with no changes to thickness or clearance. However my original measurements showed a slight increase in clearance on the virgin 702/703 bearing. For whatever reason, nobody gave much thought to this slight difference. A few weeks later, I sent kawasaki00 the bearings used in the original test, and his equipment did detect a difference between thickness of the 088/089 bearings and 702/703 bearings. So a few weeks later, I sent kawasaki00 a full set of virgin 702/703 bearings and connecting rods to take further measurements. His later findings confirmed the first ones that the 702/703 bearings were slightly thinner and showed slight increase in clearance over the original 088/089 bearings. But questions still remained because kawasaki00 wasn’t measuring at the same temperatures or with the same equipment as my original tests.

To figure out whether or not BMW changed the bearing clearances when they introduced the 702/703 bearings, it was going to require obtaining a full set of virgin 088/089 bearings and another set of virgin 702/703 bearings along with two sets of new connecting rod bolts. While I didn’t mind doing all the work, I couldn’t afford the $950 cost this would require. Thanks to many members of the m3post community, they donated the $950 to buy all new parts and make it possible to re-measure and re-confirm and/or refine the original findings.

Even though all of the original measurements are accurate and valid on their own merit (and will remain on record), they’re not made with virgin parts as if there were installed on the factory assembly line. To obtain these types of measurements would require measuring more than one virgin bearing of each 088/089 and 702/703 bearings on OEM connecting rods with brand new rod bolts. So I purchased and re-measured and entire new set of bearings and new connecting rod bolts. This will allow me to obtain as close to factory specifications as possible.





Bearing Thickness Revisited

During the “redo” clearance measurements, I was also committed to using the best practice to obtain bearing materials thickness. Many of you following this thread recall the debates regarding the use and accuracy of ball-anvil micrometers versus test indicators to measure bearing thickness. In my last post on the topic, I demonstrated the effects of “cosine error” when using a test indicator to measure bearing thickness. Cosine error occurs when the test indicator is placed at an angle with respect to the measuring surface instead of the proper method of placing it 180-degrees with respect to the measuring surface. To compensate for cosine error, an adjustment factor is applied to the measurements to adjust them back to their original values as if the test indicator was used properly.

Just to settle this topic once and for all, I purchased a few more fixtures for the test indicator to help mathematically verify the cosine error. Then as one final test, I wanted to use the test indicator as it was designed to measure bearing thickness to see if I could get the ball-anvil micrometer and test indicator to agree without any cosine error correction.

Verifying “cosine error” with a test indicator

Before verifying the effects of cosine error on the test indicator, I first needed to verify the accuracy of the test indicator instrument on a flat surface. I set up the test indicator on a granite calibration block with the needle as perfectly parallel to the surface as I could obtain. Then using this technique, I verified the measurements of various calibration blocks.

Calibrated Thickness	Measured Thickness	Images
0.10000	0.10000
0.10100	0.10100
0.11000	0.11000
0.12000	0.12005
0.13000	0.13010

Following the initial flat-surface calibration (above), I changed the angle of the test indicator to 23-degrees. If such a cosine error exists, then the measurements of the test indicator will show significant changes.

Calibrated Thickness	Measured Thickness	Images	Measuremts at 23-degrees	Cosine Calculation	% Error	Images
0.10000	0.10000		0.10000	0.10000	0.00000
0.10100	0.10100		0.10015	0.10014	0.00853
0.11000	0.11000		0.11015	0.10934	0.00597
0.12000	0.12005		0.12135	0.11965	0.00331
0.13000	0.13010		0.13250	0.12992	0.00141

The results clearly show the results of cosine error when using the test indicator in a non-standard position (180-degrees to the surface of measurement). Although not completely accurate, using the cosine error calculation, the results were adjusted back to their near-correct values. NOTE: There are some calculations missing from the tables above. The base measurement was 0.10000 inch above the flat surface. So adjustments were necessary to calculate this "translated" offset. These additional calculations are missing from the tables above and only apply to the % error values.

Re-measuring the bearing thicknesses on 702/703 bearings

Now that we got the formalities out of the way of mathematically proving the cosine error of the test indicator, we can use the ball-anvil micrometer and test indicator to measure the same sets of bearings to see how closely they match. The electronic test indicator has a very nice "MIN" mode for these types of measurements. It is designed to capture the minimum thickness of an items such as a curved bearing surface. When programmed in this mode, once you place the bearing under the test indicator, you simply move the bearing back and forth a quarter inch or so while the test indicator measures and freezes the lowest ("MIN") point it finds. Using this technique, one can verify the measurements between the ball-anvil micrometer and test indicator.

Before I get started, I always calibrate the instruments. Then I take the measurements.

702 Bearings	1	2	3	4	5	6	7	8	Mode	Min. Dev.	Max. Dev.
Ball Anvil	0.07850	0.07855	0.07850	0.07850	0.07855	0.07855	0.07850	0.07850	0.07850	0.00000	0.00005
Test Indicator	0.07850	0.07855	0.07845	0.07850	0.07855	0.07855	0.07855	0.07850	0.07855	-0.00010	0.00000
BA Photo
TI Photo

703 Bearings	1	2	3	4	5	6	7	8	Mode	Min. Dev.	Max. Dev.
Ball Anvil	0.07865	0.07865	0.07870	0.07865	0.07865	0.07865	0.07865	0.07870	0.07865	0.00000	0.00005
Test Indicator	0.07870	0.07865	0.07870	0.07865	0.07865	0.07865	0.07865	0.07870	0.07865	0.00000	0.00005
BA Photo
TI Photo

As these measurements show, whether using a ball-anvil micrometer, or properly used test indicator, the measurements will end up the same. I hope this settles this debate once and for all. For all future bearing thickness measurements, I will only use the ball-anvil micrometer.

Official 702/703 Bearing Thickness Measurements

Engine	S65B40
Bearing Dimensions (702/703 Bearings)	Metric Dimensions	SAE (Inch) Dimensions
Nominal Bearing Thickness (Top, Blue)	1.9977 mm	0.07865 inch
Nominal Bearing Thickness (Bottom, Red)	1.9939 mm	0.07850 inch
Bearing Thickness Variance (Top, Blue)	1.9977 - 1.9990 mm	0.07865 - 0.07870 inch
Bearing Thickness Variance (Bottom, Red)	1.9939 - 1.9952 mm	0.07850 - 0.07855 inch
Bearing Tolerance (Top, Blue)	-0.0000, +0.0013 mm	-0.00000, +0.00005 inch
Bearing Tolerance (Bottom, Red)	-0.0000, +0.0013 mm	-0.00000, +0.00005 inch

Re-measuring the bearing thicknesses on 088/089 bearings

Since the 088/089 bearings are discontinued, the hardest part was to find the bearings in the first place. The BMW computers showed 70-pieces of 089 bearings in stock somewhere in the USA, but every time I placed the order, 703 bearings would show up instead. This same pattern repeated itself three times until my local BMW dealer was able to pull the right strings to purchase a full set of 089 bearings.

089 Bearings	1	2	3	4	5	6	7	8	Mode	Min. Dev.	Max. Dev.
Ball Anvil	0.07890	0.07885	0.07890	0.07885	0.07885	0.07885	0.07885	0.07880	0.07885	-0.00005	0.00005
BA Photo

In spite of my good fortune purchasing the 089 bearings, purchasing the 088's was not possible. I had BMWNA search the entire dealer network and distribution chains. There were zero 088's available. I then tried Canada, Mexico, South America, Germany, and Britain. I never got responses from Mexico or South America, but all the other responses came back negative. 088 bearings were no longer available.

Since the upper bearings (089's) are the ones that take all the abuse, they are the ones most important to obtain. Thankfully, that's exactly what I was able to purchase. Obtaining virgin lower bearings was much less important because they seem to wear very little with respect to the tops. To obtain the best set of lower bearings I could for this test, I went through my collection of bearings and found the most pristine 088 bearings. I chose them with these criteria in mind: 1) low uniform wear across the shell; 2) Low-side variance shouldn't be greater than 089 bearings; 3) High-side variance should include as much range as possible to help simulate tolerance stack up. These are the bearings I selected from the collection.

088 Bearings	1	2	3	4	5	6	7	8	Mode	Min. Dev.	Max. Dev.
Ball Anvil	0.07875	0.07880	0.07860	0.07860	0.07865	0.07865	0.07865	0.07865	0.07865	-0.00005	0.00015

Official 088/089 Bearing Thickness Measurements

Engine	S65B40
Bearing Dimensions (088/089 Bearings)	Metric Dimensions	SAE (Inch) Dimensions
Nominal Bearing Thickness (Top, Blue)	2.0028 mm	0.07885 inch
Nominal Bearing Thickness (Bottom, Red)	1.9977 mm	0.07865 inch
Bearing Thickness Variance (Top, Blue)	2.0015 - 2.0041 mm	0.07880 - 0.07890 inch
Bearing Thickness Variance (Bottom, Red)	1.9964 - 2.0015 mm	0.07860 - 0.07880 inch
Bearing Tolerance (Top, Blue)	-0.0013, +0.0013 mm	-0.00005, +0.00005 inch
Bearing Tolerance (Bottom, Red)	-0.0013, +0.0038 mm	-0.00005, +0.00015 inch

[Yellow Snow]

Not sure about the quality of that clock, but I get no such errors using the Tesa Clock.

[///M Power-Belgium]

Measurements this precise should be done with Zeiss !

[Yellow Snow]

Originally Posted by ///M Power-Belgium

Measurements this precise should be done with Zeiss !

Speaking solely about bearing thickness, I would still rather rely on a properly setup comparator test using high quality equipment in the correct way.

That Zeiss CMM would give a good no nonsense measurement of the full assembly though.

[GIdriver]

Thanks to all our experts involved in the discussion(s) of this issue. A special thanks goes to Regular Guy for the massive effort he has put into this. It is very much appreciated.

[GIdriver]

So, being my 2011 E90 produced on May 2010, that means I'm screwed with the old thicker bearings, correct?

[kawasaki00]

Originally Posted by GIdriver

So, being my 2011 E90 produced on May 2010, that means I'm screwed with the old thicker bearings, correct?

No because I know of many engines with newer bearings that have gone out too. Luck of the draw.

[aussiem3]

Originally Posted by Mike Benvo

You can see more about people with the click noise here:

http://www.m3post.com/forums/showthread.php?t=412843

(Page 12 is where I posted the pictures of the valves that came out of my car).

My check valves were replaced 10-15k miles ago as customer pay because I wanted to see if it got rid of that noise. It didn't. A waste of $1100 but was worth knowing that they were not faulty.

Many 08's when left cold for awhile make a not so pleasant noise on start. I can record a few others if you like.

This thread has nothing to do with my motor and I'm simply sharing my experience, it's not intended to start an argument with you. I'm happy not to post anymore if sharing my experiences will have such a result. As an M3 owner myself of 5 years, I'm just as interested as everyone else in learning more about these issues.

I've spoken with three shop foremans and they advised that new motors are filled with 10w60.
I have been unable to find Bmw documentation that specifies that the break-in oil for the S65 as something different than 10w60.

Thanks

Isn't this noise related to the VANOS. I use to get it when the car has been parked for over a week or so. BMW said it was normal. I think the noise increases with the age and mileage. It is some check-valve related.

[thekurgan]

If the bearings are harder, it makes sense for added clearance. It's possible BMW HAD to increase clearance when forced to remove the lead.

[FogCityM3]

Thanks, RG and the thought leaders on this thread, as well as those who donated to the cause. Really interesting and useful information and I personally learned a ton in the process!

[Yellow Snow]

Originally Posted by thekurgan

If the bearings are harder, it makes sense for added clearance. It's possible BMW HAD to increase clearance when forced to remove the lead.

I don't think there is any confirmation in the thread yet of a difference in the bearing overlay hardness. It could be exactly the same

[regular guy]

Originally Posted by Yellow Snow

Not sure about the quality of that clock, but I get no such errors using the Tesa Clock.

Call me crazy, but I'll take the big name and big reputation of Fowler over a no-name device any day.

Originally Posted by Yellow Snow

Speaking solely about bearing thickness, I would still rather rely on a properly setup comparator test using high quality equipment in the correct way.

And that's exactly what I did. I guess you're still having a hard time dealing with your findings and techniques being shown as faulty. It's time to move on.

I have another question for you; it's been on my mind for a while. You say no S85/S65's have failed in GB, yet you describe having changed somebody's bearings and measured their rod and bearing thicknesses and clearance. You also showed a picture of measuring one of the new bearings. So naturally I'm wondering if nobody in GB ever needed a change, then what was the work you performed? Was it a race engine build? Can you provide more information about whatever you did to this engine...and why you did it? Do you care to share the build sheet with all the specs you measured?

[Yellow Snow]

Originally Posted by regular guy

Call me crazy, but I'll take the big name and big reputation of Fowler over a no-name device any day.



And that's exactly what I did. I guess you're still having a hard time dealing with your findings and techniques being shown as faulty. It's time to move on.

I have another question for you; it's been on my mind for a while. You say no S85/S65's have failed in GB, yet you describe having changed somebody's bearings and measured their rod and bearing thicknesses and clearance. You also showed a picture of measuring one of the new bearings. So naturally I'm wondering if nobody in GB ever needed a change, then what was the work you performed? Was it a race engine build? Can you provide more information about whatever you did to this engine...and why you did it? Do you care to share the build sheet with all the specs you measured?

Yes, I would call you crazy for using a test indicator that shows errors. Where is your 2mm slip gauge? It still looks like you are using it incorrectly. Heaven forbid you would ever admit to not knowing how to measure but I can assure you that you are in the wrong. You should begin by scrubbing that filthy surface plate for starters.

I don't recall ever saying that no S85/S65s have ever failed in the UK. That's just you trying to discredit yet again. I said I have never heard of any. Get my terminology correct please.

That S85 crank that I worked on had the old shells put back as was explained in the thread. I bought the shells myself to get an accurate measurement for my own benefit

[swamp2]

Despite the fact that clearance, obtained from installed diameter, is the critical measurement here and despite the fact that we've seen an almost 200% difference in maximum measured/calculated clearances vs. minimum ones (recall kawasaki00s "special installation technique" and results in this post.) I think this is absolutely firm evidence that the newer 702/703 bearings are thinner (at least in one location) than the older 088/089 bearings.

A quick look at the differences in mean (again RG your use of mode is odd to say the least) values compared to the standard deviation shows that this is NOT about measurement noise. Furthermore the solid comparison between the two techniques should lay all questions to rest about that (Yellow Snow). The difference in the mean of the ball anvil measured, new, 703 upper vs. ball anvil measured, new, 089 upper divided by the (larger) standard deviation is a factor of 6! This tells us right away we are measuring a real difference here.

For those that prefer the formality of a rigorous statistical test I've run the Student's t-test which tell us that there is a 4 in 1 billion chance that the (upper shell) parts have a nominally identical mean thickness value. Now we can not yet conclude that BMW purposefully designed in a dimensional change. We probably never will be able to reach such a strong conclusion. Perhaps existing supplier manufacturing tooling with the new material resulted in the difference and BMW then just observed the difference and accepted it. That's just one of a plethora of possibilities...

It is also interesting to do a quick analysis of how much change in clearance would result assuming no change in thickness along 90 degrees from the part line. This is just subtracting the sum of the sizes of both old bearings from the sum of the sizes of the new bearings and dividing by a nominal clearance (again we don't really know that nominal value yet, but we have some idea of a range of values, perhaps 0.0014 - 0.0023). The percent differences are quite significant, 15-25%, depending on which nominal clearance value you use.

I don't mind being proven wrong about speculation, and although a bit too early to firmly conclude so, it appears that along with the bearing material change BMW did specify or permit a change in rod bearing clearance, opening it up significantly.

Let's see what RG has for us next with installed diameter numbers. Those are radically more important than these. That should be the final proof.

[swamp2]

Originally Posted by thekurgan

If the bearings are harder, it makes sense for added clearance. It's possible BMW HAD to increase clearance when forced to remove the lead.

Neither conclusion follows from the observations. There are many non-intuitive things about journal bearing design. BMWs bearing supplier clearly could make a bearing identical in fitted (or unfitted) size to the older one. One can make any size one prefers...

[regular guy]

Originally Posted by swamp2

A quick look at the differences in mean (again RG your use of mode is odd to say the least) values compared to the standard deviation shows that this is NOT about measurement noise.

If MODE isn't the right thing, then tell me what is. I don't really know statistics. I thought you said to use MODE. Just let me know what's right, and I'll repost.

[aussiem3]

My head is spinning after this 79th page.

[swamp2]

Originally Posted by regular guy

If MODE isn't the right thing, then tell me what is. I don't really know statistics. I thought you said to use MODE. Just let me know what's right, and I'll repost.

It's not that mode is incorrect, it is just non-standard and less mathematically useful. My response suggesting mode was only putting into words your preference for choosing the most common value, even back then I suggested mean. Either way, since you are generous and thorough enough with the original measurement, its not at all a big deal if mean and standard deviation and the like are missing. Cheers.

[regular guy]

Originally Posted by Yellow Snow

Yes, I would call you crazy for using a test indicator that shows errors. Where is your 2mm slip gauge? It still looks like you are using it incorrectly. Heaven forbid you would ever admit to not knowing how to measure but I can assure you that you are in the wrong. You should begin by scrubbing that filthy surface plate for starters.

The surface plate is brand new; purchased just for this test. Clean as a whistle. The discolorations you see are the oils from hands and using brand new equipment that also had oils on their surfaces. The plate surface was wiped before every use even though wiping wasn't necessary. To comment on it presents yourself as an expert. I would have expected such an expert to be more aware of what he was viewing before commenting incorrectly.

If I measured wrong, then please describe what I did wrong and prove it with your own measurements and photos. I think we are all waiting for you to do this instead of playing Mr. Gotcha Gadfly. You didn't know about cosine error when it was first discussed, and I've shown exactly how it works with measurements, equations, and photographic proof. The only photo you've shown demonstrated the misuse of your own equipment. Swamp along with other more qualified engine builders than yourself (if you're a qualified engine builder) have also told you that you and your technique is wrong. I think we will all benefit if you heed that advice, learn from it, and move on.

I don't recall ever saying that no S85/S65s have ever failed in the UK. That's just you trying to discredit yet again. I said I have never heard of any. Get my terminology correct please.

OK, whatever. But that's not even true is it? Others have repeatedly pointed to threads where people in GB were discussing their blown motors. I can recall four or five other times you've been caught red-handed with fabricated data in this thread. There is no need for me to discredit you and I have no desire to do so. I could not do any more damage than you've done to yourself. I only desire accurate information. I'm sorry if the accurate information exposes your techniques as fatally flawed and caused you discomfort.

That S85 crank that I worked on had the old shells put back as was explained in the thread. I bought the shells myself to get an accurate measurement for my own benefit

If I'm reading that right, you're saying that you actually put old shells back in an engine and sent the customer down the road that way? Are you serious? I may have missed that the first time. I would never go to a shop that does such a thing or even thinks it's OK to do that.

You showed one shell that was improperly measured on substandard, no-name equipment and you call that your claim to fame and basis of proof. Yet you demand even greater proof from me than you're willing to do yourself. I've given you eight shells that were measured properly with better equipment with photographic evidence. The measurements all match. That's the end of the story for you and this saga. Learn from it, and move on.

At the end of the day, the only thing that will matter is the installed clearance. Stay tuned...that data is coming in the next two days.

[catpat8000]

Originally Posted by Yellow Snow

Yes, I would call you crazy for using a test indicator that shows errors. Where is your 2mm slip gauge? It still looks like you are using it incorrectly. Heaven forbid you would ever admit to not knowing how to measure but I can assure you that you are in the wrong. You should begin by scrubbing that filthy surface plate for starters.

Originally Posted by regular guy

The surface plate is brand new; purchased just for this test. Clean as a whistle. The discolorations you see are the oils from hands and using brand new equipment that also had oils on their surfaces. The plate surface was wiped before every use even though wiping wasn't necessary. To comment on it presents yourself as an expert. I would have expected such an expert to be more aware of what he was viewing before commenting incorrectly.

If I measured wrong, then please describe what I did wrong and prove it with your own measurements and photos. I think we are all waiting for you to do this instead of playing Mr. Gotcha Gadfly. You didn't know about cosine error when it was first discussed, and I've shown exactly how it works with measurements, equations, and photographic proof. The only photo you've shown demonstrated the misuse of your own equipment. Swamp along with other more qualified engine builders than yourself (if you're a qualified engine builder) have also told you that you and your technique is wrong. I think we will all benefit if you heed that advice, learn from it, and move on.

I have to say, I'm a little bit puzzled by Yellow Snow's continued focus on the measurement procedure. The fact that RG has followed two different patterns of measurement and the agreement between these is almost perfect, suggests that you, YS, are tilting at windmills.

I gave RG $200 to help pay for the Great Bearing Measurement of 2014 and I'm quite satisfied with his work so far.

Pat