Fellas, there is a VERY precise method of working through speedo and gearing etc. that I have basically figured out.

Principles:

1) Speedometers/Odometers like to have the speedo cable at 1000 rotations per mile. At 60 mph, of course, the speedo cable is rotating at 1000 rpm or "rotations per minute" as well. This is the way that speedos have been for quite a while.

2) DRIVE gears, the ones on the tail of the tranny, are spiral gears. Spiral gears work like this in the calculation: 6 "tooth" spiral = 6000 "teeth" go by per 1000 rpm of tailshaft. 7=7000 etc. The larger the tooth count on a spiral geat, the FASTER the speedo runs (FYI -- the spiral gear in my gallery is a 7 toother, and teeth are counted by looking at the end of the gear and NOT by the number that are present on the "side" -- Its like counting petals on a flower -- look at it like it was a donut/life-ring)

3) DRIVEN gears (at the end of the speedo cable) vary the speedo by 1/15 to 1/19 in a logical fashion per single tooth change. That amounts to a 5-7% change in speed shown, per tooth change. 15 teeth are the fastest gears (show fast speed), 19 are the slowest. You would use a 19 with a rear end with very high gear ratios (like 4.50), or very small tires. A 15 is useful with large tires or a very low (numerical, like 2.75) rear.

4) There are good tables that give the effective circumference of most tires. Using that data, in inches, you can calculate the RPM that your driveshaft will be turning to get you 60 mph. Manufacturers often public rotations per mile data -- that is the very best, as it accounts for "squat" that effects tires actual rotations.

5) Once you have that #, you can then determine the DRIVE and DRIVEN gear sizes that will give you close to a perfect 1000rpm in the speedo cable.


So, for instance: I have 205/75R15 tires which have a rolling circumference of 85.16" 1 mile = 63,360 inches So my tires rotate 744 times per mile.

Since I have a 3.25 rear ratio, the driveshaft is rotating 3.25 times as fast as the tires, or 2418 rpm.

Since I have a 7 tooth DRIVE gear, there are 16,926 "teeth" engaging per mile (2418 x 7).

To get to the desired 1000 rotations per mile of the odometer, I can use a 17 tooth DRIVEN gear. (16,926 divided by 17 equals 996 which is pretty close to correct)



Here's a link to a decent "rolling circumference table" for future reference if anyone needs it.

http://www.gnttype.org/techarea/susp...irechart2.html


Finally, when a speedo and odometer head is working correctly, the speed and the odometer BOTH ought to show correctly. Personally, if I had to accept that one or the other is off, then I'd rather have the speedo off. When a speedo is off, assuming that the odometer is correct, then the problem is usually either the magnet is weak, or there are rough spots in the head that need lubed. But if there are rough spots, then usually the needle "bounces" -- which is also the case when the speedo cable is the problem.

The odometer is a *relatively* simple set of gears -- the speedo has more vague parts such as springs and magnets. There are speedo shops still that can fix the speedometer. The are usually quite reluctant to fix or reset odometer readings, for obvious reasons.

There........

 

cdherman, a highly informative post thanks. One thing, I would rather have speedo reading correctly and odie off as I have never had a ticket because odie was 7 mph off. A mental note of odie error, reads 11% slow for fuel mileage. This is because I don't want to pull the dash apart and send speedo/odie to a repair shop and get told "can't repair or adjust" as this has happened in the past.
.....=o&o>.....

 

To each his own, with regards to the speedo versus odometer error. At least in some parts, there are still enough junked slicks laying around that one can find a few speedo heads pretty cheap -- I have a couple spares. Swapping them out is not much trouble.

I got lucky once upon a time and found one with only a few hundred miles showing -- it was obviosuly turned over, but close to zero. So I ran it back to zeros and installed when I finished my frame off restore. Now I always know how many miles I have on her since the restoration!! Running the old speedos backwards does not harm them, FYI -- but of course the usual caveat regarding legal issues should be mentioned.

 

That would be like the 68 when it had to have smog 10 years ago and odie said 69K miles, smog computer stated "too low of mileage for year of vehicle" and would not acknowledge mileage entered, Screw 'em and DMV as this was original mileage. Got a 81 DeLorean with 314 original miles, same thing. All I want is a speedo that's half a$$ accurate as they are the craps above 70 mph and read slow. Above 100 they run out of mph numbers anyway. "Each is own".
.....=o&o>.....

 

OK, I've been reading these posts with interest since my own speedo is off (PO switched from a 3-sp to a T-98, speedo reads almost 10 mph slow).

But simple mechanics says that when 2 gears mesh, merely changing one gear's tooth count will either cause binding or incomplete engagement, since the 2 gear's pitch diameters don't match anymore.

How does the speedo drive get around this? Don't you have to switch both gears at once?

 

they do it by using the same pitch but different number of teeth thus the centerline of the shaft is moved to get the right mesh. It's not just a gear you get its a gear on a shaft with a bushings that goes in to the trans the hole in the bushing is moved "off center" to make room for more teeth

 

Brider

I am a mechanical engineer and was thinking the same thing but so many people told me to just change the one gear I figured they were talking from experiance. And it worked fine. I did take a close look and the gears are actualy the same diameter same length and the only differance is the amount of teeth.

jd

 

There is a limit in the speedo gear tooth numbers before you have to change the metal gear on the tail shaft.
.....=o&o>.....

 

That is what we were trying to say 11 days ago. Don't worry about the color or colour just the tooth count and pitch.

There are way too many variables to get one exact. Close is all that you should hope for. IMHO

John