JohnnyBs68S

Just finishing up my 200-4R trans swap in my '68S and was checking my speedometer against my GPS to see how close it is. I used all the proper "calculators" to determine the speedo gear that was needed in the trans:

-Tires: 245-60-14 (~26" tall)
-Rear: 3.90 (Richmond gears, not factory)
-Trans speedo drive gear: 10 teeth
-Trans speedo driven gear: 31 (from calculator)

At the indicated speeds of 40, 50 and 60 MPH, the GPS was reading 44, 54 and 64 respectively (4 MPH higher in each case).

It appears that my speedo gear ratio calculations ("gain") are correct (a 10 MPH increase in actual speed results in a 10 MPH increase in indicated speed), but I have an "offset" of ~ 4 MPH in my speedo needle position relative to the dial. If I was able to rotate the needle movement relative to the dial ~4 MPH, it'd be nuts-on. So my question is, is there any way to adjust this "offset" to correct this? I know in the electrical world that there is a "zero" adjustment for most analog (electro-mechanical) meters that allows you to zero the needle with no stimuluation applied (this corrects for stray magnetic fields acting on the needle movement or temperature effects on the movement return spring). I searched the service manual and can't find anything that goes into that level of detail for the speedometer (only how to remove and replace it in the dash). Any help is appreciated.

Yellowstatue

It would be much easier to change the driven gear, and cheaper too.

hookem horns

Mechanical speedos are ususally off by a percentage, not a set amount as you say. Yours is pretty close so maybe you cannot tell the ~1 MPH difference on the speedo comparing 40 to 60 with just a 6% error overall. Agree you can easily swap the driven gear to get closer to actual speed.

JohnnyBs68S

Indeed, a speedo driven gear change would be simplier than taking the speedometer out of the dash and re-setting the "zero" point of the needle to be 4 MPH higher. However, changing the speedo driven gear will result in the indicated reading being too high at higher speeds and too low at lower speeds. The engineer in me wants to resist this compromise if there is a way to avoid it. My hope was that there was a "zeroing" trim screw adjustment on the back of the speedo that I could just adjust w/ a screwdriver w/o taking anything out of the dash that would remedy this issue. Apparently, there is no easy way to accomplish this?

I did some calculations, if I switch to a 29-tooth driven gear, the error at 64 MPH reduces to 0.13 MPH (too high), while at 44 MPH the error is 1.2 MPH (too low) and at 84 MPH (assuming the offset at that speed is still 4 MPH) the error is 1.5 MPH (too high). This is probably the best compromize that can be achieved w/o re-working the speedometer itself.

RandyS

There isn't a way to create an offset in the speedo head. Since you are working with a mechanicaly driven speedo, the error in speed indicated vs true speed will always be the same %. It isn't a variable drive. The optimum way to work toward 100% accuracy is to change the driven gear. You CAN chage the drive gear, but you need to disassemble the tranny to do that. If you are dead set on getting as close to 100% as you possibly can, you may have to look at different rear tire sizes as well. .........

JohnnyBs68S

I would agree that there isn't an easy way to create (or correct) an offset in the mechanical speedometer head, but it is possible to remove the indicator needle from the shaft that turns it, reattach it at a slightly different "zero" position, and you've effectively changed the offset of the readout.

And yet in my case, a 4 MPH error at 40 MPH is a 10% error, while the 4 MPH error at 60 MPH is a 6.7% error. So this is not always the case.

Agreed. The drive to the speedo via the cable is very linear and is "scaled" by tire size, rear axle ratio, and the speedo drive/driven gear ratio. What is non-linear in this case is the readout (speedometer indicated speed). If I were to extrapolate the data points I obtained, my speedo would be reading "0" when I was moving 4 MPH, and it also reads "0" when I'm stopped. My guess is that for whatever reason (46 years of aging?), that my speedo's "zero" point (where it relaxes with no rotation from the speedo cable) is at -4 MPH, but it's prevented from pointing at anything below "0" MPH by some internal mechanical stop. Thus, I have to be moving a 4 MPH before the needle moves off of "0". Once the needle moves off of "0" though, it responds very linearily to any additional speedo cable velocity (i.e.: a 10 MPH increase in actual speed results in a 10 MPH increase in indicated speed). Thus, the slope (gain / sensitivity that is affected by tires/rear gear/speedo gears) is correct, but everything is offset by 4 MPH.

Well, if there is an inherent offset in the readout, changing the "gain" (aka: slope of the transfer function) of the system that converts road speed to speedo cable rotational velocity (i.e.: the tires/rear gear/speedo gears) will allow you to null-out the error at a specific speed (as I've shown above), but cannot null out the errors over the entire range of speeds of interest. While changing something in the tires/rear gear/speedo gears may reduce the average error over the range of speeds of interest, it's not going to result in removal of 100% of the error over the entire speed range since it does nothing to reduce the inherent offset.

Maybe I need to draw a picture of what I'm talking about?

Most folks who "calibrate" a speedometer by changing the speedo gears in the transmission are merely doing a "2-point" calibration (actually, they assume that one of the 2 points, the one a 0 MPH, is inherently correct, thus they end up w/ a 1-point calibration at the speed where the error is measured). This works in most cases since any errors at 0 MPH are usually small and as long as the error up in the speed range of interest (say 40-80 MPH) is small, who cares.

To be truly accurate over the entire range of the readout, a multi-point calibration must be performed (i.e.: error measured at several points from the low end of the readout to the high end of the readout). This data is then used to determine a best-fit straight line (assuming a "linear" readout, Y = mX + b) whose slope (m) and offset (b) can be determined and compensated for either though math or adjustments of the readout. This is what I've attempted to do w/ my speedometer readout (a "3-point" calibration if you will, and ignoring zero as one of those points), and the data indicates that the slope is correct, but a -4 MPH offset exists.

So it appears that my choices are:
1) Take the speedometer out of the dash, disassemble it, and attempt to correct the -4 MPH offset
2) Change to a 29 tooth driven gear in the trans and try to remember that the speedo is reading slightly too low below ~64 MPH and slightly too high above it
3) Do nothing and try to remember that the speedo is reading 4 MPH too low regardless of the indicated speed.

RandyS

Jeez, try to help a guy out .........
And no professor, no pics needed..........

pcard

Is it just the speed that is being indicated incorrectly, or is the odometer showing a difference from the GPS as well?
The 4mph error throughout your speed range indicates a "translation" error, which would not happen if your gearing was faulty. So I would think your odometer will be correct in which case I would leave the gearing alone and mentally adjust.
I cannot imagine what would cause this other than a misalignment of the speedo dial or indicator.

JohnnyBs68S

I know! Right? The nerve of some people......

Seriously, I appreciate the engaging discussion. But you're a little too late (see attached pic).

JohnnyBs68S

Good point on the ODO, I'd have to drive a few miles on a road w/ mile-markers in order to discern a "distance traveled" error that would throw off the ODO (haven't done that yet). Not sure if my GPS can show distance traveled that precisely, haven't played w/ all the settings / modes / options yet. But yeah, a 6.7% error on distance would translate to 6,700 miles over 100,000 miles, so this error could accumulate to be a significant value over time. I'm pretty sure the ODO is geared to show 1.0 mile w/ a 1000 rotations of the speedo cable (1000 is the constant used for this in the speedo gear calculator I used).

hookem horns

$20 and 20 minutes for a driven gear and you will be 98% accurate . Still think its possible you have a % difference across the range and just can't see it indicated on the speedo. You say 40 at GPS 44 but maybe that needle is really idicating 39, same error you have if 60 is GPS 64.

jag1886

You just need one of the little boxes that screw onto the out put of the transmission, "I'm at a loose for it's name right now. Lots of older cars and trucks used these to correct the speedo when there was a gear in the car or truck. I had one installed at the shop with a dyno, they ran it and got my speedo within 1% at 60 MPH.

pcard

I would agree, as long as the odometer is also incorrect, based on the graph I would change my stance and the gearing. But if the odometer is correct now I would leave it alone.

James Colvin

Web site for that is www.cpttransmission.com saw it in a mag. $75

Yellowstatue

And then there is the "Certified Police Speedometer". One for a '68 on eBay for $1500.00

JohnnyBs68S

LOL, "NOS" w/ 33,000 miles on the ODO. I'd still trust the GPS more than that.

I did find that my GPS has a trip mode that records distance traveled, however I think the resolution is 1 mile (not tenths), so I'd still have to travel a decent distance (20 miles or so) to discern an error. If I take a trip, I'll use this to check the ODO. I agree, ODO accuracy is more important than speedo accuracy to me, and if its OK, I'm leaving the speedo gearing alone.

Those are used if you want to avoid changing the speedo gears inside the trans, or if the right combination of drive/driven gears are not available for your tire/rear axle ratio combination. I had to use one of these w/ my Jetaway when I switched from a 2.78 to a 3.90 rear gear, and I didn't want to remove the tailshaft housing to change the driven gear. A combination of a 26.x% external reducer and a 1-tooth change in the driven gear got me to the right combination (although I never checked the speedo accuracy of that setup w/ GPS since I knew I'd eventually be upgrading the tranny). With my new trans, I am able to get to the proper combination of drive/driven gears without using an external reducer (barely, the 10-tooth drive gear is the smallest available for a 200-4R, and the 31 tooth driven gear is the largest........If I had a 4.10 rear gear or smaller tires, I'd have to use an external reducer). As I showed in an earlier post, I can "fix" the ~6.7% error at 60MPH by changing to a 29-tooth driven gear, but that's a 2-tooth jump from what I have and I'd still be off (although less-so) at other speeds.

For sure, "quantization error" of the GPS readout is at LEAST +/- 1/2 MPH, and "eyeballing error" the needle position on the speedo dial is probably not much better. Combine those inherent errors in my measurement technique I could still be off a % or so in either direction. I could mitigate some of this error by making numerous repeated measurements over a wider operating range and averaging. Or I could ignore these possible errors and just aim down the center of the fairway.