Inaccurate speedo

[MagicMatch]

Hi,
When I bought the car the speedo needle dance around all over the place, and the reading was way out – been driving on the rev counter. I've just had the dancing around fixed, but the reading is still inaccurate. I'm wondering if the right diff is on there? It's a '76 auto, what should I have and how do I check what's fitted. Numpty alert! I'm useless at mechanics, so please excuse my ignorance.

[Goldstar]

With the diff on the car it’s not easy to check the ratio but if you’re not using a standard 3.7:1 ratio diff it will make a difference.

What size tyres do you have fitted as that also affects the speedo. Has the speedo ever been changed as I’m pretty sure that different speedos have different final drive ratios.

[Neil Chatterton]

My speedo under reads quite significantly but if you use Google maps on your phone it will show the correct speed based on GPS. Once you know the difference at different speeds you can be mindful of it where you need to watch your speed.

[trunt]

First I would check the speed shown at low speed and high speed, if it’s constantly high or low it may be a simple matter of “slipping” the needle a little to get it correct. The needle is a press fit on its shaft, you can remove the Speedo from the casing hold the rear disk and “slip” the needle. Eg it’s showing 60 at real speed of 50, hold the disk at 60 and turn the needle back to indicate 50.

[Philip Wardle]

From my archives ... this Practical Classics article about speedo calibration published in January 1985. The text will be difficult to read in the image, so here's what it says: -

Following on from what I was saying about buying starters or dynamos 'blind' at a breakers or autojumble, a number of readers have asked me to give some pointers on buying second-hand instruments. Most home restorers fight shy of repairing instruments though it's possible with care to do quite a bit of repair yourself, and maybe I'll go into that in a future issue. Speedometers, though, don't respond kindly to amateur exploration in their innards, so this month we'll assume yours has packed up and you're looking for a second-hand replacement that will fit straight in and work.

Speedometers can sometimes be a chancy buy — not so much because they may not be working but because they may be calibrated wrongly for your car. Don't assume that just because the speedo head looks the same as yours that it came off an identical model. It could have come from the van which is derived from the car, it could have come from an up-market badge-engineered version with a higher output engine and a different rear axle or, if you've got the up-market version, Sod's Law says that the speedo head on the autojumble table will have come from the bog-standard version.

Either way, don't expect .too much help from the stall holder. You might be lucky, and he might be able to tell you that it definitely came from a 1963 1.6 litre Puffmobile, but more than likely he won't be certain or, and I've noticed this as an unfortunate tendency recently, he'll say anything to make a sale such as: "That'll fit, mate. Austins, Morrises, Wolseleys, MGs, they were all the same round about that time. "

Of course, they weren't. And one thing in which they often differed was the rear axle ratio which, in turn, meant that the speedometer calibration had to be different to indicate the speed with any reasonable accuracy.

To help sort things out on the production line, where mistakes are not unknown, as well as for their agents and repairers, Smiths and other speedometer makers often marked the dial with a code number indicating the calibration.

This number indicates the number of turns per mile, usually abbreviated to TPM, which the cable makes. You're looking for a figure such as 1180 or 1250 or something of that order.

If it isn't marked on your speedo head, and you want to know what it should be, there’s quite a simple way to check. All you need is a piece of cardboard, a piece of chalk and a helper.

First of all, check that you've got the same size rear tyres (or front tyres if it's a front wheel drive car) as the original specification — changing from cross ply to radial ply will make a difference because the number of wheel- turns per mile is not the same for a cross ply tyre and its equivalent sized radial. So, if you're keen on getting things as accurate as possible and you've got radials when the original specification used cross ply, see your local tyre specialist, and ask if you can look in his technical manual to get the relevant wheel turns per mile of equivalent sizes of cross ply and radial tyres, and then apply a correction factor.

Now check the TPM of your car. Take the drive cable off the back of the speedo and push a cardboard pointer over the end of the inner cable so you can count how many times it goes round. Then find a bit of level ground, check that your tyre pressures are right and mark the lowest point on one of the rear tyres with a piece of chalk. Then, while you lord it in the driving seat to count the number of turns and part turns which your cardboard pointer makes, get one of your mates to push the car along so that the marked wheel makes exactly six turns.

To get the number of TPM of your speedo cable from this information is a simple matter of putting your figures in the simple formula:

TPM = ( 1680 x N ) / R

In this formula, N is the number of turns your cardboard pointer made for six turns of the marked wheel, and R is the distance from the centre of the wheel to the ground, measured in inches. Try to count the turns to the nearest eighth of a turn.

So if, for example, your pointer made 9 1/8 turns (or 9.125 to use a calculator and save the longhand maths), and the distance from the centre of the wheel to the ground was 12.5 inches, the formula becomes:

TPM = ( 1680 x 9.125 ) / 12.5

To the nearest whole number this works out at 1226. Don't expect to be absolutely spot on because of the practical errors in what is, after all, a pretty rough and ready way of experimental working compared with a laboratory. But you ought with care to be within 20 or so of the right answer, which is around 2%, and ordinary speedos don't indicate any more accurately than that.

Apart from checking that the TPM is right for your car, you can't do a lot more at an autojumble stall except to watch out for speedos which have got oil in them - you can usually see signs on the casing somewhere. It got there because someone poured oil in the drive cable instead of greasing it. If you do this, the oil works its way up into the head and gunges is it up. The only lubricant that you should use in a speedo cable is grease, and then only sparingly, particularly near the top.

speedo calibration.jpg

[gbb483]

As the milometer uses the same feed, check the distance travelled as indicated by GPS against the distance travelled as indicated by the trip meter.

[trunt]

To find the diff final ratio you can just mark the prop shaft and a tyre and push the car one full revolution of the wheel. the prop will turn either 3.4 times or 3.7 times, you should be able to see the difference i.e. is it an extra 1/2 turn or 3/4 turn..

Or check rpm at a particular speed.. The rom has rpm for 10mph in top gear for the 3.7 in 3 tyre sizes. so e.g. 50mph on 185-14 is 2525 rpm.. assuming your rev counter is accurate!

[Goldstar]

To find the diff final ratio you can just mark the prop shaft and a tyre and push the car one full revolution of the wheel. the prop will turn either 3.4 times or 3.7 times, you should be able to see the difference i.e. is it an extra 1/2 turn or 3/4 turn..

just mark the propshaft and a tyre, we're not all as supple as you Terry

[trunt]

just mark the propshaft and a tyre, we're not all as supple as you Terry

If you mark the diff input flange whilst lying on the floor next to the car you could mark the tyre at the same time.. no contortions required!

[MagicMatch]

My word everyone, thanks for your input. It seems I have a few things to check so I'll test the consistency of the speed, tyre size, and distance travelled. I've obviously not told the full story yet.