If you keep souping up your engine, eventually you will find a weak link.  In my case, the first weak link was the clutch.  In my never-ending search for power, the clutch started to slip.  Thankfully, nothing was damaged.

A quick inspection revealed that all the clutch components were in factory-new condition.  All the fiber and steel plates were in spec, and there was zero evidence of overheating.  The springs had the same force as they did at 1000 miles, and their free-length was in spec.  The release system was adjusted properly and there was adequate free-play.  The poor little clutch just couldn’t hold the power.

The immediate fix for my slipping clutch was to install a Barnett clutch kit.  The Barnett kit fixed the problem, but it did so by using springs that are about 13% stiffer than stock.  Those stiffer springs increase stress in the release mechanism.   I would prefer a solution that does not increase stress in a failure-prone component.

It was time to take a hard look at the Savage clutch.  What can be done to beef up this weak link?

When the release cam fails, the damage won’t necessarily be confined to the cam itself.  The cam fails, falls apart, and proceeds to wreck up other stuff.  Like this release plate.

The stock Savage clutch is a typical motorcycle wet clutch.  It uses a total of eleven plates (six fiber drive plates and five steel driven plates).  There are four coil springs that clamp all the plates together and lock up the clutch.  In general, there are three ways to increase the torque limit on the clutch.

Increase the clamping force (i.e., stiffer springs)

Increase the coefficient of friction (i.e., change the friction material)

Increase the surface area of the friction material (i.e., more plates, larger plates, etc.)

I have seen ideas posted on the forum regarding clutch mods.  The addition of another friction plate, adding a steel plate, sanding down plates to allow room for an extra plate, replacing the special outer friction plate with a standard friction plate, eliminating the wave washer, shimming springs, stiffer springs…..  Did I miss any?

We have this crazy powdered-metal release cam that breaks unexpectedly.  Get the geometry a little out of whack and you end up with a snapped release cam.  It’s usually a long walk home.  Any modifications must consider the release cam.

I initially figured that the broken release cams were the result of coil-bind and/or poor clutch pushrod setup.  After a closer look, I think there might be a little more to it.

The release mechanism has a 7:1 ratio, meaning for every 7mm that the lever arm travels the release cam pushes the pushrod 1mm.  I checked the release travel several ways to get an idea how much the clutch springs would be compressed when the lever on the handlebars was pulled all the way in.
 
First adjust the clutch cable to remove all the slack.  Then loosen all the clutch cover bolts.  Then pull the clutch lever on the handlebars all the way in and lock it in place.

I double checked the release travel with the clutch cover removed.  When the release lever was moved ½” (12.7mm), the release cam pushed a dial indicator .065” (1.65mm).  Pretty close.  I think we have that piece of data nailed down.

If you assemble the clutch with a couple of light springs, you can cycle the pressure disk with your fingers.  It makes it easy to check a few things.

When I push in on the spring bolts, I can move the pressure disk inward until it hits the stops in the clutch basket.  Then I can use feeler gages to measure how much it moved.  On the stock clutch, the pressure disk moves about .157” (4mm).  That’s well shy of coil bind, and way more than the available release mechanism travel.  Nothing will interfere with clutch operation.