Originally Posted By: PeterPolyol
I used to say yes more torque = higher visc by default before I considered that there are new cars out there pushing 77ish ft-lbs per crankpin at 1700rpm using 0w20 oil.
Again, you can't just generalise across platforms and designs, and you can't use one manufacturer's design to justify an oil selection in another...and as the OP said "all else being equal"
Back to the bearing design curves...the dimensionless number along the bottom axis is the "Somerfeld Number", or "Bearing characteristic number"
So = (r/c)^2 x u x N / P
r = shaft radius
c = shaft/bearing clearance
u = Kinematic viscosity (must be high shear kinematic for multigrades)
N = shaft speed in RPS
P = the vertical load/the projected area of the shaft.
The first point to notice is the massive difference in MOFT between the line l/d=1/4, and l/d=1/2, that's the bearing axial length to diameter ratio, and as can be seen, going from 1/4 to 1/2 doubles the film thickness..."all else being equal" in the Somerfeld Number.
That's a mechanical parameter, dictated by engine design.
The next mechanical parameter is (r/c), the radius/the vertical clearance...
increase the radius, you increase So, and push up MOFT.
reduce the clearance, you increase So, and push up MOFT.
Honda have offered in some of their papers that they are doing all three, bigger, longer bearings with less radial clearance...actually measures that increase friction, but the premise is that they are gaining more in reduced piston and skirt friction.
You can't do the above in isolation in the design...longer bearings mean that a lower level of crank/block flex causes edge interaction, and wear. Smaller bearing clearances similarly mean that less flex is required to create edge interaction and wear.
So the modern OEM is going for a deeply skirted (rigid) block, often crossbolting or girdles to reduce block flex, and the bigger journal diameters in turn stiffen the crank.
Once you've got your engine design parameters fixed (that's the all else being equal), the designer has load, speed and viscosity to contend with.
for 40% more power, it can be 40% more torque at the same speed, or the same torque at 40% more speed.
The former will DROP the Somerfeld number by 30%, and push the MOFT to the left, i.e. smaller, the latter will INCREASE the Somerfeld number by 40%, and increase the MOFT (both cases ignoring inertial effects, but simplified to enhance the picture).
Thus my recommendation "all else being equal", on a V-8 fitted with a supercharge that has more torque at the same RPM...a high revving Astra engine in a 500kg car can get away with a lot lower viscosity than an engine that can pull hard from off idle.
Gokhan's introduction of the Stribeck curve...
is the epitome of "all else being equal"...the r/c, and the l/d have been left out, and only the viscosity, speed and load interaction have been left.
If the OAM has designed it in the RHS region of full hydrodynamic, using those ratios will keep it pretty much there.
If they have designed it at the low point, to provide acceptable life, as many are doing these days, changing the viscosity will keep it at the design point.