ST-202 gets deeper into the workings of the magnetic speedometer drive and ways to affect callibration changes there.

So, you're still here? In this session we get into something akin to playing god with magnetism. Please set your brain to the anti-meltdown mode.

Referring to the spinning magnet mentioned in ST-101, the magnetic torque is directly proportional to the speed of the spinning magnet. The reverse torque from the clock spring is directly proportional (almost) to the angular motion of the needle winding it up. Doubling the speed of the input doubles the motion of the needle, so the face scale is linear (again almost). The primary cause of non-linearity is the fact that the spring is not quite at zero torque when the needle is at zero mph.

You cannot change the magnetism of the spinning magnet. You can change the alignment of the spinning magnet relative to the cake pan, but not a good idea. Most of the magnetic "drag" that is acting on the cake pan happens at the flange of the cake pan. Changing the face spacing distance between the disk and the cake pan therefore has only a small effect. But this is one source of error. As the input bushing wears (on the ends of the bushing) the disk can float axially, causing a little flutter in the needle, maybe +/- 2 mph, just enough to make you think it should work better.

The larger source of error caused by wear comes from the input bushing wearing out to a larger bore size, allowing the spinning magnet to waffle back and forth from its' optimum center position. The magnetic drag is inversely proportional to the distance of the magnet from the flange of tie pie tin, but this effect is exponential (non-linear). Cut the gap on one side by 20% and you get about 44% more drag. At the same time you open the gap on the opposite side by 20% and you get about 64% of the original drag. 64% + 44% = 108%. The net affect is that the speedo will then read higher than it should, and the error will be proportional to the speed, like +3 mph at 30 mph, + 6 mph at 60 mph, etc. When it gets obviously worse than about 10% error (usually accompanied by a little flutter), it's time to replace the input bushing. Sometimes wear on the bushing around the needle shaft can have the same effect and may also need replacing.

Referring to the prior paragraph, correctly realigning the spinning magnet to run concentric with the pie tin (if it was misaligned) should reduce the output reading proportionately all the way up the scale.

Trimming the ends of the magnet strap to increase the air distance between the cake pan flange and the strap would make the greatest changes to the speedometer reading. Taking material off the ends of the magnet would cause the speedometer to read slower. As the magnetic drag is inversely exponential to the air gap between the spinning magnet and the cake pan, very small changes in the air gap will make very large changes in the speed reading. Changing the gap from 0.020" to 0.023" will decrease the attraction by about 24%. And here I'm mentioning a change in gap about equal to the thickness of a piece of paper.

Suppose you want to reduce the speed reading by 10% by increasing the air gap. Take the square root of that ratio and multiply by the original gap to get the new desired gap. Suppose that the original gap was 0.020". Then (1.0/0.9)^0.5 x 0.020 = 0.0211, meaning that you only need to increase the air gap by a bit more than 0.001". So don't take off too much, because it isn't easy to put it back.

And you can't add magnetism to the spinning magnet, because it is a permanent magnet. A piece of iron can be magnetized to varying strength, but the effect is not permanent. The iron magnet will eventually loose its magnetism. A permanent magnet on the other hand has a fixed saturation density for the magnetic field. This means that it will always have a magnetic field of a fixed strength, depending only on the material it is made of. If there are impurities in the material, these impurities might be magnetized to varying amounts, thereby affecting the strength of the field slightly. But, those impurities would not retain permanent magnetism, and would eventualy either relaps to non-magnetic, or would assume the field strength of the permanent material arond it. Either way, the total magnetic field of the alloy would be stable with time.

If you actually manage to tinker in some improvement after all this, I would die a happy man. If you now decide that the delicate internal adjustments are a bit much for your ham fists and thick fingers and send the unit out to be repaired, I think this may also have been a good use of my time.

First time I sent mine out for service was because the needle was off, the needle shaft was seized, and the input bushing was shot. I think the results were satisfactory, but then I wasn't so picky in those days.

Second time was because the ratchet gear had lost some teeth and the needle was wavering. By this time I had my rally computer and could easily tell the exact error. I told the shop that the odo reading was right on (100.4 miles indicated in 100 miles), but the speed was reading about 10% high and wavering. When it came back as repaired, the speed was still reading 5% high, and I was a bit disappointed.

I suspect that the repair shops do little or nothing to calibrate the speed output reading after completing the repairs. I think they just install the specified OEM spring, set the needle to balance at zero, and depend on the spring to be accurate from there. If they had made any attempt to calibrate and adjust the speed reading, they would have moved the needle back about 2 mph. I'm pretty sure they spun it up to watch the needle move, but I seriously doubt that they used any known speed reference for comparison.

So, it is my concern that one may not be completely happy with the end result of sending parts and $$$$$ off to fantasy land with a prayer. It is my impression that the repair shop is satisfied if the odo ratio is correct and the speedo needle moves when finished, accuracy be damned. Then again, sometimes I just have a bad day.