General Tech Speedometer Calibration

[Geo Hahn]

I just had a speedometer rebuilt and calibrated by one of the Big 3 firms that do this work. The work was needed to accommodate a change in the diff ratio. The results are mixed.

I did their prescribed test - roll the car 1/100 of a mile and count the cable turns. This part went well as the odometer is now accurate and the speedometer is accurate at 60 mph. This suggests to me that the unit was spun at the speed indicated by the roll test (cable turns x 100) and the unit either indicated (or was adjusted to indicate) 60 mph.

The problem is that the speedometer reads low (compared to a GPS) at low speeds - for example an indicated 35 is an actual 40, and indicated 40 is an actual 44, etc. It becomes correct at 61 mph and starts to read too high above that. I was limited in being able to test at very high speed as I was too far from a suitable stretch of road.

I thought this non-linear error meant an issue with either the strength of magnetism in the flying magnet or the strength of the spring. The reply from the rebuilder was this:

Because of the wide variation between the original and current calibration of the Speedo, any further fine tuning of the calibration is limited to weighting the pointer and playing with spring rates both of which will produce results that are in all honesty, a compromise, I.e., weighing the pointer will drop the Speed down at the lower range but will also slow it down in the upper range, changing out the drum spring to one with a higher rate will help solve this problem, however there are not a broad range of different spring rates available.

Are things as hopeless as that seems to suggest?

Is the strength or weakness of the flying magnet not a factor in this calibration?

I'm not criticizing the rebuilder at this point - just trying to understand if the limits of what they say can be done are, in fact, reasonable.

Disclosure: This is unit from a Jaguar, not a TR -- but the design is the same and I am asking here because this is where I usually find answers.

 

[sail]

I'm of no help but would like to bring a chair and watch the next roll test.

 

[CJD]

Hey Geo!

To fix that issue, you will have to turn the needle CCW on the post, which will slightly increase the resting spring tension at "0" mph. This will have the affect of lowering the indication at the lower speeds. The problem is, this will also throw off the calibration at "60" mph, making it show slower than the car actually is going. So, then you will have to increase the magnetism on the spinning magnet to bring the "60" calibration back up to spot on. If you search my thread from last year, you will see that this merely involves brushing a medium strength magnet across the rotating magnet to magnetize it to a stronger state.

As you can tell, all this involves a bit of trial and error. But, with patience, you can get the speedo within 2mph throughout the "0" to about "80" mph. I guess I'm trying to say...they could have got it closer when they calibrated it. Their response is just a fancy way of saying they didn't want to spend the time. Springs do not lose their rate, so the spring you have can calibrate the speedo to the same accuracy as it did when new...so long as it is not damaged.

 

[TR3driver]

The strength of the magnet is very much a factor, and it can be adjusted. That is exactly how I recalibrated the speedometer portion of my TR3 speedo, and the results came out very close across the range of interest. But it was a bit of a learning curve, after several attempts at weakening the magnet, I got carried away and weakened it too much. So I had to remagnetize it and start over.

Here's a shot of my ad-hoc degaussing tool, along with a printout of a simple spreadsheet converting rpm (at the head) to desired mph indicated, using the measured no-load shaft speeds from my lathe and calibration number from the speedo face. I used 24 mph and 108 to check the magnet strength while I was changing it; then double-checked afterwards at 40 and 65. I got all 4 points to be within 2%, which I felt was pretty good. (Of course, the calibration doesn't match my tires, but I plan to work on that in the future by changing the rear axle ratio. Then I'll revisit the speedo calibration.)



And here is the remagnetizing tool. The AGM batteries weren't enough, so my final configuration was the pair of 300,000 mfd caps in the foreground, charged to roughly 50 volts by the bench power supply (not shown) through the headlight bulb (for current limiting).


I used the lathe for a speed source (the cable is an old tach cable from the junk bin), but you could rig up almost any small motor as long as you have some way to control the speed (and know what it is). Or even use the tach connection to your TR motor, and an electronic tach to check actual rpm.

 

[ed_h]

A linear instrument, which the speedo approximates over a certain range, needs a two-point calibration to read correctly over its range. Your rebuilder apparently only did a single point calibration at 60 mph.

A two point calibration sets not only the height (or "zero crossing") point of the response curve, but also the slope. Your slope is too steep. It crosses the correct curve at 60 mph, where both curves would be correct.

The slope of the response curve is determined by how strongly the magnet tries to pull the disk, and how strongly the spring resists it. In your case, you need a weaker magnet or a stronger spring. The rebuilder's suggestion that weighting the pointer would have a positive effect is just silly. Weighting the pointer will damp it so that is moves more sluggishly, but won't change its final position.

I believe that changing the magnet strength is easier than trying to mess with the spring. Randall's approach is a good and effective one. I did something similar, but with permanent magnets. You can read about it here, starting about halfway down the page: https://bullfire.net/TR6/TR6-40/TR6-40.html

Ed

 

[CJD]

Good point Ed...I totally missed the "needle weighting" they recommended. The needle should only be weighted so as to balance the needle at the center post. In other words, the weight counterbalances the needle, so the needle doesn't try to hang downward from gravity. Assuming the needle hasn't been gobbed with thick, heavy paint, there should be no reason to touch the factory needle balance. A symptom of an out of balance needle is that it bounces the indication up and down when you go over bumps while driving.

I am sure Randall's magnetizing rig works great, but really, all you need is a decent, small, rare earth permanent magnet. Just draw it across the speedo spinning magnet and it re magnetizes it just fine...and much easier!

This tiny magnet is all it takes! Brush it across the speedo magnet and that is enough to "over" magnetize the spinner. Then, to reduce the strength, slowly move the little magnet near the spinner, while it is spinning, and it will overpower the spinning magnet and demagnetize it. Unbeleivably easy!! The case has to come off the speedo, but no further disassembly is required.

https://www.britishcarforum.com/bcf/showthread.php?94885-Instrument-Faces/page2