Softop

Miata Forum Member
Posts: 230
From: Australia - Classic Red '99, Loch Stewart system, mud spats and K&N Drop-in, Magnecors and lowered front plate Registered: SEP 1999

posted 10 August 2000 18:31

---

The inexpensive fix will work as long as damage is limited to the keyway and the crankshaft is not bent or cracked. The right epoxy is a good tool for rebuilding the keyway - but it doesn't address the reason that it failed in the first place. The fact that "none have failed yet" is not decisive - remember, originally some of these engines went 100,000 miles without failure in the first instance. The company I work with, Loctite, are acknowledged around the world as the experts at this type of operation. This is the fix that our engineers came up with:

Read Lance Schall's article at http://www.miata.net/garage/crankshaft.html

Read Skip Cannon's article at http://www.socalm.org/crank/
6/13/2002: editors note: the socalm link is broken.

Disclaimer: I work for Loctite. I suspect - but have not been able to prove to date that a retaining compound of some kind was used in the original assembly. If so, its failure during service or absence in subsequent re-assembly would likely be the root cause of the problem.

1. Dismantle the crankshaft and pulley and check for damage. If there is no damage then you can re-assemble. TIP - This is a good time to change the oil seal. Use a new bolt and a new key if the old one is in any way damaged.

2. Without inserting the key, slide the pulley onto the crankshaft and determine the bore clearance - plastigauges are good for this.

3. If the diametrical clearance is less than 0.2mm put a coating of Loctite 242 or Loctite 243 threadlocking compound around the shaft where the pulley seats. If the gap is greater than 0.2mm then coat the shaft with Loctite Quick Metal instead.

4. Re-assemble using a new key (make sure it is fitted in the right direction). Use a new bolt. Before inserting it put a couple of drops of the Loctite 242 or 243 on the threads and torque to specification.

5. DO NOT START THE ENGINE FOR 24 HOURS.

If when you dismantle the engine you find damage which is limited to the keyway as shown in Skip Cannon's article then here are the steps you should take:

1. As far as possible using a small file clean up the keyway damage, removing burrs etc. from the area where the pulley seats. Take care not to file the undamaged side - this will be your reference for correct alignment during re-assembly. If the pulley keyway is damaged then it is best to use a new one. Using emery tape (not to fine because a coarser surface works better with this repair method) linish the crankshaft to a smooth slip-fit.

2. Clean the bore of the pulley and mating shaft area very thoroughly using a cloth and solvent or methylated spirits and let dry for 5 minutes.

3. Without inserting the key, slide the pulley anto the crankshaft and determine the bore clearance - plastigauges are good for this.

4. Squeeze a little Loctite Quick Metal into the keyway and fit the key. Fill the damaged part of the crankshaft keyway with Quick Metal.

5. If the diametrical clearance is less than 0.2mm put a coating of Loctite 242 or Loctite 243 threadlocking compound around the shaft where the pulley seats. If the gap is greater than 0.2mm then coat the shaft with Loctite Quick Metal instead.

4. Re-assemble using a new key (make sure it is fitted in the right direction). Use a new bolt. Before inserting it put a couple of drops of the Loctite 242 or 243 on the threads and torque to specification. Remember to keep a clockwise pressure on the pulley during assembly to ensure correct alignment. Torque bolt to specification (80-87 foot pounds).

5. DO NOT START THE ENGINE FOR 24 HOURS.

Author's note: Keyways of this type are normally used for the purpose of alignment. Assembly forces are usually transferred to other areas such as tapers - which from the photo's seem to be absent. This means that the force locking the assembly is the clamp load exerted by the bolt then holding the pulley against the shaft shoulder and any interference between the pulley and the shaft. Even a partial loss of this clampload (bolt stretch, thermal cycling, vibration) would lead to the entire load being transferred to the key - and subsequent key failure. It would seem that this design is only good engineering practice if a retaining compound is added during assembly. That's what I suspect originally shold have occurred and that's what I have proposed.

The Loctite 242/243 or Quick Metal that you have added to the shaft cures to a tough thermoset polymer between the mating faces, keying into machining marks and micropores of the metal. This process turns the whole of the mating face area into a keyway (after alignment you could actually remove the key altogether and it would make no difference to the performance of the assembly). This assembly will dismantle with the standard puller.

The threadlocker on the bolt is also very important. Torque wrenches work on a theory called torque/tension relationship. Basically this theory determines that by applying a certain amount of twisting force to the fastener, a predictable amount of clamp load will be generated. This is only true if friction can be controlled. The quoted torque specification was valid for a NEW fastener in "as received" condition. This means that it was coated with a substance with lubricity equal to that of 30W oil. Less lubricity than this and the resultant clampload will be too low resulting in loosening. Higher lubricity will lead to more clampload which leads to thread failure and fastener stretch. The Loctite threadlockers mentioned have a lubricity controlling agent which provides lubricity in the correct range. Beware - this is a patented property of Loctite threadlockers and cannot be found in any other brand. It will also lock the bolt and prevent it from loosening.

I told you this fix was engineering quality and here's the number. After using the 242 or 243 threadlocker as described, to dismantle the assembly you would need to apply a force of 2000 pounds for every inch of mating suface area.

Sorry about the long post.

[This message has been edited by Softop (edited 12 August 2000).]