Club 80-90 Forums

My speedo reads much too high, and it was starting to bug me. So on a recent trip I decided to record the indicated speed against the 'true' speed as given by my satnav. I made this plot.



It's not a straight line, but a best-fit straight line is close enough. It tells me that there's an offset of 5.3mph and a scaling error of 1.08. I have standard wheels and tyres, so it's definitely the speedo's fault. The odometer seems fine, but although both the speedo and the odometer are driven by the same cable, they use different mechanisms - a nice direct geared drive for the odometer and a spring for the speedo needle.

So with a rush of blood to the head I decided to see if I could adjust the speedo. When you remove the speedo face you can see a spring that controls the needle.



Since my needle moves 8% too much in response to the cable rotation rate, I figured the spring was too loose. I guess it makes sense that over the years the spring could have relaxed a little bit. I counted about 6 coils in the spring, and figured that the 8% adjustment I needed to make corresponded to about half a turn.

If you look at the captive outer end of the spring you can see that it's trapped against the alloy casting by a brass pin. I levered the pin out by pushing a flat-bladed screw driver under it, below the casting. Then I pulled the spring through by about half a turn, and wedged it back in place with the brass pin.



I also adjusted the needle position to try and account for the 5.3mph offset, i.e. I just pushed it onto the spindle so that it's rest position was a few mph lower than before, then I lifted it over the end stop. I reassembled it and put it back in the van. I took it for a test drive and it's not bad. Unfortunately I've gone a bit too far the other way - at an indicated 65mph I'm actually doing 70mph. That's not good - I don't feel safe from speed cameras if my speedo under reads. But at least it's proven the concept, and with another iteration I should be able to get it close enough.

That's really interesting - thanks.
I always thought that if my speedo was registering 10% high, the odometer was always counting high too. So possibly not the case.

I'm not sure I will go to the trouble of adjusting as you did. I can work out the actual speed quite easily as 55 mph indicated is actually 50 mph and 33 indicated is 30 etc. etc.

a1winchester wrote: I always thought that if my speedo was registering 10% high, the odometer was always counting high too. So possibly not the case.

Yeah, me too. But having seen the mechanisms with my own eyes I'm happy that they are somewhat independent. The odomoter counts the number of turns of the cable, and that's directly related to the number of wheel rotations. The speedo needle moves in response, I think, to a rotating magnet - the faster it spins the more it pulls against the spring, so the spring directly affects the indicated speed. Of course, a change in wheel size will affect both, but with standard wheels the odometer should be correct.

I tweaked the adjustment again this evening, now it’s almost perfect - within plus or minus 1 or 2 mph all the way to 70mph. What with that, and curing the ‘bounce’ a few months ago with a new cable, it feels like a different van!

Only 70mph?, you do have a DJ, we need graphs up to top speed and we want to know the incline, wind speed and road surface coefficients

Its when nobody replies you have to worry

Your research is excellent Chris, I only wish I had the time.

It's ok - no offence taken. Received in the spirit in which it was intended.

If I was worried about being considered a nerd I wouldn't post nerdy stuff.

Nerdy or Geeky?

That's me on the right.

That's Paul on the left................

Great post - mine reads about 5mph low. It’s this sort of research that gives relatively new members the confidence to undo screws and delve into things ( well, me anyway). Thanks

I took the van out to record the indicated speed against my satnav speed again, post-adjustment. The red line is the original, the blue line is after adjustment.



It's not as good as I thought. The green line is the theoretically perfect result. Although my speedo is within 1-2mph everywhere above 30mph, at an indicated 20mph the van's only doing 16mph, and at high speeds it's starting to read a little low. I'm quite happy with it this way round - slightly over reading at low speeds where I don't want to trigger a speed camera, and only under reading at high speed where a couple of mph would be quite safe.

But that chart shows that there's still a trend in there that could be removed with a bit more adjustment. The dots fit a straight line quite nicely. That best-fit straight line equation, y = 0.91x+5.08, means the needle is moving 91% as much as it should do (so the spring is now too tight), and there's a 5mph offset.

The offset is tricky, because the needle has to come off each time the face comes off to adjust the spring, so there's a bit of guesswork each time it's refitted. The 91% is actually a bigger error than the 108% I started with! So I think I'll have one more go - I'll slacken the spring by half as much as I tightened it, and I'll take a bit more care with the needle position.

It's a bit of a nuisance having to remove the dash completely each time I want to adjust the speedo - there are so many fiddly electrical connections. Last night I spent a good hour or so trying to get my LED dashboard lights to work. I used my multimeter to try and trace a drop of about 6v at the bulb connections. When I traced it to the dashboard light switch it was one of those D'Oh! moments - I'd nudged the rheostat for adjusting the brightness!

Next time I might refit everything but leave the speedo separate from the binnacle, just in case I want more than one 'final' iteration.