The purpose of scraping new bearings upon installation into a motor is NOT to make them rounder or smoother or to the correct radius (as some people seem to think) .... the bearing manufacturer is perfectly capable of doing that as long as correct specs are used.
The purpose of scraping bearings upon bearing installation into electric motors is to compensate for any misalignment between as-installed position of the shaft/bearing induced by misalignment between the two bearing housings or perhaps gravity sag of shaft. (sleeve bearings are not particularly tolerant of misalignment, especially as ratio of length to clearance increases).
If you have misalignment in the horizontal plane between bearings of horizontal motor, then you may see diagonal pattern. (the angle shows up much more dramatically than you’d think... if you have 3” diameter shaft with for simplicity 3 mils clearance and a horizontal angle misalignement of 3 mils per three inches, then angle you see is NOT 1/1000 which would be invisible.... it is closer to 1:1 since the pattern goes from one side of bearing to the other over the length of the bearing). [I added a new message below with example of this].
If you have misalignment in the vertical plane between bearings of horizontal motor, then the blue-check/dry-roll check pattern will be only on one end of the bearing, and not the other. That is the idea of the 80% contact spec for blue pattern/roll pattern.... it should be at least 80% of the LENGTH of the bearing (NOT the width!)
There seems to be a common mis-conception that 80% means you want 80% contact over the entire bearing (including width). That is just plain silly because the journal has smaller radius than the bearing and digging out the bottom of the bearing to form similar radius of curvature for full width would destroy the ability to form a convergent wedge.
How common is this misconception? See attached excerpt from article “Sleeve Bearings: A Modern Use for an Old Technology” by Richard Nailen (normally very accurate imo), published in Electrical Apparatus magazine, February 2008. On the 2nd page of the pdf, he says
[Nailen (2nd page of pdf):]
In assembling the bearing, then, the essential step is what’s sometimes called “blueing and scraping.” The make sure the babbit contacts the journal as uniformly as possible throughout the contact angle of Figure 2 calls for a fitting procedure like that given in the accompanying box [Figure 15].
Figure 2 is on the first page of the pdf and shows contact angle extending for 120 degrees centered on bottom of bearing.
The accompanying box [Figure 15] is dead wrong. The “before” of Figure 15a looks like pretty good contact to me. The after Figure 15b is full contact (ouch!). The caption for figure 15b says:
[Figure 15b caption, 2nd page of pdf]
Final fitting accounts for the difference in radius between the two curved surfaces, leading to the more uniformly distributed contact at (b).”
All I can say is ouch.. dead wrong. The difference in radius is a design feature and we don't want to change that.
He goes on to talk about various repair specifications with imprecise terms [I agree], but he never tells us that Figure 15 is dead wrong and in fact the entire context of how he refers to figure 15 and the caption of Figure 15 suggest this is supposed to tell us the correct pattern in Figure b.
I think I have an idea why there misperception that you want 80% or 90% over the entire bearing (rather than just a narrow strip down the center whose length is 80 – 90% for vertical alignment as it should be). When machinists work with FLAT surfaces they use percentage of contact of the ENTIRE AREA as determined by blue check to describe how flat and smooth the surfaces are. Sleeve bearing 80% contact should never be confused with that type of check... it is completely different. The after shown in Figure 15b would kill a bearing.