The MGA With An Attitude

ST-102 defines different methods of measuring and calibrating the error of the instrument.

I have a TSD rally computer in my car, so this part is easy for me, although the explanation may at first seem useless to you, bear with me, you'll get the idea. I set the ODO calibration factor to something close to what I think is correct. Then I drive on an interstate highway with mileage markers every mile. After 10 miles I figure the error ratio between the actual traveled distance and the recorded miles and adjust the correction factor accordingly, so the computerized ODO is dead on.

Next I enter the intended travel speed into the computer (say 60 mph), set the computed time to match the time of day clock, and proceed to drive at the set speed. The computer does a calculation like 60 mph = 1 minute per mile, and continuously adds time to the calculated time clock as I drive. An error meter on the dash displays the difference between the time of day and the calculated time, so if you drive at exactly the intended speed the error display stays on zero. Positive error, slow down. Negative error, speed up. When I can hold the error on zero for several seconds at a time, the speed is correct, and I note the speed reading on the speedometer.

For other speeds higher than 75 or lower than 55 I just find an empty stretch of country road, set the speed in the computer to whatever, set the computed time to equal the time of day, drive the right speed to keep the error display on zero, and note the speed indicated on the speedometer.

At 60 mph my speedo reads 63 (5% high), and it's pretty much proportional to the scale, like 21 at 20, 42 at 40, 84 at 80. This is obviously a weakness of the coil spring, or a misalignment of spinning magnet, or too much strength in the magnetic field. But it's not too far off, and once I know what the error is, it's no big deal. If I wanted an easy fix, and I didn't mind a small error, I wouldn't bother with the spring rate. I would rotate the spring or the needle backwards by about 2 mph so the output would read 19 at 20, 40 at 40, 61 at 60. 82 at 80. Plus or minus 1 mph within the legal speed limits is not a bad result.

You can still do this calibration without a rally computer, it just takes a little more time and more driving. Take a watch with seconds indication and go drive on an interstate with mileage markers. Adjust you speed until you can consistently record flying miles of exactly 60 seconds each. This gives you your speed reference (and error) at 60 mph. While you're at it, watch your odometer and record the indicated distance readout in 10 miles of travel. Read between the tenths if you can to record the nearest hundredth of a mile over 10 miles. While you cannot adjust your odometer to be dead on like I do, you can still use your odometer to calculate and measure correct distances and times once you know what the error is.

Suppose you have smaller tires than original or you have changed your final drive from 4.3 to 4.1. Your odometer, if originally correct, should now read out 10.49 miles in 10 miles of travel on the interstate (4.3 / 4.1 x 10). For a more accurate reading you can drive 20 miles for the first check.

Now you want to check the speedo error at 40 mph? 60 / 40 = 1.5 minutes per actual mile = 1.5 minutes for 1.049 recorded miles. Also at the correct speed you should get 1.43 minutes (85.8 seconds) for a recorded mile (1.5 / 1.049). This is easier to deal with. Make up a little time sheet in 1/10 mile increments for 3 miles of travel (30 lines). 0.1 mile = 8.58 seconds (rounded off to nearest 1/10 second = 8.6 seconds). 0.2 mile = 17.2 sec, 0.3 = 25.7, ...... all the way up to 3.0 miles = 257.4 seconds. Now it's nice to have a navigator to read the chart and keep an eye on the watch. For a normal watch you probably want your calculated times in minutes and seconds, like 257.4 seconds is 4 minutes 17.4 seconds.

Find an empty stretch of country road, flatter the better. You drive at what you think is 40 mph, keep it as steady as possible. Keep one eye on the speedo, one eye on the odo, and at least one eye on the road (nice trick). At zero tenths on the odo you call out "Mark", and the navigator starts the clock. As each tenth of a mile rolls up on the odo you call out "Mark", and the navigator checks the clock against the chart, telling you only if you are a little ahead or a little behind schedule. You adjust the speed accordingly until you are calling "Mark at just the right time for several tenths of a mile. Then you note and record the reading on the speedometer. All this amounts to is a fairly accurate human feedback cruise control.

Make a data list for other speeds, test drive with the assistance of a navigator, and record the speedo readings, so long as you patience holds out. The results of all this effort will give you the speedo error at each of the calibrated speeds. For the purpose of adjusting your speedometer, 10 or 15 mph increments up to 80 mph will generally suffice, so you only need maybe a half dozen data points for this. My MGA speedometer is numbered every 10 mph, so the even 10's make good reference points, the odd 5's are a little harder to read.

Suppose you take just one trip on the interstate to find your actual odo reading (and error). Then you can make up time charts for several different speeds in advance, and you should be able to complete the rest of the calibration in the second outing. Once you have figured out what adjustments you need to make to the speedometer, the actual mechanical tinkering is just a little tricker than brain surgery. Then you get to go out and drive the second part of the calibration run all over again to see if you did it right.

For the brave of soul and the devout tinkerers, ST-201 gets into ways to effectively change the calibration of the instrument to improve the accuracy of the speed readings.

Barney Gaylord

Thank you for your comments -- Send e-mail to <Barney Gaylord>
© 1998 Barney Gaylord -- Copyright and reprint information