All references to the tensioner are from the perspective of a person standing in front of the engine compartment, looking at the tensioner. Thus, right/left and CW/CCW are opposite to the perspective of someone siting in the driver's seat.

The subject serpantine belt tensioner is from a 1992 300E, which only has 35,000 miles.

The tensioner consists of an aluminum housing, which has an arm for the pulley and an aluminum inner bushing whose axial movement is limited by snap rings, The front face of this bushing has a steel insert, which contains the "rachet" for the pointer and is the load bearing surface for the lock bolt load against the mounting strut (the "crowfoot bracket" or "Y-bracket") Looking at both the front and rear faces of the tensioner, from the inner aluminum bushing to the housing you observe the inner bushing, a steel ring, a rubber ring, another steel ring, and finally the housing. The steel ring/rubber ring/steel ring visible on both ends are sealed ball bearings that keep the inner bushing firmly fixed, radially, relative to the housing. Viewed another way, since the 19 mm hex lock bolt fixes the inner bushing to the engine structure, the sealed ball bearings fix the housing radially and the snap rings fix it axially, but allow it to rotate relative to the inner bushing, which swings the tensioner pulley in an arc of about 2.75" radius.

Job number 13-3452 describes a tensioner functional test. It impllies that in the unloaded state, imaginary lines from the two flats on the back side of the inner bushing should pass equidistant on each side of the pulley axis. (Confirmed by KermitF observing a new tensioner - thanks!) This should be the unloaded state of a good used or new tensioner. To test, mount the tensioner in a vice with soft (hard wood or equivalent) jaw inserts clamping the flats. Then, using the pulley as a hand hold, rotate the housing about 30 degrees off line in each direction, which should take a good deal of force. (This is the approximate angle the pointer moves through during tensioning, which indicates the relative angular displacement between the inner aluminum bushing and housing). If the pulley does not return to its original position after being displaced this angular amount in each direction, it is not serviceable. DO NOT USE IT!

On the car, tensioner failure is indicated if the pointer does not move back to the bottom of the ramp as belt tension is released - assuming that the pointer started out at the top of the ramp or just beyond. Such a tensioner is suspect, and it should be removed for the above test and be prepared to buy a new one. Using a failed tensioner will just cause you more heartache and may damage other parts if you attempt to use it to tension the belt.

Since this tensioner had failed - the center bushing could be rotated to any position without springing back - I decided to disassemble it and figured that this could not be done non-destructively.

After removing the snap rings I used a hammer and wood block to knock out the front bearing by pounding on the back side of the inner aluminum bushing. Then I hammered on the front side and knocked out the rear bearing and inner aluminum bushing as a unit.

What is contained inside the housing - between the inner bushing and housing and the two sealed ball bearings is the "tensioning element" - a big rubber bushing. Of course this was destroyed by my disassembly, but it was already completely broken internally before I started.

It works like a suspension bushing. As the 13 hex mm tensioner nut is rotated CW to the point of taking all the slack out of the belt, continued CW rotation of the nut continues to push the tension rod down (remember, they are LH threads), which rotates the inner bushing CCW while the housing rotates very little due to belt stiffness holding it in place. All of the relative rotational displacement between the inner aluminum bushing and housing is absorbed by bushing deflection or what I will call "strain" - just like what happens when you twist or bend a rubber eraser - and the greater the torsional strain of this rubber bushing, the more torsional resistance it offers, but it's probably highly non-linear.

Suspension bushings work the same way, HOWEVER, the bushing bolts should only be torqued at normal ride height so that the bushing is not strained in torsion when the car is sitting still. Bushing torsional strain only occurs with suspension movement while the car is moving. This is also the reason why you should not store a car on jackstands as the bushings will be strained in torsion, which can damage them. Permanent full time strain also damages the tensioner rubber bushing. In fact, failure is virtually guaranteed!

Rubber is subject to "creep" - strain it long enough, and it will not return to its original shape when the load is removed. Also, rubber is subject to degradation from thermal effects, and this degradation increases exponentially with temperature. Tensioners that see a lot of low speed drive time in hot weather, which usually creates the highest underhood temperatures are not going to last as long as one that sees primarily highway speed driving, and this car spent a lot of time in Palm Springs.

The bottom line is that this design belt tensioner is a recipe for failure. The individual parts are not cheap or shoody. In fact they appear to be of very high quality. The problem is that using a torsionally strained rubber bushing to maintain belt tension is just a piss-poor design architecture, and I'm utterly dumbfounded that Mercedes used a rubber bushing instead of a mechanical spring tensioner, which, if properly designed, should last virtually forever.

It's also too bad for us that the aftermarket has not come up with a substitute mechanical spring design.

Failure of this rubber bushing - as indicated by the pointer not moving down the ramp as belt tension is released can lead to consequential damage when you install a new belt and attempt to tension it. (The tensioner can also fail in service - either slowly due to creep or thermal degradation or suddenly if the rubber breaks.)

When you begin tensioning the new belt, the inner bushing starts out rotated farther CCW than if the tensioner was good and the inner aluminum bushing returned to its original unloaded angle relative to the housing. As a result, attempting to tension can cause the tensioning rod to interfere with the inner aluminum bushing as you continue to rotate the bushing further CCW, which will bend the rod. In extreme cases you could even run the nut off the top of the adjuster rod thread. (Remember, it's a LH thread so as you turn the adusting nut CW it pushes the rod down, which rotates the inner aluminum bushing CCW.)

So the bottom line is observe that pointer as you remove belt tension. If it's not between the top of the ramp and the thick solid line, set it there, then watch to see that the pointer slides down the ramp as you release belt tension. If it doesn't, replace the tensioner. DON'T attempt to reuse it!

To be sure, set it up in a vice and run the above mentioned test. Observe the radial position of the inner aluminum bushing and housing by drawing those imaginary parallel lines from the flats toward the pulley. Are they equidistant from each side of the pulley axis? If not you've probably got a paperweight. Run the angular deflection test anyway, and I'm sure you will find that it fails.

Buy and install new tensioner.

Duke