To my other TR6 pages

April 13, 2014

Flywheel and Clutch

With the engine and transmission finished, the only remaining pieces between them are the flywheel and clutch.

The flywheel is a massive piece of iron that rotates on the rear end of the crankshaft.  It serves to damp the torque pulses from the  individual cylinders, smoothing out the delivery of energy to the drive train. However, one of the downsides of a heavy flywheel is that it takes significant energy to increase its rotational velocity.  This can take away from the linear acceleration the car is capable of.

One popular performance modification for these and many other engines is to lighten the flywheel by an amount that will allow better acceleration, but not impair the smoothing action.  Unfortunately, while simple in principle, it's a little more complicated than just hacking pieces of the flywheel off.  First, the flywheel must maintain sufficient strength to withstand the very large forces generated at high RPMs.  Second, the flywheel has to remain balanced so that it doesn't contribute vibrations of its own.  

Beyond these things, it turns out that it is very important where the weight is removed.  A flywheel resists acceleration by its moment of inertia (MI).  The total MI of a flywheel depends not only on the mass of the wheel, but how the mass is distributed.  Mass near the periphery of the wheel contributes much more to the total MI than mass nearer the spin axis.  So to be most effective, flywheel material has to be removed as far from the spin axis as possible, consistent with strength and balance.

One interesting way to look at the effect of lightening a flywheel is to calculate the equivalent weight that would have to be removed from the car to give the same increase in acceleration.  The basis for this comparison goes something like this:  When accelerating, some of the energy developed by the engine must be used to accelerate the flywheel.  This energy is not available to drive the wheels.  If we lighten the flywheel so that it takes less energy to accelerate it, the more the drive train will see.  There's more, and here is where it gets interesting.  Since there is a gear reduction between the engine and the drive wheels, there is a torque multiplication.  This means that any increase in torque into the drive train will be multiplied by the gear ratio before it gets to the wheels.  This increased torque at the wheels translates into a larger linear force on the pavement, which means a higher acceleration for the car.  One other way to get higher acceleration for a car for a given linear force on the pavement is to lighten the car.  Given these two ways to get a car to accelerate faster, we can compute the weight reduction of the car that would give the same acceleration increase as a certain flywheel weight reduction.  Depending on where the weight is removed from the flywheel, the exact drive train gear ratios, and the size of the drive wheel, the multiplier can easily be more than 10--that is, a pound of well placed metal removed from the flywheel will give the same acceleration increase as lightening the car by 10 pounds or more.  Those afflicted with a knack for math can see more detail on this here.

My flywheel as removed was dirty and rusty.




First task was to clean it up to see if there were any obvious problems.  After this, I thought the flywheel was in pretty decent shape.  The friction surface still showed some evidence of the last grind marks.




Then I weighed the wheel so I'd know the starting point.  It came in at 27.24 pounds.  Many consider this to be a little heavy for these cars anyway.




After some careful measuring, I came up with a plan for material removal.  I calculated that this should remove around five pounds, but being mostly towards the periphery, the effect should be maximized.  


**DISCLAIMER**Please note that this plan for material removal is not a recommendation.  It is a description of what I did.  Please don't adopt what I have done here without a full understanding of the principles involved, or professional advice.


This would normally be done on a lathe, but the flywheel is just a little too big for mine, so the material was removed on a mill, as evidenced by the tooling marks.




The final weigh-in showed a new weight of  22.44 pounds--a loss of 4.8 pounds, pretty close to my estimate.  I know my vacuum got a lot heavier.

 


Then it was off to a shop to have the flywheel surfaced and balanced.  Unexpectedly, I found that most shops, at least in my area, did't have the small arbors necessary to balance TR6 flywheels.  Finally found one, but it was 30 miles away.




One other mod  for a flywheel that is sometimes considered is stepping.  Stepping a flywheel consists of machining away a little material on the part of the flywheel where the clutch mounts.  This in effect moves the pressure plate a little closer to the flywheel.  The motivation for doing this is often stated as increased pressure on the clutch disc.  This sounds intuitive, but in reality, the result may be just the opposite.  The problem is that the belleville type springs used in diaphragm clutches have a very non-linear rate curve.  In fact, it is so non-linear that it often has a region where the rate is actually negative.  In this region, increased displacement actually results in less force.  You can sometimes actually feel this in pushing the clutch pedal--the required effort peaks early in the travel, then gets smaller.  It is in this negative rate region that a new clutch operates.  As the clutch disc wears thinner, there is less displacement of the pressure plate, and the operating point moves up the curve to higher pressure until it finally passes "over the hump" to a region of positive rate where further wear reduces the pressure, eventually to the point of slipping.

It could be that the idea of stepping a flywheel is something left over from the days of coil spring clutches with their simple linear rate springs.  Also, it is true that some stock flywheels are stepped, but in these cases, the step was presumably a design parameter used to get the desired operating point for the clutch.  On the other hand, stepping a flywheel as a modification has to be undertaken with a good understanding of the spring rate curve for the particular clutch.  I didn't have any good reason for wanting to change the disc pressure, let alone reducing it, so I didn't consider stepping any further.


Clutch Linkage

The TR6 uses a standard diaphragm clutch.  The release (throwout) bearing rides on a carrier that is moved by a fork.  The fork is fixed to a cross shaft that runs horizontally through the bell housing and carries an arm on its drivers side end.  The arm is actuated by a hydraulic slave cylinder.  There are a number of joints in the overall mechanism, and each can contribute a little slop.  Most of the parts in this chain of linkages needed some kind of attention.

Back in the early 80s when I had the transmission out to fix the third gear synch ring, I also modified the bushes for the clutch cross shaft where it runs through he bell housing.  I took out the short steel bushes and installed full length bronze bushes, and then drilled and tapped the bottom side of the bush bosses for grease zerts.  This seemed to work really well for maybe a year until the clutch started getting stiff and sticky.  It would release OK, but was slow to engage.  It appeared that the cross shaft was binding in the bushes.  It finally got so bad that one day I pushed in the clutch and it stuck there.   I didn't have time to fix it, so in the garage it went.  My theory at the time was that clutch dust mixed with the grease and hardened.  Fast forward 30+years, and it was still stuck when I parted the transmission from the engine a few months ago.  When I started working on the transmission a few weeks ago, the shaft would not budge.  I had to put a 3 foot cheater on the cross shaft arm to get it to move.  one bush was seized to the shaft and I had to cut it off.  This is still a bit of a mystery to me since I know that earlier TRs had zerts for their bushes.




On the way to removing the cross shaft, I encountered a heartache that many TR owners know all too well.  When I unscrewed the tapered pin holding the fork to the shaft, only part of it came out.  The pin had broken in its bore and the lower part was keeping the fork from coming off. This well known problem is caused by a bad design of the bore for the pin.  The bore in the shaft is tapered, but the bore in the fork is not.  This leaves the bottom end of the pin not registered to anything on the fork so all the shear force is concentrated at the upper intersection of the shaft and fork.  The pin can't take it, so it breaks.

The most immediate problem was to get the remains of the pin out so the fork could be separated form the shaft.  Since the shaft won't come out of the bell housing with the fork on it, all operations to remove the pin had to be done in the housing.

The solution, not original to me, was to drill a small hole in the fork such that it would open into the bottom of the tapered pin bore.  This would allow a small drift to be inserted to drive the pin fragment out.  




While on the subject of the fork, it has two pins which engage a groove in the release bearing carrier.  These pins were pretty worn, especially where one of them rests against the anti-rotation pin in the carrier.  These pins are replaceable, and are just driven out, and new ones pressed in.




The wear on the fork pins was matched by wear in the release bearing carrier groove.  There is a pin in the groove that prevents the carrier and the inner race of the bearing from turning.  Torque transmitted by the bearing keeps the carrier up against the pin most of the time, so the carrier wears in the areas where the fork pins rest.




The wear bothers me, and if there were no alternative, I'd replace the carrier.  However, my mistrust of most aftermarket parts led me to another option.  Since the areas of wear are determined by the position of the anti-rotation pin, it seemed simplest to just move the pin.  I drilled another pin hole 90 degrees from the original.  This renewal could actually be repeated quite a few times.




Another area that was asking for attention was the arm on the cross shaft.  There are three holes in the arm to connect to a pushrod from the slave cylinder, presumably to allow a little adjustment of the lever ratio between the slave cylinder and the release bearing fork.   It's the middle hole that's nearly always used, and mine was elongated.  To fix this, I welded up all the holes, and just redrilled the middle one.  I've never even been tempted to use either of the other holes, so they seemed superfluous.




Next up was the slave cylinder pushrod.  Like the cross shaft arm. its clevis holes were distorted.  




Rather than trying to weld up the rather thin arms on the pushrod, I decided to just make a new, slightly beefier clevis end for the rod.  I'm particularly proud of that silver solder job.  I don't think I've ever had one come out that nice before.  The last pic is after zinc plating the rod and making a new stainless clevis pin.




My slave cylinder was toast.  I think I took it apart back in the 80s when I made a feeble attempt to diagnose the clutch problem, and it was sitting under one of the seats for 30 years.  It's bore was rusted and some of the parts were missing.




I got a new slave cylinder.  It has a smaller body diameter than the original, but the one inch bore is the same.  It had a pretty lousy paint job, so I repainted it and the original mounting plate.




The last job was to fix the design flaw that causes the taper pin on the fork to break.  There are a number of decent solution that have been applied here.  Some involve giving some support to the small end of the stock pin, while others provide a second pin or bolt to take some of the shear stress off the original pin.  I found a nice writeup on one of the TR6 forums that describes what I consider to be an ideal solution.  It consists of installing a new taper pin in a fully tapered hole.  There is even a boss on the fork that looks like it was created for just this kind of solution.  The thing that makes this approach so sweet is the unique taper pin used.   The pin has a threaded section on its small end which allows a nut to draw the pin in and keep it from backing out.  Known as an "AN386" type pin, it is most often used in aircraft applications (I believe the "AN" identifies it as a military specification).




The pin has a 1/2" per foot taper, and is compatible with a Brown & Sharp (B&S) #2 taper.  After screwing in the top remnant of the original pin to locate the fork, I drilled through the boss on the part so that the hole would be centered in the shaft.  I then followed with a slightly larger drill from the other side to rough out the taper.  The final taper was accomplished with a B&S #2 hand reamer.  There isn't a much more accurate way to make a taper pinned joint than this "in situ" process.




The pin can be pushed in by hand, but the nut draws it in another 1/8" or so.  To guarantee that the pin is in full double shear, the bore is usually sized so that some of the small end of the pin sticks out a little when the pin is fully seated.  That's why the special recessed washer is used under the nut.




I had ordered a new stock pin, and though I think the new pin makes it redundant and unnecessary, I didn't see any reason not to install it anyway.  The tapered spring goes on the shaft between the lever arm and the  bell housing.  It is apparently no longer available, and mine was torn up during the violence in removing my shaft.  The one in the picture was donated by a generous member of one of the popular TR6 forums.





Comments to:  elhollin1@yahoo.com

To my other TR6 pages