Don M

The stock cranks in the Dodge/Cummins are Forged Steel. A huge percentage are all Forged Steel. I think a few may have had a cast crank in very low power density engines. The cheapo China stuff I see vendors dealing are all cast, but not a bad part for the money if the HP levels you want is not real high. They use the Vickers system to test their products so expect a low Rockwell test number. A number low enough to fail the Rockwell test by Cummins spec, but once you convert Vickers to Rockwell..its considered OK by many.


The fillets radius used on the flange is best left the way Cummins did it. Good design/spec.

They are fully fillet hardened. Grinding is hard to do correctly without burning the fillets on the grinder. This extends into the journal area and should never hardness test less than a 36 Rc.

Stroking these cranks is doable, but should remain a job of a shop with experience and the knowledge to get it right.

Never exceed 325 rpm when you balance the crank. There is enough mass there to whip the crank around and change it on the machine.

Do not use bob-weights. The stock throwing index provides inherent balance already.

Many of the high powered engines also have a shot penned fillet area. If you grind it away with stroking or just undersizing the journals....you have to re-peen that area again. Not all shot peening is the same. Be certain it is done by a shop listed by Cummins. There are only 10 in the country. These shops have been provided the proper shot and process guidelines to do the work the correct way. If yours is not shot peened in the first place....you have not to worry about getting it done at all.

Don~

GO 4LO

Thanks for the info, Don. Are these issues that a crankshaft builder will need to take into account when building a billet crank? i.e. should the stock "fillet radius used on the flange" be stronger than an aftermarket fillet of different radius, if you're starting from a blank billet, or are you speaking with regards to offset grinding the stock crank?
When you say "Stroking these cranks is doable, but should remain a job of a shop with experience and the knowledge to get it right. " Do you mean its physically possible to build a crank that is stroked, or that it can be made to run well? Do you know of any shops that have experimented with increasing the stroke on the Cummins B?
Thanks again for all the info, everyone!
Chris

HOHN

I don't know of Winberg's web presence. It's also weird to me to see Crower and Scat mentioned as a comparison. I've been led to believe that Scat cranks are much cheaper. I don't have any experience with Crower OR Scat, so I'd recommend you find someone who has for comment on that.

The Winberg cranks are used by the biggest name NASCAR, Pro Stock, Top Fuel, teams etc. It's big $$ to get them to do your work.

Henry Velasco is the only other crank "guru" that I would look up for a serious high end project. I know there are a lot of talented crank grinders out there, but Velasco and Winberg are among the most respected, if not them most.


Loopy and jet-lagged in Germany,

Justin

GO 4LO

I didn't mean to offer those two companies as an alternative to Winberg. It's just these are the only two companies I've been able to find on the internet that will custom build a billet Cummins crank. Do guys like Winberg and Velasco do one-off work, or is it more like they're each associated with a specific nascar team and only do work for them? Thanks again for all the info, and if you want to wait till you get back Stateside to answer, I guess that'll be okay.
Chris

Don M

A custom crank grinder might just grind his own fillet radius and match that to a set of bearing that would be custom as well.

The Cummins fillet and bearing design are IMO the best way to go.

A one off crank is not cheap. To forge something like that will cost in the "hood" of probably 5000 bucks or more.

Not all forged cranks are alike either. Some are forged and twisted into shape. Twisted cranks are generally weaker than a crank that is forged into shape.

The stock Cummins crank is a sweet piece. Its forged and stronger than you might think.

GO 4LO

So given all this info above, what would be the ideal setup for a dragracing (and occasional sledpulling) Cummins:

Displacement numbers assume 10thou overbore.

4.37"___5500rpm___334cid___1.75
4.52"___5250rpm___346cid___1.69
4.72"___5100rpm___361cid___1.62
4.92"___4900rpm___377cid___1.55

Not sure how to calculate the differences in cylinder wall loading between these setups, though. Any help?
Chris

GO 4LO

By the way, the number correspond to a 4000 fpm max piston speed and were originally calculated by Missouri Mule in a previous bore/stroke post.
Chris

HOHN

You can calculate the cylinder wall loading based on the the BMEP that you would assume and the rod angularity. Say, for example, that you have an angle of 15° between the connecting rod and the bore centerline when the crank is at 90°. If you assume that the the cylinder pressure is 2000psi at this point, you would have the 15° vector of that component acting on the cylinder wall.

Picture a right triangle where the crank stroke is one leg, the piston height is another, and the rod is the hypotenuse. If cyl pressure is 2000psi, then you divide by the piston surface area to give you the FORCE acting down on the piston. At a 15°, 96.6% of the piston's downforce acts along the rod. The remaining 3.4% goes into the cylinder walls as friction's normal force.

This is just the Cosine of the angle between the rod and the bore centerline.

At 25°, the side loading goes from 3.6% to 10%.

At 45°, it's 29.3%

At 60°, it's 50%!

So you can see how it rapidly gets REALLY bad as you get into the extreme rod angles I don't think that any engine uses an angle over 35°, but I could be wrong.

Short rods cause more wall loading.

Justin

HOHN

To visualize loading, picture a sine wave with 90° at the top and 0° at the bottom. The rate rapidly passes from one extreme to the other. You'd think that the 50% loading point is 45°, but it's not-- it's 60°! 45° is still 71% of your initial value (or conversely, 29% loading).

Or visualize the "bell curve" that you've seen before, representing a "normal" distribution. It's also similar. As you head towards the extremes, things change drastically.

jlh