The MGA With An Attitude
Variations and Adjustments

ST-201 gives a more detailed description of the internal workings of the speedometer, how certain variations and adjustments can effect the calibration, and how to make some coarse adjustments to the dial readout.

The speedometer and odometer are not completely independent. They can be adjusted independently inside the unit, but as they are both driven by the same cable from the gearbox, any change outside of the unit will affect both proportionally.

As for odometer error, the only way to adjust for that is to change the gear reduction ratio somewhere along the line. Things you can change are:

-- Tire size,
-- Final drive ratio,
-- Speedo drive gear in the gearbox (sometimes you can get a plastic gear for the bottom end of the cable that has one tooth more or less than the original one),
-- Gear reduction box inside of the speedometer,
-- Number of teeth on the ratchet gear (these gears may be available through the rebuild houses).

Another way is to purchase a speed change box to install between the top end of the cable and the back of the speedometer. This item may be quite pricey when you find it. The main advantage of these little reduction boxes is that you can get practically any ratio you need, and then you can change the ratio later at minimal cost if you change something else in the car that screws up the odo again.

If you are going to do anything outside of the speedometer case to correct for the odo error, be sure to do this before attempting to correct the speedo reading. Any change in the drive rate of the cable will also affect the speed reading.

Remember that trying to calibrate a worn out speedometer is a lost cause, like trying to set ignition timing with a worn out distributor. You should repair the worn parts first. But you could attempt some of the following strategy and find some improvement, even on a somewhat worn unit. When a speedo is in otherwise good condition, there are three things you can do to affect the calibration.

1.) Making the clock spring stiffer (higher spring rate) will reduce the speed reading proportionally all the way up the scale. This is not easy to do without replacing the spring, generally requiring thicker wire. And the adjustment you can make to the existing spring is rather coarse. You could try fastening two adjacent coils together by spot welding or soldering (low temperature please), effectively reducing the length of the coil by one turn. Count the coil turns, subtract one, divide one number by the other, and you get the amount of increase in stiffness. Example: 8 active turns, cancel the motion of one = 7 active turns. New spring rate = 8/7 of the original rate, so the output will now read 70 instead of 80, 35 instead of 40, etc.

2.) Rotating the needle on the shaft has the effect of increasing or reducing the output reading by a fixed amount all the way up the scale. If you set it back by 3 mph, it will read 17 where it used to read 20, 37 instead of 40, 57 instead of 60, etc. The problem I've had with this fix is getting the press fit needle hub off the shaft without breaking the needle or needle hub (fit is pretty tight), and without bending the shaft (quite thin). The next problem is to somehow mark the position of the shaft (pie tin) so you know where it was before you removed the needle.

3.) Winding the spring up tighter has the effect of reducing the speed reading by a fixed amount all the way up the scale, just like rotating the needle on the shaft. You may find that the fixed anchor point for the coil spring can somehow be rotated around the center of the shaft. Otherwise, if you have jeweler's tweezers and a steady hand, you could fold the coil back on itself in a double bend "Z" shape to shorten it very near the fixed anchor point. This has the same effect as rotating the coil or rotating the needle on the shaft, and may be easier. Take a good look at which way the coil is wound. Usually the coil gets tighter as the needle increases reading.

Having figured all this out, you may also notice that items 2 and 3 will have another effect at low speeds. Like the needle won't come off the peg until you hit 5 mph. You may have long since noticed but ignored the fact than the speedometer is configured and positioned in such a way that the needle rests on a peg by gravity at the zero reading when the car is at rest. I believe the original design is to adjust the coil spring so it is at rest (zero tension) when the needle is on zero, or to give it a slight negative torque so that it almost supports the weight of the needle. If you hold the instrument in your hand and tap it gently from the bottom, the needle should bounce up off of the peg and then settle back, bouncing a couple of times, and just coming to rest on the peg with very little force. This should be a good starting place.

If you don't think it would be viable to mess with the clock spring, maybe because you can't get to it, think again. If your clock spring breaks, there are many repair shops who will install a new one for you. What man has built, man can take apart, and rebuild again. But it can be very dificult to get to without disturbing a lot of other parts, and at considerable risk of some damage.

If your needle reads proportionally high by at least 10% all the way up the scale, you may consider trying item 1. If it's a few mph (constant amount) high or low all the way up the scale, you could try item 2 or 3. By using a combination of these techniques you can affect the linearity of the output, much like using two different size springs in the mechanical advance of a distributor. It all takes a bit of figuring up front. FIRST you have to record the error at different points on the scale before you decide what you want to change. For that proceedure refer back to ST-102.

For the purpose of adjusting the needle by a couple mph, it may be easier to hook the speedo up to a fixed speed electric motor to check the speed readout before and after each adjustment.

For some artistic type people there's another way of getting the speedo needle to read correctly after the calibration run. Repaint the face of the speedometer so the markings match the position of the needle at any given speed. This yields a non-linear scale that matches the non-linear motion of the needle, and then the speedo is dead on correct. Unfortunately it is then also non-stock and non-standard, so it would be more difficult to recalibrate if it were to need more repairs in the future.

For the brave of soul and the devout tinkerers, ST-202 gets into more technical aspects of the spinning magnetic drive for the speedometer needle and some devious ways to affect the calibration of those parts.

Thank you for your comments -- Send e-mail to <Barney Gaylord>
© 2002 Barney Gaylord -- Copyright and reprint information