Official S65 Bearing Specification/Clearance Wiki

[Shredicus]

Originally Posted by Mvy

A collective oh shit from all of use who installed Calico coated bearings.... Did the builds you looked at also use the ARP connecting bolts?

+1 curious about that as well.

Also welcome back rg

[regular guy]

Originally Posted by Shredicus

+1 curious about that as well.

Both were factory rod bolts. But let me take a moment to comment on the rod bolt theory.

1. I've never seen any examples of rod bolt failure. But this is a little nebulous. When a rod bolt fails, it's likely going to puke a rod out the side of the block. We've seen plenty of rods puked out of the side of a block. But when the pieces are available for investigation, they all show the same thing: big end of the connecting rod turned blue due to excessive heat. Those aren't rod bolt failures, those are bearing failures.
   
2. I found it odd that many of the folks who to steadfastly denied BMW would have screwed up the bearing clearance design were quick to believe that the manufacturing line was habitually undertorquing rod bolts. Let's say there's been 50 sets of bearings inspected now, and approximately 49 of them showed excessive wear. To me that sounds an awful lot like "we don't believe a design defect could cause excessive bearing wear in 98% of all bearings we've seen, but we do believe 98% of all engines are assembled incorrectly." That kind of reasoning has me scratching my head a little.

I sent Kawasaki some factory rods, virgin factory bolts, Carrillo rods and M-CARR bolts. He's performing analysis on them and will post the results when they become available.

Also welcome back rg

Thank you.

[kawasaki00]

Originally Posted by regular guy

I sent Kawasaki some factory rods, virgin factory bolts, Carrillo rods and M-CARR bolts. He's performing analysis on them and will post the results when they become available.

We are doing clamp load testing for the stock new/stock used and Carrillo bolts to compare all. It is tedious work, will post in this thread when finished. RG was gracious enough to send these over, it is appreciated.

[Acree]

Originally Posted by regular guy

Thank you.

I've been on automotive forums for almost 15 years, and as an engineer, I have to say this is the most impressive post I've ever seen. I kept thinking to myself "holy shit" as I kept reading and discovering the lengths you've gone to document this for the community. I just wanted to say thank you for all of the time and effort. I've done a few engine builds and swaps myself but this is above and beyond. Thank you for what you've done.

[Mvy]

Originally Posted by Acree

I've been on automotive forums for almost 15 years, and as an engineer, I have to say this is the most impressive post I've ever seen. I kept thinking to myself "holy shit" as I kept reading and discovering the lengths you've gone to document this for the community. I just wanted to say thank you for all of the time and effort. I've done a few engine builds and swaps myself but this is above and beyond. Thank you for what you've done.

+1 million, Thanks Kawasaki and Regular Guy!!

[BMRLVR]

Originally Posted by regular guy

Both were factory rod bolts. But let me take a moment to comment on the rod bolt theory.

1. I've never seen any examples of rod bolt failure. But this is a little nebulous. When a rod bolt fails, it's likely going to puke a rod out the side of the block. We've seen plenty of rods puked out of the side of a block. But when the pieces are available for investigation, they all show the same thing: big end of the connecting rod turned blue due to excessive heat. Those aren't rod bolt failures, those are bearing failures.
   
2. I found it odd that many of the folks who to steadfastly denied BMW would have screwed up the bearing clearance design were quick to believe that the manufacturing line was habitually undertorquing rod bolts. Let's say there's been 50 sets of bearings inspected now, and approximately 49 of them showed excessive wear. To me that sounds an awful lot like "we don't believe a design defect could cause excessive bearing wear in 98% of all bearings we've seen, but we do believe 98% of all engines are assembled incorrectly." That kind of reasoning has me scratching my head a little.

I sent Kawasaki some factory rods, virgin factory bolts, Carrillo rods and M-CARR bolts. He's performing analysis on them and will post the results when they become available.



Thank you.

Originally Posted by kawasaki00

We are doing clamp load testing for the stock new/stock used and Carrillo bolts to compare all. It is tedious work, will post in this thread when finished. RG was gracious enough to send these over, it is appreciated.

See, this is how proper testing and data compilation is supposed to be done. Not making a post saying: " We tore apart an S65 with x brand rod bearings and y brand rod bolts and the bearings looked good. Without any clearances, or clamping force numbers we can conclude that X brand rod bearings and y brand rod bolts are the solution to the bearing wear problem".

[Hujan]

Originally Posted by BMRLVR

See, this is how proper testing and data compilation is supposed to be done. Not making a post saying: " We tore apart an S65 with x brand rod bearings and y brand rod bolts and the bearings looked good. Without any clearances, or clamping force numbers we can conclude that X brand rod bearings and y brand rod bolts are the solution to the bearing wear problem".

I assume you're referring to MRF's thread. If so, the first of the two boldfaced sentences is an accurate characterization of MRF's comments. The second is not, and it is quite unfair of you to suggest it is.

Having read your past posts, I expected better from you.

[Malek@MRF]

Originally Posted by BMRLVR

See, this is how proper testing and data compilation is supposed to be done. Not making a post saying: " We tore apart an S65 with x brand rod bearings and y brand rod bolts and the bearings looked good. Without any clearances, or clamping force numbers we can conclude that X brand rod bearings and y brand rod bolts are the solution to the bearing wear problem".

Originally Posted by Hujan

I assume you're referring to MRF's thread. If so, the first of the two boldfaced sentences is an accurate characterization of MRF's comments. The second is not, and it is quite unfair of you to suggest it is.

Having read your past posts, I expected better from you.

I am sure this comment was directed toward that particular thread that was started, and unfortunately based on speculation.

Yes, you are right, a couple S65's are here and have been torn down, a third one is on it's way. You are assuming that I definitively asserted that rod bolts were the cause of the problem. You are wrong to assume, in thinking that none of these engines have been fully spec'd for clearances and what the findings have been. They have been, 3 times each, for the sake of redundancy and repeatability.

[regular guy]

Originally Posted by Hujan

I assume you're referring to MRF's thread. If so, the first of the two boldfaced sentences is an accurate characterization of MRF's comments. The second is not, and it is quite unfair of you to suggest it is.

Having read your past posts, I expected better from you.

I know Malek plans to post all of the measurements when he has time. I wish I could have come visit Malek last week on my visit to SoCal. I even brought all my measuring gear so I could lend a hand and help finish the job. But I was seriously under the weather and just didn't feel like staying out in Orange County any longer than I had to. Sorry Malek, I wanted to come by!

[regular guy]

Originally Posted by Malek@MRF

I am sure this comment was directed toward that particular thread that was started, and unfortunately based on speculation.

Yes, you are right, a couple S65's are here and have been torn down, a third one is on it's way. You are assuming that I definitively asserted that rod bolts were the cause of the problem. You are wrong to assume, in thinking that none of these engines have been fully spec'd for clearances and what the findings have been. They have been, 3 times each, for the sake of redundancy and repeatability.

But on the other hand Malek, many people in your thread were saying that and no corrections were made over there. So it's not a totally unfair criticism IMO.

I have full measurements of two engines with Calico bearings. And like you, I will post them when I get time. Time is always a precious commodity. :-)

[BMRLVR]

Originally Posted by Hujan

I assume you're referring to MRF's thread. If so, the first of the two boldfaced sentences is an accurate characterization of MRF's comments. The second is not, and it is quite unfair of you to suggest it is.

Having read your past posts, I expected better from you.

You can expect better all you want and assume who it was directed at all you want. I made the post without mentioning any names or brands of components so any conclusions drawn will be purely speculative.

Basically my point is when posting about something like this rod bearing issue numbers are paramount or the information is speculative and not founded nor beneficial to the community.

My post was not to make anyone in particular look bad but rather to say that numbers are crucial when trying to draw any conclusion like this......... Look at the information in this wiki thread and how it is documented and layed out, yet many discredited the findings. Others have made posts purely speculative with little to no information and people are so quick to jump on the bandwagon.

[regular guy]

Anybody who wants a sneak peak at the updated data, here's the place to find it. I created a wiki-diy to store all of the data. The wiki is a bit more flexible for page formatting than a forum and has the benefit of being a true wiki page where anybody can update and participate with their knowledge.

Here's what I've added to the wiki-diy:
1. Updated clearance specs.
2. Full set of VAC/Calico coated bearing measurements
3. Full set of WPC treated bearing measurements
4. Full measurements of 12 crankshafts
5. Measurement standard-deviations

http://wiki.rcollins.org/core/index....5_Rod_Bearings
http://wiki.rcollins.org/core/index...._Main_Bearings

The wiki-diy is pretty raw, and actually has some errors in it (reference notes are out of date). I'll fix those in the next day or two. There's also graphs to add to the wiki page. Right after I fix the wiki page, I'll update all of the data here as well.

EDIT: Updated main page tables to reflect the same data as the wiki-diy. Only difference, is wiki-diy has all the measurement breakdown and standard deviations as well. That will get added here in the next day or two. But for now, here's the updated tables.

NOTE: These new measurements and clearances reflect the addition of the twelve crankshafts measured. Nominal clearance got slightly better, but worst-case clearance got much worse.

Engine	S65B40
Cylinders	8
Piston Bore	92.00 mm	3.6220 inch
Crankshaft Stroke	75.2 mm	2.9606 inch
Displacement	3999 cc	244.05 sq. inches
Crankshaft	Forged
Conecting Rod Length	140.72 mm	5.5402 inch
Connecting Rod Weight	621.5 grams	21.92 ounces
Piston Weight	487.0 grams	17.18 ounces
Crankshaft Main Journals & Bearings	Metric Dimensions	SAE (Inch) Dimensions
Main Journal Dimensions	59.9758 mm	2.36125 inch
Main Journal Variance	59.9694 - 59.9821 mm	2.36100 - 2.36150 inch
Main Journal Tolerance	-0.0063, +0.0063 mm	-0.00025, +0.00025 inch
Bed plate + Bearing Dimensions	60.0164 mm	2.36285 inch
Bed Plate + Bearing Variance	60.0126 - 60.0202 mm	2.36270 - 2.36300 inch
Nominal Main Bearing Clearance	0.0406 mm	0.00160 inch
Bearing Clearance Variance	0.0305 - 0.0508 mm	0.00120 - 0.00200 inch
Bearing Clearance Tolerance	-0.0102, +0.0102 mm	-0.00040, +0.00040 inch
Crankshaft Rod Journals	Metric Dimensions	SAE (Inch) Dimensions
Rod Journal Dimensions	51.9811mm	2.04650 inch
Rod Journal Variance	51.9786 - 51.9887 mm	2.04640 - 2.04680 inch
Rod Journal Tolerance	-0.0025, +0.0076 mm	-0.00010, +0.00030 inch
Connecting Rod Dimensions	Metric Dimensions	SAE (Inch) Dimensions
Rod L.E. Bushing Dimensions	20.9995 mm	0.82675 inch
Rod B.E. Bore Dimensions	56.0121 mm	2.20520 inch
Rod B.E. Variance	56.0121 - 56.0146 mm	2.20520 - 2.20530 inch
Rod B.E. Tolerance	0.0000, +0.0025 mm	0.00000, +0.00010 inch
Rod B.E. Thickness	18.4150 mm	0.72500 inch
Rod B.E. Variance	18.4010 - 18.4201 mm	0.72445 - 0.72520 inch
Rod B.E. Tolerance	-0.0140, +0.0051 mm	-0.00055, +0.00020 inch
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
Con Rod Clearance (088/089 Bearings) (1)	Metric Dimensions	SAE (Inch) Dimensions
Rod + Bearing Dimensions	52.0179 mm	2.04795 inch
Rod + Bearing Variance (1)	52.0154 - 52.0192 mm	2.04785 - 2.04800 inch
Nominal Rod Bearing Clearance	0.0368 mm	0.00145 inch
Bearing Clearance Variance (1)	0.0267 - 0.0406 mm	0.00105 - 0.00160 inch
Bearing Clearance Tolerance	-0.0102 - +0.0038 mm	-0.00040 - +0.00015 inch
Bearing Clearance per Journal inch		0.00071 inch/inch
Bearing Clearance per Journal Inch Variance (1)		0.00066 - 0.00073 inch/inch
Bearing Eccentricity	0.0216 mm	0.00085 inch
Bearing Eccentricity Variance	0.0178 - 0.0229 mm	0.00070 - 0.00090 inch
Bearing Eccentricity Tolerance	-0.0038 - +0.0013 mm	-0.00015 - +0.00005 inch
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
Con Rod Clearance (702/703 Bearings)	Metric Dimensions	SAE (Inch) Dimensions
Rod + Bearing Dimensions	52.0243 mm	2.04820 inch
Rod + Bearing Variance	52.0230 - 52.0294 mm	2.04815 - 2.04840 inch
Nominal Rod Bearing Clearance	0.0432 mm	0.00170 inch
Bearing Clearance Variance	0.0343 - 0.0508 mm	0.00135 - 0.00200 inch
Bearing Clearance Tolerance	-0.0089 - +0.0076 mm	-0.00035 - +0.00030 inch
Bearing Clearance per Journal inch		0.00083 inch/inch
Bearing Clearance per Journal Inch Variance		0.00081 - 0.00093 inch/inch
Bearing Eccentricity	0.0508 mm	0.00200 inch
Bearing Eccentricity Variance	0.0444 - 0.0584 mm	0.00175 - 0.00230 inch
Bearing Eccentricity Tolerance	-0.0063 - +0.0076 mm	-0.00025 - +0.00030 inch
Rod Bearing Alternatives
VAC (Calico) Coated Rod Bearings (2)	Metric Dimensions	SAE (Inch) Dimensions
Rod + Bearing Dimensions	52.0141 mm	2.04780 inch
Rod + Bearing Variance	52.0090 - 52.0141 mm	2.04760 - 2.04780 inch
Nominal Rod Bearing Clearance	0.0330 mm	0.00130 inch
Bearing Clearance Variance	0.0203 - 0.0356 mm	0.00080 - 0.00140 inch
Bearing Clearance Tolerance	-0.0127 - +0.0025 mm	-0.00050 - +0.00010 inch
Bearing Clearance per Journal inch		0.00063 inch/inch
Bearing Clearance per Journal Inch Variance		0.00054 - 0.00063 inch/inch
WPC Treated Rod Bearings (3)	Metric Dimensions	SAE (Inch) Dimensions
Rod + Bearing Dimensions	52.0281 mm	2.04835 inch
Rod + Bearing Variance	52.0268 - 52.0332 mm	2.04830 - 2.04855 inch
Nominal Rod Bearing Clearance	0.0470 mm	0.00185 inch
Bearing Clearance Variance	0.0381 - 0.0546 mm	0.00150 - 0.00215 inch
Bearing Clearance Tolerance	-0.0089 - +0.0076 mm	-0.00035 - +0.00030 inch
Bearing Clearance per Journal inch		0.00090 inch/inch
Bearing Clearance per Journal Inch Variance		0.00088 - 0.00100 inch/inch

[Malloy]

RG: few questions about what you call defects with the coated bearings:

Originally Posted by RG Wiki

Manufacturing Defects: There's two primary manufacturing defects associated with Calico coated bearings.

Calico coating is also placed on the parting lines. The material on the parting lines will affect the "bearing crush" and may wear unevenly or in a different manner than a standard OEM bearing.

100s of thousands of engines are running bearings with Calico coated bearings.

Calico says on their site, "Calico Coatings proudly partners with ACL, Mahle/Clevite and Durabond to provide the high performing coated bearings for the motorsports industry."

http://www.calicocoatings.com/bearings

How do you conclude their method is defective?

Originally Posted by RG Wiki

BMW bearings are sold in upper (blue) and lower (red) bearing shells. When the Calico coating is applied, this color distinction is lost. The upper (blue) bearings are slightly thicker than the lower (red) bearings. However, Calico habitually mixes these shells up and returns them to the customer mislabeled: uppers are now lowers, and visa-versa.

Since Calico coats thousands of bearings a week, I'm sure they can make mistakes. That said, saying 'habitually' is a bit if a stretch, no? If they made so many mistakes, they wouldn't be the leading supplier of coated bearings to the motorsports industry.

[regular guy]

Originally Posted by Malloy

RG: few questions about what you call defects with the coated bearings:

100s of thousands of engines are running bearings with Calico coated bearings.

Calico says on their site, "Calico Coatings proudly partners with ACL, Mahle/Clevite and Durabond to provide the high performing coated bearings for the motorsports industry."

http://www.calicocoatings.com/bearings

How do you conclude their method is defective?

The effects of coating the parting lines has previously been discussed. However, I'm not an expert on bearing crush, so I don't think I could adequately explain it. I'll let kawasaki and others comment on it. I do know Van Dyne has mentioned it as a flaw in coated bearings. I did speak to Calico guys once on the phone. I was calling to ask them about the thickness of their coating (they graciously pointed me to their FAQ which answered the question). During the same conversation, I also asked them about coating the parting lines. They mentioned they do coat it, but also said the engine builder should remove it before use. I also know Clevite in their catalog specifically calls out coated parting lines in their description of Tri-Armor bearings as why their process is superior to other coatings. I also know Clevite does not use Calico for their Tri-Armor coating.
http://www.mahle-aftermarket.com/mah...ngine-bearings, Clevite Engine Bearing Performance Catalog (EB-40-14), page-15.

We’re particular about parting lines
At MAHLE Aftermarket, we know that bearing crush is critical, especially
in high performance engines. So you can imagine that coating the parting
lines would adversely affect bearing crush and fit. And you shouldn’t
have to sand off material that never should have been applied to these
surfaces in the first place. So we don’t put it there. It’s extra effort to do
the job right, but that’s the only way MAHLE engineers know how.

Since Calico coats thousands of bearings a week, I'm sure they can make mistakes. That said, saying 'habitually' is a bit if a stretch, no? If they made so many mistakes, they wouldn't be the leading supplier of coated bearings to the motorsports industry.

I've personally measured four different sets of Calico bearings, each from different manufacturing lots. Two sets were VAC coated bearings, one taken before a restock, and one taken after a restock (two different manufacturing lots). The other two sets were sent by end-users to Calico for coating, months apart (also two different manufacturing lots). So that's four sets of bearings from four different manufacturing lots. All four sets of bearings were wrong. I also checked with another SoCal installer who has used the Calico coated bearings. He's had the same experience, but he added the caveat that he has seen some measure correctly. He mentioned the importance that he checks everything they do. I know he's watching this thread, so maybe he can comment directly on his experiences.

So for me, with four different sets coming from four different lots and seeing them all wrong, I think it's a fair criticism to call that a habitual problem.

[ticat928]

Originally Posted by Yellow Snow

How does the size between the parting lines compare to a regular lower revving BMW motor such as an M54?

The S65 shells only give a .0016" clearance across the smallest point ie 90 degrees to part line. The clearance across the part line is at least .002" greater, thus giving an oval bore with .0035" - .004"clearance at the widest point.

From this I'm assuming that BMW did some stress analysis at 8,400rpm and found the rod bore stretched by .001". This would mean the bearing bore would be perfectly round at 8,400rpm resulting in a clearance of approx .0025" all around the bearing?

You have a perfect understanding on the reason behind bearing eccentricity.

On S85 engines with over 160,000 miles this big end rod bore stretch has actually been observed (by Troy Jeup) to have taken a permanent set even in an unloaded state.

[regular guy]

Here's some updated clearance specs and measurements. While I was on "vacation," I was busy working on the new Clevite aftermarket bearings. Clevite wanted the crankshaft journal, rod, and rod bearing specifications as close as I could get them. I didn't want to trust this investment to the measurements of two engines, so I reached out to a very experienced S65 builder who documents and blueprints every engine build to ask for a collection of bone stock measurements They are sponsors on this forum and everybody knows them, but I don't know if they want to be identified. So I will leave it up to them to identify themselves in this thread if they so choose.

I had a simple request: go through your blueprinted, bone stock engines and give me the specs of 10 crankshafts. I will combine them with the two I've already measured, and that will give us 12 crankshafts and 96 rod journals to infer the BMW rod journal specifications. A few days later, I received 10 text messages with 10 sets of specifications.

The most important thing I wanted to see was a "peace of mind" comparison to my own measurements. For peace of mind, I wanted to know if all of our hard work would hold up to another independent shop doing their own measurements with their own equipment. Answer: YES! Our measurements matched theirs.

Secondly I wanted to see the spread of measurements to see how far the tolerances went in either direction of our nominal value. This was going to give us hints and evidence in favor of or against the "tolerance stacking" theory. Answer: As everybody will see, these measurements very strongly support the tolerance stacking theory.

Finally, with all of these new measurements, I wanted to see if the specifications we originally generated would stay the same, or change. Answer: they do change slightly for the better, but they also show a much worse case for tolerance stacking than we originally knew.

Raw Measurements:

	Cyl-5	Cyl-1	Cyl-6	Cyl-2	Cyl-7	Cyl-3	Cyl-8	Cyl-4
Engine-1	2.04650	2.04655	2.04655	2.04650	2.04660	2.04655	2.04650	2.04660
Engine-2	2.04650	2.04655	2.04655	2.04650	2.04660	2.04660	2.04650	2.04650
Engine-3	2.04650	2.04650	2.04650	2.04650	2.04650	2.04650	2.04650	2.04650
Engine-4	2.04660	2.04670	2.04650	2.04660	2.04640	2.04640	2.04650	2.04650
Engine-5	2.04650	2.04650	2.04650	2.04650	2.04640	2.04640	2.04640	2.04640
Engine-6	2.04670	2.04680	2.04670	2.04670	2.04670	2.04670	2.04670	2.04670
Engine-7	2.04650	2.04650	2.04650	2.04650	2.04650	2.04650	2.04650	2.04650
Engine-8	2.04650	2.04650	2.04650	2.04650	2.04640	2.04640	2.04640	2.04640
Engine-9	2.04650	2.04650	2.04650	2.04650	2.04640	2.04640	2.04640	2.04640
Engine-10	2.04650	2.04650	2.04650	2.04650	2.04640	2.04640	2.04660	2.04660
Engine-11	2.04680	2.04680	2.04650	2.04650	2.04640	2.04650	2.04660	2.04670
Engine-12	2.04650	2.04650	2.04650	2.04650	2.04650	2.04650	2.04650	2.04650

Engine-1 and Engine-2 were the original two engines used for this article. You can see these measurements are pefectly in line with the ten other sets.

Specifications:

With these new measurements, the nominal measurements and tolerance spread has changed. The nominal measurements and clearance slightly improved, but the tolerance spread got much worse.

Previous Specifications:

	Min	Max	Official	Min Dev.	Max Dev.
Journal	2.04650	2.04660	2.04655	-0.00005	0.00005
Bearing	2.04815	2.04840	2.04840	-0.00005	0.00020
Clearance	0.00155	0.00190	0.00165	-0.00010	0.00025

Updated Specifications:

	Min	Max	Official	Min Dev.	Max Dev.
Journal	2.04640	2.04680	2.04650	-0.00010	0.00030
Bearing	2.04815	2.04840	2.04820	-0.00005	0.00020
Clearance	0.00135	0.00200	0.00170	-0.00035	0.00030

As these measurements show, nominal clearance improved slightly from 0.00165 inch to 0.00170 inch, as did max clearance improve from 0.00190 inch to 0.00200 inch. But due to tolerance stacking, the worst case measurements reduced by 2/10000ths of an inch from 0.00155 inch to 0.00135 inch. Ouch! (Wait until you see how that affects the measurements of Calico coated bearings!)

Tolerance Spread:

To me, this is the most interesting part of taking all of these measurements. Determining the tolerance spread was going to tell me if the "tolerance stacking" theory was a bunch of nonsense, or if it had legs. The "Raw Measurements" presented above were all rounded up to 1/10000ths of an inch, then binned to find the spread. These are the results:

Journal Size	Deviations
2.04630	0
2.04640	17
2.04650	53
2.04660	14
2.04670	9
2.04680	3
2.04690	0
2.04700	0

To me, that table says it all. Just like a bell curve, you see 53 measurements right in the center, and the deviations going down almost in parity at 1-sigma, and fewer deviations at 2-sigma, and even fewer deviations at 3-sigma. There were 3% of all measurements at 3-sigma too large. That means 3% of all crankshafts will have 3/10000ths an inch less rod bearing clearance than nominal. Coincidentally or not, this seems to come pretty close to the number of engines that fail early.

Is the tolerance stacking theory nonsense or not? These are the numbers, you decide.

[Hujan]

Originally Posted by regular guy

Is the tolerance stacking theory nonsense or not? These are the numbers, you decide.

Interesting stuff, RG, and thank you for sharing. One question I have is: What is your definition of "tolerance stacking"? I've seen a lot of discussion as to whether it's a viable theory or not, but I never saw it defined. I probably missed it or it may be too obvious to state, but I thought I would ask.

By the way, I hope you type these posts out on Word and then copy and paste them into M3Post. I can only imagine what would you would do if you tried to hit "Submit Post" after typing all that in and the post failed to load.

[regular guy]

Originally Posted by Hujan

Interesting stuff, RG, and thank you for sharing. One question I have is: What is your definition of "tolerance stacking"? I've seen a lot of discussion as to whether it's a viable theory or not, but I never saw it defined. I probably missed it or it may be too obvious to state, but I thought I would ask.

Basically, tolerance stacking goes like this. Each crankshaft, rods, and bearings have a +/- tolerance associated with them. The parts chosen at random during manufacturing can be +sized, or -sized. We're most interested in those combinations which create a dangerously small oil journal clearance and lead to early engine failure.

Let's say you take a crankshaft with a +deviation journal but still within factory spec, and mate it to a connecting rod and bearing with a -deviation but still within factory spec. This combination potentially creates a dangerously small oil journal clearance and is what we're calling "tolerance stack up" for the purpose of this thread.

I was very intrigued by bone stock motors blowing while others survive for no apparent reason. I first started this thread to bring attention to the journal clearance issue. But as the thread progressed, the tolerance stack up theory came up as a possible explanation for these blown motor. Seeing a 3% 3-sigma deviation could be just a coincidence, or it could also be significant evidence that this is a likely cause of these blown motors.

By the way, I hope you type these posts out on Word and then copy and paste them into M3Post. I can only imagine what would you would do if you tried to hit "Submit Post" after typing all that in and the post failed to load.

I type the text on the fly. I prepare the tables offline. I have an excel spreadsheet with a Visual Basic (programming) back-end that converts the tables into a format compatible with this web site. I have a similar program that does the same thing for the wiki-diy. The only thing I do is cut-paste from my spreadsheet that holds the data into the spreadsheet that runs the program to format the output. Press a button, and viola, I've got instant tables with color, links, and formatting.

[Sephiroth]

RG, i have some Qs which i am sure have been answered somewhere but i am going to ask anyway:

Is there evidence to suggest that even at nominal clearance we see abnormal bearing wear? Meaning BMW's spec was just bad in the first place and any deviation from this nominal (which is natural in any manufacturing process) towards lesser tolerance just makes the symptoms or nature of the failure more disastrous?

There were 3% of all measurements at 3-sigma too large. That means 3% of all crankshafts will have 3/10000ths an inch less rod bearing clearance than nominal. Coincidentally or not, this seems to come pretty close to the number of engines that fail early.

There are also engines that fail later in life too. Does that mean this clearance is changing over time and then has bearing failures? Or that even a 1 sigma deviation from nominal eventually leads to failure?

And to echo the comments of others, your efforts to help out the community are greatly appreciated

[swartzentruber]

Originally Posted by 1love

Is there evidence to suggest that even at nominal clearance we see abnormal bearing wear? Meaning BMW's spec was just bad in the first place and any deviation from this nominal (which is natural in any manufacturing process) towards lesser tolerance just makes the symptoms or nature of the failure more disastrous?

There were 3% of all measurements at 3-sigma too large. That means 3% of all crankshafts will have 3/10000ths an inch less rod bearing clearance than nominal. Coincidentally or not, this seems to come pretty close to the number of engines that fail early.

There are also engines that fail later in life too. Does that mean this clearance is changing over time and then has bearing failures? Or that even a 1 sigma deviation from nominal eventually leads to failure?

Not RG, but I'll take a stab. I'm sure wearing to the point of failure is a function of several things, not just simply clearance. My suspicion is that wear to the point of failure is somewhat highly correlated to how much time your engine spends at the upper RPM range, with maybe a few other things like cold starts, oil change intervals having lesser secondary effects, along with the clearance. A 3 sigma deviation with a lot of higher rpm driving would probably fail fairly quickly in life (i.e. 30k miles or less). However, it would be rather difficult to know for a higher mileage failure if you were a super conservative 3 sigma or an average 2 sigma or a really bad 1 sigma.

RG, do you have any way of differentiating your clearance ranges on these 12 engines for pre-2011 leaded bearings vs the 2011+ lead free bearings? If would be interesting to see if the deviations from standard are better or worse for the two types, particularly since BMW switched suppliers. SFP's data (I know, your favorite guy) suggests a trend of the 2011+ engines to fail earlier, and this could easily be explained if the lead free bearings had a slightly wider spread (more variation), which would suggest a higher proportion of 3 sigmas and more tendency to fail earlier.

By the way RG, nice work and welcome back (this coming from somewhat of a former RG critic). I think this data collection does move the conversation forward.

[BMRLVR]

Originally Posted by 1love

RG, i have some Qs which i am sure have been answered somewhere but i am going to ask anyway:

Is there evidence to suggest that even at nominal clearance we see abnormal bearing wear? Meaning BMW's spec was just bad in the first place and any deviation from this nominal (which is natural in any manufacturing process) towards lesser tolerance just makes the symptoms or nature of the failure more disastrous?

There were 3% of all measurements at 3-sigma too large. That means 3% of all crankshafts will have 3/10000ths an inch less rod bearing clearance than nominal. Coincidentally or not, this seems to come pretty close to the number of engines that fail early.

There are also engines that fail later in life too. Does that mean this clearance is changing over time and then has bearing failures? Or that even a 1 sigma deviation from nominal eventually leads to failure?

And to echo the comments of others, your efforts to help out the community are greatly appreciated

If you don't mind I will try and answer the question you asked RG.

The nominal clearance which RG has deduced for the S65 is on the tight side in it self for a high revving high performance, mass produced engine, especially with the 10W60 oil that is used in the sump.

The other issue with the S65 is the fact that no +\- bearing sizes are available to adjust the final oil clearance due to the inevitable production variance of the rods and journals.

If the S65 used an oil clearance closer to the bearing manufacturer's / engine builders rule of thumb recommendations, of 0.001" / 1" of journal diameter which would be ~ 0.002", a 3 sigma +\- variance would put final clearance in a window of 0.0017" - 0.0023" which would be totally acceptable to me.

Basically, in my opinion the major issue with the S65 is too tight of an oil clearance for a mass produced street/daily driven engine with a redline of 8400RPM. If the S65 were hand assembled with measured and hand selected bearing sizes for each journal so that the clearances were more closely controlled, we wouldn't be discussing this issue right now

[regular guy]

Originally Posted by 1love

RG, i have some Qs which i am sure have been answered somewhere but i am going to ask anyway:

Is there evidence to suggest that even at nominal clearance we see abnormal bearing wear? Meaning BMW's spec was just bad in the first place and any deviation from this nominal (which is natural in any manufacturing process) towards lesser tolerance just makes the symptoms or nature of the failure more disastrous?

For the first part, yes. As a rough guestimate, we've seen 50 engines disassembled and bearing photos posted. Of those 50, all but one of them showed excessive wear. Others may disagree and they are free to do so, but that's how I've seen them.

For the second part, that's more speculation. It's a hypotheses that tight bearing clearance plus tolerance stacking can (and does) lead to early engine failure. So as part of the scientific process, one looks for evidence that either supports it or refutes it. Finding 3% and 3-sigma is the type of evidence that supports the hypothesis.

There were 3% of all measurements at 3-sigma too large. That means 3% of all crankshafts will have 3/10000ths an inch less rod bearing clearance than nominal. Coincidentally or not, this seems to come pretty close to the number of engines that fail early.

There are also engines that fail later in life too. Does that mean this clearance is changing over time and then has bearing failures? Or that even a 1 sigma deviation from nominal eventually leads to failure?

No, I don't think it means that. I would think it's natural for a 2-sigma engine to last longer than a 3-sigma; and 1-sigma to last longer than a 2-sigma, etc. Does it eventually lead to failure? Probably not; or if it does, it takes longer to wear the bearings to the point they do damage. But as others have said, there's other factors at play too.

And to echo the comments of others, your efforts to help out the community are greatly appreciated

Thank you sir!