GPW Engine rebuild - crankshaft issue

[YLG80]

As advised by artificer, I opened a new topic about that issue.
The discussion has started here :
viewtopic.php?f=4&t=257077&p=1543134#p1543047

Issue :
GPW engine block dated from July 1942. (Already rebuild once - serial number grinded and re-stamped # 512133.
Crankshaft seized when pre-assembling the bearing without the rear seal.
Cannot tighten the middle and rear bearings to the torque without blocking the crankshaft.

@ Joël:
Here are the pictures of the bearing caps starting from the front bearing.
The casting numbers are oriented in the same direction.
Front bearing cap = GPW-14 6329

Front_bearing_cap.JPG (106.26 KiB) Viewed 1957 times

Middle bearing cap = GPW-9 6330

middle-bearing_cap.JPG (134.01 KiB) Viewed 1957 times

Rear bearing cap with A mark crossing the block and the cap. No casting number on top that part.

rear_bearing_cap.JPG (148.77 KiB) Viewed 1957 times

The front and central bearings have two marking lines on top of the F script.
Both are oriented towards the front of the engine block.
I'm not sure how these caps could be assembled incorrectly.
The top and bottom shells oil channels have to be in line as well as the oil holes and they cannot be assembled incorrectly due to the pin.

@Gindi:
This a GPW block.
I've given a phone call to the shop.
They tell me that the ground bores are likely slightly out of alignment.
This would cost about 200€ to rebore.
So they propose what you mentioned in an earlier post : add tiny metal shims under the middle and rear bearing caps.
I guess it's going to be tough to find tiny shims.

I've measured the crankshaft with a digital meter. Unfortunately it's not accurate enough (3 digits) to make absolute measurements.
So only relative measurements are relevant.
Front journal : 2.295'' (3 measurements - edges and middle)
Middle journal : 2.295" (3 measurements - edges and middle)
Rear journal : 2.296"" (3 measurements - edges and middle)
It looks like the machinist did a perfect job.
There was a metal plate attached to the crankshaft with .040'' for the main journals and .050 for the connecting rods.

However I'm questioning myself about what he told me when I picked up the crankshaft.
One of the journals was originally grounded to a different dimension than the two others.
He did not like that and machined the 3 journals @ the same dimension.

When I check the old bearings, I cannot find their dimensions.
Only the following markings are visible on the shells:
Front bearing :
Fm
4 51
Central bearing
Fn
2 5
Rear bearing
Fm
2 51

[YLG80]

Hello Gindi,
Thanks for your very interesting point on the grinding stone radius.
I'm going to check that.
The measurements given above have been taken a 3 different places : left, middle and right.
I've found exactly the same readings @ the 3 locations on each journal.
But I will check if I was close enough to the edges.
Now I better understand Joël remark on that aspect of the problem.
Yves

[YLG80]

What bothers me is that the middle bearing shells (upper and lower) and the rear bearing shells (upper and lower) are equally "polished on the edges" .
If there was a misalignment, I would have expected the shells to be polished only on the upper or the lower side or on one edge.
The almost perfect symmetry looks strange to me.
Yves

[Joe Gopan]

Standard Crankshaft Size is 2.3331/2.3341 in inches for the Main Bearings.
It appears that your crankshaft is now 0.040" undersize and the factory tolerances for 0.040" undersize is 2.2931/2.2941"
NOTE:
The measurements you have given indicate the journals have been ground slightly oversize.

Here are the dimensions for the wall thickness (shell thickness) for 0.040" undersize bearing shells
#1 = 0.1223" (Max.)
#2 = 0.1173" "
#3 = 0.1173" "
A ball micrometer is needed to check the bearing shell (wall) thickness because of the curved shape.

Below are the Overall Lengths of the bearing shells.

#1 = 1.921"
#2 = 1.813"
#3 = 1.750"

[YLG80]

Correction : the above wall thickness values are not correct.
Bearing wall thickness must be calculated based on the STD wall thickness + the oversized value DIVIDED by TWO)

It's amazing how documented you can be ... and how you can retrieve the information so quickly.
Thanks a lot for these dimensions.

#2 and #3 shell lengths are OK
As far as I can measure the thickness, they are just below your dimensions with 0.0015''.
But I agree: it's difficult to measure and I will do that with a comparator + a ball termination.

But that point is very interesting, because it's related to something I've found with the shell dimensions.

I've made some progress towards the solution, obviously w/o adding shims.
That's really a "Sherlock Holmes job".

I've examined the old shells and compared them with the new shells.
I've also compared the assembly with the old shells + caps and new shells + caps.

I've immediately found a strange mark on the cap side of the middle and rear top shells.
So I've double checked the dowel pin position and I've found both of them above flush.
The same mark exists on the old shells (0.030'').

bearing_dowel_pin_above_flush.JPG (88.12 KiB) Viewed 1893 times

The picture shows the pin after correction, below flush
That mark is on the new shell (0.040'')

That issue was easy to resolve and now the pins are OK.

Analyzing that issue a little further, I've found something odd with the shells.
If I fully torque bearings #2 and #3 in their respective positions and then disassemble them, both ends of the shell are not flush with the bearing cap surface near the bolts.
So both shells are pushed very hard against each other when torqued.
I think they should be flush with cap surface, isn't it? The torque should be applied to the caps, not the shells, I guess.
That's indeed the case with the old shells which are flush.

Do I have to somewhat grind both ends on a marble to have them flush with the bearing cap surface ?

That crankshaft job is really a delicate job !
Yves

PS Shells are from Federal Mogul.

[Joe Gopan]

Here are the diameters of the housing bores for the main bearings measured with the caps in place and torqued to specification.

#1 = 2.4995/2.5000"
#2 = 2.4895/2.4900"
#3 = 2.4795/2.4800"
I can do this in metric, if need be.

[artificer]

Sorry this was being typed when Gindi posted, so there may be some duplication....

So both shells are pushed very hard against each other when torqued.
I think they should be flush with cap surface, isn't it? The torque should be applied to the caps, not the shells, I guess.

Many mechanics & non mechanics miss this fact totally!
The bearing shells should not be flush.

The correct term for not fitting flush is bearing crush.
Not fitting flush, is something that is supposed to be measured on all mains & con rod bearings when re-assembling an engine, it doesn't take long & is done in the same operation as checking the bearing clearances with plasti/flexi gauge.
Checking bearing crush is achieved by tightening down both cap bolts to torque then releasing one bolt & measuring the gap between the cap & block or rod as the case may be with a feeler gauge or other more sophisticated measuring device.

Crush ensures the bearing halves are held very tightly in their respective bores for good shellback contact & heat transfer.
None or not enough is why bearings eventually spin in their bores &/or fail.

Too much crush will cause the bearing to distort & grab the crank....BUT....in your case it should be the same on all mains not just 1 or 2. That said check it & let us know what you actually have. Having too much crush will not necessarily be evidenced on a plasti/flexi gauge reading.

The measurement for crush should be 0.001" per 1" of diameter so you should have about 0.004" on main bearings & 0.002" on big ends.

Something mentioned earlier that should be clarified:
Steel backed, layered slipper bearings should not be scraped under any circumstances to fit cranks better. Scraping was done on poured babbit bearings, not slipper replaceable types.

http://www.musclecardiy.com/performance ... ngs-guide/

[artificer]

Yves if this link works go to page 128 & you might like to go through the whole section on crankshafts...JG.
http://www.slideshare.net/mcfalltj/chap ... _search=42

Also in this one about page 172 on
http://www.slideshare.net/mcfalltj/hald ... m_search=5

[YLG80]

Yesterday afternoon, I've read many posts on that bearing crush.
Thanks to your information and the links provided here above, I now understand what it is.
I'm really surprised that a shell can support such a high pressure.

The crankshaft seizure occurs when I torque the bolts above 60 ft lbs.
So it's likely related to bearing crush.

Already done :
Resolved the pin partially above flush issue. (#2 and #3 caps)
Inspected and polished the bearing caps surface.
Observed that the seizure occurs at the edges of the shells.
Measured the crankshaft. Looks OK but slightly oversize.
L shells are located in the caps - U shells in the block.
Measured the shell thickness against Joël dimensions.
Rear bearing #3 : 0.107'' . OK
Middle bearing #2 : 0.110'' . OK

To do:
Measure the internal bore diameter with bearings @ torque as proposed by Joël.
Double check my torque wrench against another one.
Test bearing crush as advised by Artificer.(Will use a feeler gauge.)
Try to find why bearing #1 does not cause any problem.
Measure again the crank journals at the edges. As the crank has been grounded slightly oversize, I suspect kind of a stack-up tolerance problem.

BTW I've found a very interesting document about bearing problems by Mahle.
http://www.mahle-aftermarket.com/media/ ... ochure.pdf

[artificer]

Observed that the seizure occurs at the edges of the shells.

Joel brought up the crank fillet radius [grinding] & you need to eliminate this, before going off chasing irrelevancies. Always remember, the most simple, no real cost issues are those overlooked.

[YLG80]

@ Artificer,
Yes I've doubled checked the radius.
There are indeed radius near the edges on the crank journals.
But they are by far outside the working surface.
There are chamfers on both sides of the shells.

I've also made your recommended crush test with the feeler gauge.
1-fully torqued the middle bearing with the shells but without the crankshaft.
1-released one side
2 checked the gap between the block and the cap with a feeler gauge.
I've found much higher values than 0.004''.
The 0.008'' feeler passes through form one side to the other side above the bolt hole.
0.009'' does not pass through from one side to the other.
That's twice your calculated value. Over crushing ?
[Correction: Sorry, I guess I've to divide that value by two, so it's correct]

I've also measured the middle bore with shells while fully torqued.
As expected I've found lower values than the crank journal diameter. About 0.0025''
Same value in 3 directions.
That's about the clearance I should find for a correct working.

How can I determine if the bearing caps have been reground or not in the past?
#1 has curved grinding traces.
#2 and #3 have crossed straight lines

[Joe Gopan]

If I remember correctly, The existing crankshaft main journal diameter measures greater than the factory spec, this indicates the main bearing journals are incorrectly ground and from the appearance of the edge of the main bearing shells the radius is in question. Also the "housing bore" or the internal diameter of the bore of the installed main bearing caps is done without the bearings in place.There is way too much detective work being done on such a simple topic. Did you compare the housing bore with the factory specification?

I have attempted to find your measurements for the reground main bearing journals but am lost in this maze of information overlload. Can you repost the diameter of your reground crank shaft and compare it with the manufacturers size for a crankshaft reground to 0.040" undersize. (2.2931/2.2941)

[YLG80]

You are right, there are so many informations that it's even difficult for myself to remember what I did before !

Using direct digital reading in inches (published in a previous post)

Front journal : 2.295'' (3 measurements - edges and middle) - (2.2931/2.2941)
Middle journal : 2.295" (3 measurements - edges and middle) - (2.2931/2.2941)
Rear journal : 2.296"" (3 measurements - edges and middle) - (2.2931/2.2941)

Using another metric analog micrometer and conversion:
Front journal : 2.2932'' - (2.2931/2.2941)
Middle journal : 2.2933" - (2.2931/2.2941)
Rear journal : 2.2933" - - (2.2931/2.2941)

The second measurement confirms what you mentioned before: The journals are on the oversize side.
But taking into account the instrument precision, I cannot blame the machinist !

I'm going to make another plastigage test, cutting the plastigage exactly to the journal dimension.
This should show if the radius is incorrect at the edges of the shells.

[dpcd67]

Can't blame the machinist? Why not; it is his responsibility to get the grind right. That is why I pay them.

[YLG80]

Here are the plastigage pictures:
Rear bearing:

rear_bearing_plastigage_radius_test.JPG (79.59 KiB) Viewed 1245 times

Middle bearing

middle_bearing_plastigage_radius_test.JPG (67.16 KiB) Viewed 1245 times

It shows a minimum oil clearance of 0.001'' in the red wide areas.
Perhaps someone else can detect something wrong from these pictures?