Cleaning the magnetic Smiths speedometer (Part 1)

by Laurent Castel (SW France)

One of the pleasures of working on a vintage car is the various activities that you can do, or at least learn a few basics. One day sewing, the morrow welding a fender, the other day electricity, plumbing or paintwork. Today, I propose a bit of watchmaking and micro-mechanic.

Don’t be afraid of getting into it if you are meticulous. You’ve guessed – I’m talking about the speedometer. The late model, with dished glass which relies on magnetic principle.

First, there is a very good paper about these speedometers written by Anthony Rhodes that you can find on several websites. ‘Repairing Jaeger and Smiths Speedometers’

Unfortunately, there is something wrong with at least my speedometer (see that in part two about odometer). You should read Anthony’s paper anyway as it brings much interesting information. I won’t duplicate its information but rather complete this knowledge. The present article will focus on the speedometer whilst the odometer will be addressed in our favourite magazine.

Smith speedometers and odometers are a great piece of engineering.

Just like many other devices in our old MG, fixing very often means disassembly, cleaning and reassembly.

Disassembly begins with removing the bezel and  glass. Just bend a little bit the 8 tags of the back face of the bezel off the housing. You will leave them in that position for reassembly. Clamping strength is enough and you will be able to disassemble and assemble many times now without bending them more.

You can now twist the bezel and the housing to take the bezel apart. Then the seals, the glass and the dial ring come off.

Gently lift the tip of needle just above the stop to observe the actual rest position. Mine is just below the stop. The needle touches the bottom side of the stop.

Now, it’s time to make this useful and very simple tool.

Use the tool as a lever to extract the needle. Two tiny screws secure the dial.

Two other screws on the back face of the casing allow removal of the mechanism after the reset command has been disconnected.

Tricky moment, you have to disconnect the two tiny springs that are attached to the two odometer pawls. They are very prone to fantastic leaps in the workshop.

The speedometer part can now be removed from the whole instrument with 4 tiny screws. Take notes, they are not identical. You need to maintain the two parts in their position with your hand. Now turn everything upside down, so that the number wheel faces the workbench and the input shaft faces the ceiling. Very carefully, you can now separate the odometer mechanism from the input casing. The aluminium cup must remain with the odometer part as it is attached with a hairspring to the odometer mechanism. We won’t go further for  this aluminium cup. I’m meticulous but I know my limits! Put the odometer part in a safe area, always keeping the odometer wheels toward the workbench and the aluminium cup on top.

Now the easy part looks like this:

The centre bearing on this side of the spinning magnet is the bearing for the rotating aluminium cup. The rotating cup directly drives the needle. Here is a mechanical assembly that is quite unusual. Bearing is spinning whilst the shaft is steady (or almost). This bearing needs to be very clean. I used no oil in it.

Remove the two screws that secure the plate. Then you can slide the locking plate which maintains the spinning shaft of the magnet plate.

Remove the magnet plate.

Remove the two fork shaped spring blades that secure the odometer command gears and remove the gears. If you get as far as there, then I remember that you’re meticulous. So, take note of the respective position of each pawl as they are different. Look carefully, one is a pusher, the other is a puller.

Now you can clean everything. Pour a drop of thin oil in the felt reservoir of the main bearing on the casing, under the washer. Smear a very small quantity (half a grape seed) of light grease on both gear shaft recesses. They also act as a reservoir. Then you can reassemble the whole puzzle. Take care of the pawls…

The pegs that move the pawls are out of centre. When inserting the magnet shaft and worm gear, set the two pegs so that they are in opposite position. Set them both away from the casing for example.  We will see in the odometer section that when one pawl acts on  its ratchet wheel, there is an additional resistant torque for the cable. Setting the two pawls in opposite positions ensures that they don’t turn their ratchet wheel at the same moment.

Coat the locking plate with a thin film of light grease. Actually, use a rag with grease and wipe off  any visible grease. Check that the mating surfaces of the casing and the plate are clean The plate might be worn on the side facing the worm gear. Reassemble the same way. Check that everything is free to move and that the end play of the magnet remains below 1/10mm. If end play is greater, then try to put the plate upside down, the worn side toward the casing.

Now, you can choose to wait for Part 2 of the article about odometer or to assemble everything back together. Assembling the speedometer and odometer together absolutely needs no effort. Just careful alignment.

Hold the odometer assembly in one hand with the cup still facing upward. With the other hand fit the speedometer assembly over the odometer assembly. Align the pegs and the screw holes.  There is only one position. No error possible. Fiddling the aluminium cup will allow its spindle to enter the bearing on the magnetic plate. The bottom end of the spindle should have remained in its fixed bearing on the odometer side. Check it. Screw in the four tiny screws that maintain the odometer and the speedometer assembly.

Part two will take you into the odometer wheels but two more topics are often discussed about the speedometer and I can add a couple of tricks that are rarely described.

About calibration.

The physics of the speedometer is based on Eddy currents (Foucault in France). When the magnet turns in front of the conductive cup, it induces a rotating current in the cup itself – like if it was a dynamo coil. This current produces a torque that tends to decrease the rotating magnetic field. So, the cup turns the same direction as the magnetic field but the hair spring stops it when the hairspring counter torque is equal to the Eddy current torque, indicating the rotating speed of the magnet. The Eddy current torque is proportional to the magnet speed. Counter torque of the hairspring is proportional to the angle of the cup.

Hence the formula :

Indicated speed = A x (shaft speed) + B
A is related to the magnetic field, the magnetic coupling with the cup and the hair spring strength factor.
B is related to the rest position of the needle and the initial length of the spring.

Depending on how you want to move your curve to match the ideal curve shown on the previous diagram (1600 TPM), there are two ways of actions.

B: the offset parameter is a function of spring initial length and rest position of the needle : set the needle at a different rest position. Easy.

A: the gain parameter is related to magnetic field, cup/magnet coupling and spring strength factor.

Spring strength factor cannot be altered. It is a function of the spring section and the material.

Coupling is not easy: it would need to change the gap between cup and magnet or the shape of the cup.

The best solution is to modify the magnetic field. This will be detailed in another complete article. 

About wavering needle.

Check your speedometer with a drill running slowly and reverse. if the needle is steady then the cable is the culprit. It is the most probable cause. The cable wears…. As new, it can only twist less than 1/8 turn between fingers in the right direction, less than 1/4 turn in the wrong direction. The external layer of the winding is built so as to tighten itself around the inner core. It is thus more rigid in the direction it is intended to turn.  With ageing, it can twist an entire turn or more in both directions. A small resistant torque at the output of the cable makes it slowing down, stores energy, increases the output torque until it becomes higher than the resistant one. Suddenly the energy is released and the cable accelerates causing wavering of the needle.

The longer the cable the greater the wavering. The furthest the resistant torque from the gearbox, the greater the wavering. So, after checking all the common recommendations, bent, chips, grease and so on, Check the cable for its twist capability and purchase a brand new one. It will be rigid as a shaft.

You’re meticulous but now, you also need to be patient and wait for the next issue of our favourite magazine that will address the odometer part, still looking at the ceiling of your workshop!

Ed’s note:

Laurent owns TD29133 which he came to England to buy. His TD started out in life as NOV 2, a Birmingham registration number, issued in 1953. This registration mark with a low number would obviously have been attractive to somebody with these initials so it was sold and a new number 3966 AD allocated. When Laurent bought the car, it still had the 3966 AD plate, but he was told by the seller that he would retain the number before selling the car to him. The car then sported the registration number 552 UYF.

It is known that the car was in the Plymouth and Penzance areas in the 1960s and also in Cardiff. By 1970 it was in Holywell (North Wales). The longest period of ownership was from 1984 to 2011, when it was ‘living’ in Croydon.

Laurent has been able to fill in one of the gaps in the car’s history i.e. during the 1970s. A request for information in TTT 2 back in 2014 bore fruit. However, he is still missing details of the first owner from 1953 to 1960.

Perhaps this latest mention, might produce something? Laurent is hoping to come to the MG Centenary at Gaydon in May and I am looking forward to meeting him.

2 thoughts on “Cleaning the magnetic Smiths speedometer (Part 1)

  1. Mike Leadbeater says:

    Thanks for the excellent advice. My TD’s speedo , as well as being a mile (sorry) out of calibration due to fitting a Ford ‘box and 4.55 axle ratio, has a greatly oscillating arm.
    The mechanism, having done 74k miles, is greatly worn. I have detected a lot of end float, maybe 2mm or so, in the aluminium cup, due, I believe, to wear on the outer cup spindle bearing which is basically a pointed needle. As the cup floats back and forth, the torque it receives from the rotating magnet will vary, causing this wavering. Do you agree?
    I am considering making a repair of this bearing , which looks fiddly.
    Another issue is the new cable , acting I guess as an Archimedean Screw, brings oil up from the gearbox, which then clogs the magnet and cup, causing excess drag and making the arm to wildly fluctuate.
    Any suggestions on how to cure this problem?

  2. CASTEL says:

    Dear Mike,
    For sure, end float of 2mm is way too big. You’re right. It is impossible to have a constant magnetic coupling. You first have to fix this. There is a slotted screw just behing the dial, close to the needle shaft. You can adjust the end play with this screw. See figure 17 of Anthony’s paper. Not sure it will be enough. Best solution to prevent oil from the cable is to fit Doug Pelton’s mod for speedo pinion : Also described in TTT2 N°1. You can also fit a speedo gearbox suited for your rear end ratio.

Comments are closed.