December, 2011


I pulled out the speedometer and tachometer a couple of weeks ago to try to decide what to do about them.

They are both the original, stock  instruments that came on the bike.  The speedometer was in bad shape.  It was weatherbeaten, the pointer was broken off, and both odometer  assemblies were crooked in their windows.  The  tach looked a little better, but I seemed to remember that it was very jumpy the last time I rode the bike.

I had several choices regarding the instruments:  Aftermarket reproduction units are apparently available.  There are also a number of outfits that would rebuild these units, or accept my units as cores, and sell me rebuilt ones off the shelf.  The research I did showed that neither of these was necessarily a bad choice.  

The other option was to try to rebuild them myself.  
Since the this option is more in keeping with the spirit of the entire project, this is the route I chose.

I did some research, and found half a dozen web sites that have some information about it, and also a book on the subject.  I ordered "Magnetic Speedometer Repair" by Graham Blighe.  Though I was already fairly familiar with working on small mechanisms like this, the book probably helped me avoid a pitfall or two.

Following is the story of what I did to my instruments.

It turns out that one of the most difficult steps is the first one:  getting the crimped bezels off.  The sources say that if the bezel is carefully pried off little by little around the perimeter with a small screwdriver, that the bezel might be reused.  Once I got the hang of it, it wasn't too bad.   When I was finished, those bezels didn't look like anything I'd want to reuse.

Most of the pictures that follow are of the speedometer.  The tachometer is identical inside, except that it doesn't have the odometers.

There is a triangle-section rubber seal between the bezel and the glass, another rubber seal under the glass, and a circular mask piece that hides the space at the edge of the dial, but allows light from the internal bulb to illuminate the dial.  The rubber seals were hard and brittle, and basically fell apart on disassembly.

Undoing the three screws on the back of the case, the mechanism can be lifted out.  When I did that, the main odometer assembly fell out onto the table.

The internal mechanism can look a little intimidating at first, but it is really pretty simple.

After the pointer needle (what was left of it) was pried off, the dial came off by drifting out the center keepers of the two small reusable plastic rivets.  This reveals the internal mechanism.  All magnetic speedometers and odometers I've had apart employ the same basic principles, and here's a synopsis of how it works (see annotated pic below):  The speedometer (or tach) cable connects to the bottom of the casting, and the cable's inner flexible shaft turns the input spindle of the instrument, which has both the brass worm gear and the dark grey magnet pressed onto it, so the gear and the magnet both spin at the same speed as the speedometer cable core.  The worm gear drives the odometers through a chain of plastic gears.  Just above the magnet, but not touching it, is an aluminum disc.  The disc is on a smaller spindle that extends up through a hole in the brass spindle plate, and down into a recess in the center of the magnet.  The hole in the spindle plate and the recess in the center of the magnet serve to locate the spindle so that the disc is always parallel to, but not touching, the magnet.  The upper part of the disc spindle has a collar pressed onto it above the spindle plate such that the disc and spindle are hanging from the plate, supported by the collar.  The collar also holds one end of a spiral return spring so that when the disc turns, it winds the spring.

Now, even though aluminum is not magnetic, it is conductive, so when the nearby magnet is spinning, electric currents start to flow in the aluminum disc. These currents create their own mangnetic field which reacts with the field of the magnet such that the disc tries to spin with the magnet.  The torque created by the disk trying to follow the magnet is opposed by the spring.  Since the torque on the disc goes up as the magnet spins faster, the result is that the disc rotates against the spring in proportion to how fast the magnet is spinning.  The indicator needle is attached to the spindle, so the needle shows how far the disc rotates.

The steel modulator plate is provided as part of the magnetic circuit so that the strength of the magnetic field that the disc sees from the magnet can be varied.  If the modulator plate is moved closer to the disc, the field is stronger, and the torque for a given magnet speed is increased,  Adjusting the modulator plate is how the speedometer (or tach) is calibrated.  There is a screw accessible on the back of the  main casting that moves the modulator plate against the spring that can be seen to the right of the magnet.

Here are some other views of the disc assembly (disc and spindle, modulator plate, spindle plate, collar and spring) removed.  The assembly is held to the casting by more of those reusable plastic rivets.

In order to inspect for wear, the disc assembly has to be taken apart.  This is done by pulling off the pressed on collar from the disc's spindle.  Shown is  a collar removal tool like those used by RC model car hobbyists.  It inserts under spindle plate, and the screw pushes the spindle through the collar.

One thing I nearly lost was a tiny washer under the spring collar on both instruments.  In effect, this is a thrust bearing that takes the weight of the disc and spindle while minimizing contact area.  

Here is one thing I found that explains a lot.  These are the spindle plates from the two instruments.  The hole near the center of the plate is the locating hole for the disc spindle.  The spindle where it passes through the hole is slightly less than one mm in diameter, and the hole diameter should therefore presumably be about one mm. The hole in the bar to the rear is very slightly oversized, but the one on the front is a little over 2 mm!  It was pretty obvious that this needed to be fixed.

I fixed the problem by using some small telecoping brass tubing from a hobby store.  The larger tube measures 0.093 OD, which was very slightly larger than the hogged out hole in the spindle plate.  A #42 drill is 0.0935, and it made a nice fit for the tubing.  Luckily, the enlarged hole was still centered in the plate, so drilling was simple.

With all the parts being brass, soldering is the obvious way to fix the nested tubes into the hole:

I cut away the excess tube and leveled the surface.  I'm not sure what the ID of the smaller tube is, but it's less than 1 mm, so I drilled to 1 mm.

There is one additional part in the disc assembly.  There is a tiny plastic sleeve that goes on the disc spindle just under the spindle plate.  It is a sliding fit on the disc spindle, and it has an arm that sticks out into a slot on the spindle plate.  The arm keeps the sleeve from turning with the spindle, so it provides a little friction to the turning of the spindle,  This serves to damp rapid fluctuations in the spindle (and thus the needle).  Well, some regrettable haste led to breaking the tiny arm off one of the sleeves.  

One thing that occurred to me was to get a junk instrument and take the sleeve from it (maybe an indicator needle, too).  I found after calling around, though, that since even junk units are valuable as cores for rebuild, no one had junk speedos or tachs around.  If I had looked harder or if there had been any swap meets handy, I might have found one, but I took another path and had a fixed sleeve an hour later.  The OD of the sleeve is about 0.062", so a 1/16 drill through an old credit card, some Exacto work, some cyanoacrylate glue formulated for plastic, and I had a functional sleeve again.

Replated the steel modulator plates and other steel parts:

Now I could put the disc assemblies back together.  The modified C clamp tool helped press the spring collar back onto the spindle in a very controlled way.  The collar should go on far enough so that there is no more than a few thousandths of vertical play in the spindle.

The disc assembly ready to go back on the main casting:

There was one more bit of unpleasantness in the speedometer to take care of.  The main odometer assembly had come loose from its moorings, but this was really just because there wasn't enough of a deformation of the metal over where the axle of the assembly fit into the main casting.  I replaced the odometer and put a little RTV over the recesses to hold it in.  The trip odometer was a different story.  First, this didn't look right:

Took the trip odometer assembly out, then took it apart to see how it worked.  There was a while there where I was sorry I did that.

The problem turned out to be that this shaft is supposed to have that toothed wheel peened onto the end of it, and the wheel had fallen off.  Luckily, it wasn't lost, and I was able to fix it back on.  The wheel is part of the reset mechanism.

Now for reassembly.  Here's the cleaned up casting.  All the plastic gears and such were in good shape, so I didn't remove them.  

Next picture is with the input spindle installed and the worm gear and magnet pressed on.  I realized one thing in looking closely at these parts.  I've often heard that if a speedometer or tach inner cable is too long, it will ruin the tach or speedo by intruding too far into the guts of the mechanism.  At least on these instruments, that can't happen.  First, the square recess in the input spindle that accepts the inner tach or speedo cable is a blind hole. Nothing can get into the works through that hole.  What can happen, though, is that since the input spindle has a flange at the bottom that limits how far it will go in, excessive force on the input spindle from an inner cable that's too long would put pressure on the flange, and cause increased wear.  This would eventually but surely cause the spindle to move upward to contact the aluminum disc.  Intermittent contact of the magnet with the disc causes the pointer to jump wildly, and will eventually break something.  

Then the trip odometer.  There is a little gearbox that receives the odometer reset shaft.  Smiths loved to use those little tapered pins to hold things.

Here is the speedometer with the main odometer and the disc assembly mounted.  The tach is simple by comparison.

Then there was the matter of the broken pointer needle on the speedometer.  Since it didn't look like I was going to find a junk unit, and I couldn't find anything suitable online, I had no choice but to try to dummy one up.

The pointers were made of aluminum, about 0.010 thick.  I found some flashing tha fit the bill.  I had to buy 10 square feet of it.

Cut away a little more of the old pointer under the hub, and fastened the new pointer with metal filled epoxy.

New fluorescent orange paint:

Put the faces and pointers back on , and now it's time to calibrate.  I built a jig to hold the instrument, a variable speed drive motor, and an optical tachometer.  The stand is from a little Dremel drill press. At first I tried to use the Dremel tool as a motor to spin the instruments, but after I got everything made and set up, I found the Dremel turns counterclockwise.  I took it apart to see if I could remedy that.  It was a permanent magnet motor, and while they are easy to reverse in principle, this one was so tight inside, there wasn't a good way to do it.  Next try was a Craftsman chainsaw sharpener.  It too ran counterclockwise, but was a series wound universal motor, so all I had to do was reverse the brush wires.  The shaft from the motor to the instrument is 1/8" key stock turned to 1/8" round to fit the motor's collet.  I ended up cutting the shaft in half and joining it with a piece of rubber tubing to accommodate a little misalignment better. I used an electronic motor speed control to vary the speed.

First picture shows the adaptor to mount the instrument on using the three threaded holes in the back ofthe casting.  What looks like maybe 8-32 threads are really 4BA, which is 0.142 x 38.5 tpi.  The calibration screw is also visible.

The tach is a 4:1 instrument (as indicated on the face shown in the second picture at top), so 1000 rpm on the input shaft should read 4000 rpm on the tach dial.  Turning the calibration screw on the back of the unit moves the modulation plate, varying the magnetic field the aluminum disk sees.  Smiths speedometers like this have a marking on the bottom of the face that indicates the number of revolutions of the cable that will rack up one mile on the odometer.  On mine, it was 1600 (see very first picture above).  This implies that 1600 rpm on the input cable should indicate 60 mph (one mile per minute).  Both instruments calibrated fine, and appeared to be working well.

So with the guts in servicable order, I turned to the "cans", or shells that the instruments mount in.  They were in good shape, but the paint and the rubber parts had seen better days.

The cans cleaned up well with a little blasting and paint.  The inside is painted white so that more of the light from the internal instrument light makes it to the dial face.  There was a greyish area painted on the originals, which I assume is to mute the light right above the lamp, so as to even out the light on the face.  It hadn't been applied yet in these pics.

The rubber parts weren't so easy.  There are six  grommets in each instrument that serve to isolate the internal mechanism from vibration or shock. These were all badly distorted and deteriorated.  There was also a rubber disk  with a slit to seal around the trip odometer reset shaft.

I found a place claiming to sell grommet/hardware kits, but they didn't really have them available.  The fallback position was to make new ones, but since that process was fairly involved and not pertinent to this story, I'll skip over it here.  If anyone is really interested, follow this link.  Here are the new grommets, and a new little gasket for the trip odometer reset tool.  It was punched out of some rubber sheet with that one-use punch made from a piece of sharpened electrical conduit:  

The replated the special mounting hardware.  Those screws are 4BA threads.  Not much chance of finding those at Home Depot.

Everything ready to go together.

There is also a little foam cushion that goes in the bottom of the can.

Made the trip odometer reset tool, since mine was long gone.   Blighe's book shows a dimensioned drawing of the tool.

Ready for bezels:

Bezel kits are available, and I ordered a couple.  As anyone who fusses with old cars or bikes knows, reproduction parts are always a bit of a crap shoot. The first bezels I got for a good price were slightly too small, and would have required grinding a bit off the flanges of the instrument cases.  I might have done that, but I also thought the quality of the rubber gaskets was shoddy.  The next sets I got from a different supplier were much better.  Even with the second pair of kits, one of them had a very visible scratch in the glass.  My 40-year old original glass was better, so I re-used it.


The bezels must be crimped onto the instrument cases sufficiently tight to compress the rubber gaskets enough for a water seal.  There are sources, including Blighe's book  that suggest that the crimping can be done using simple hand tools.  I was skeptical that a reliable seal could be guaranteed this way, so I did a little more research.

One very nice blog showed a picture of the "Smiths factory tool" for rolling the crimp:

I considered building the tool, but the blog also describes a method using a lathe to press and turn the instrument while the crimp is rolled.  I have a small lathe, so this seemed to be the better approach.  First, simple adaptors must be made to fit the front and back of the instrument to hold it in the lathe.

Also a ball bearing is fitted to a piece of key stock that is held by the lathe's tool post.  The bearing is what rolls the crimp.

I didn't use power in the lathe--just turned the chuck by hand as the tool was slowly advanced.  The crimping operation flakes some of the chrome plating off the brass bezel where the tool rubs on it.  Not sure how to fix that, but it is pretty much out of sight anyway.

The finished speedometer.  On the tachometer, which I did first, I put a  thin smear of RTV on the gasket between the case and the glass to ensure water tightness.  This was  a bad idea, since the RTV made the gasket slippery enough to extrude out of place when the compression pressure was applied. Of course this couldn't be seen until the crimped instrument was taken out of the lathe.  I had to remove the bezel, and order another one.

The instrument lights are pretty crude little lamp holders that snap into the back of the instrument cans.  They were grungy, and one of the hot wires had broken off the internal lamp contact, but they were otherwise undamaged.  These aren't really designed to be repaired, and replacements aren't all that expensive.  On the other hand, it appeared that I could bring these back to life with an hour or two of attention, and I'd save 20 dollars.  Besides, judging from experience, the replacements would likely be of lower quality than the originals.

The most challenging part was to pry open the tiny hot contacts so I could re-crimp and solder the wires to them.  Then I balsted and painted the holders and put them all back together.  I put real solder-on bullets on the ground wires instead of those tacky little quasi bullets that the bare wire just wraps around.  Also extended the heat shrink up farther toward the instrument.

The instrument mounting bracket is fastened to the steering top lug with two resilient "Metalastic" rubber bushes to isolate vibration.  The bushes are pressed into housings on the bracket.  They have to go in far enough so that when mounted, the bracket is held away from the top lug.

The finished instrument cluster:

To Other Pages

Send comments to: