Speedometer recalibration procedure ?

[MountainRoads]

The speedo in my trusty -6 crapped out. Big clue that the problem was with the speedo and not the cable or transmission drive unit was the extreme latency before it died. Both coming up to speed and then winding down again. (I was pretty sure I wasn't going 60 mph when I pulled into the garage.) Not really a huge tragedy. PO had changed the final drive ratio and both wheels and tires are non-stock size, so it was about 5-7% optimistic and needed to be recalibrated anyway. So I called the places I know - Palo Alto and North Hollywood. Both have good reps, are easy to deal with, and offer comparable price and services.

North Hollywood kindly e-mailed me instructions to follow to give them the information necessary to accurately recalibrate the speedo. I would post the sheet they sent here and I'm pretty sure there are no real trade secrets involved, but I haven't asked for permission. I will gladly post for future reference if they give me the OK. The procedure essentially consists of marking off a known distance (52 ft, 9.5 inches), disconnecting the speedometer cable at the speedo end, counting the number of cable revolutions required to cover the known distance, and then repeating this procedure two more times to get a good average. This seems reasonable, but has a couple of drawbacks for me: 1) I'm not to keen on finding a suitable location and doing this in inclement weather right now, and 2) it seems like you'd almost need three people to do this. One to pay attention to the covered distance, one to pay attention to the number of resulting speedometer cable revolutions, and one to push/steer the car.

So I'm wondering about a simpler procedure. The covered distance is almost exactly 1/100th of a mile. Likely not a coincidence. I'm guessing they average out the number of cable turns required to cover a mile and then use that for calibration purposes. The tire manufacturer specs the tires (205/60/15) at 847 revs per mile, which works out to 8.47 revs per 1/100th of a mile. Wouldn't it be a lot easier to jack up one side of the car, spin the rear wheel 8.47 times, and count the number of resulting cable revs instead of pushing or driving the car to get this info? The rolling circumference of the tire won't change because it's in the air and leaving the other tire on the ground just means you divide the number of cable revs by two, since the other tire isn't rolling along in tandem and therefore all single wheel spinning motion is transferred to the transmission mainshaft instead of being shared by both wheels. Can anyone provide additional input our point out flaws in my logic, assuming no LSD is involved to interfere? Or, do I have to multiply the number of cable revs by two, instead? Thanks very much in advance.

- MR

[304065]

A fun question.

First let's talk about sources of error. Your sixer came originally with 165 HR 15 tires, same as an elfer. According to the book

Outside Diameter 650mm
Nominal tire width 177mm
Effective static radius 295mm
Effective dynamic radius 314mm.
That works out to 816 revs/mile for a dynamic radius of 314mm. 1/ ((314 x 2* Pi) /25.4 /12 /5280)

Let's pause there for a second. For a tire with a 650mm nominal diameter, you would think it would have an effective static radius of 325mm. In fact, the weight of the car bulges the sidewall so you end up with about 9.25% less radius. When the car is moving, centrifugal force expands it somewhat, so the dynamic radius is about 3.4% less.

So the first thing I would be concerned about is whether or not your revolutions per mile is a laden or an unladen spec. I mean, is that just the nominal diameter divided into a mile? What NH is asking you to measure is certainly 1/100 of a mile, which they will then use to come up with the revolutions per mile. since they are asking you to move the car, I would say it's somewhere closer to the static reading, given that the tire probably doesn't bulge out that much when the car is only moving 52.79 feet.

They will then take this revolutions per mile number and rotate the speedometer shaft at that precise number of revolutions per minute, which should then indicate 60 MPH on the speedometer. If it's off either way they fiddle with the magnet and spring to get it lined up.

Old VW VDO gauges use 825 revs per mile as the spec, I think they used the same 165/15 tires. A tire that generates 825 revs per mile has a dynamic radius of 310mm. That's pretty close to the Porsche spec of 314mm.

Anyway, what needs to happen is you need to measure the revolutions per mile of the tire at a given speed. Why not go out on a level surface, use the dashboard app in your iphone to set the speed to precisely 60 mph and note the speedometer reading, then give that to NH speedo so they can calibrate? I realize that means another shipping round trip but it would give you more accurate readings if you decide to stick with those shorter tires.

Where it really gets interesting is using a VDO pickup in place of the cable to eliminate the mechanical speedometer entirely. . . rally anyone?

[MountainRoads]

Thanks much for the detailed response. PM sent with further detail to avoid boring the others

The idea of going all electric is intriguing, but for now I'm trying to stay as stock as possible.

BTW - to further muddy the waters, I believe 914-6s originally came with 14 inch "comfort" wheels. At least the Fuchs were. Can't speak to the steelies. I'm not sure about the stock tire size.

- MR

[MountainRoads]

Resurrecting this post for documentation purposes. I had a very nice conversation with Kevin at North Hollywood about my *alternate* procedure for proper speedometer calibration. Post by 304065 (John, above) is absolutely correct - they take the number of revolutions per mile and use that as revolutions per minute which should match 60 mph. They also do a check at what should be 30 mph, 90 mph, and I think 120 mph, as well. The big revelation was Kevin told me that unlike many cars which use the transmission mainshaft for the speedometer drive, all Porsches with mechanical speedometers took their reading from the ring gear. Taking the reading from the ring gear is independent of the final drive ratio and means the final drive gear ratio is irrelevant. This simplifies things greatly, because you don't have to worry about what rear end gear ratio is in the car and all that matters is knowing the number of tire revolutions per mile, which most tire manufacturers specify. I believe he told me this was also true for Porsches with electronic speedometers, because the sensors (magnets) for those were also on the ring gear.

I'd be really interested if anybody has heard anything different from the above. My guess is North Hollywood knows what they are talking about, if anybody does.

Thanks,

- MR

[MountainRoads]

Adding to this post in case somebody else runs into the same issue / conundrum. Although North Hollywood's provided procedure undoubtedly works, it is generic (applies to any car), and is problematic unless you are willing to brave the elements this time of year. The procedure I used let me stay in the garage while it was snowing outside and proved Kevin's assertion that 1 "axle" rotatation produced 1 speedometer cable rotation, since the speedometer drive comes from the differential ring gear and not the transmission main shaft.

Raising the rear of the car on jack stands, I immobilized one of the rear wheels. Turning the other wheel two full revolutions in the forward direction produced a single speedometer cable rotation. Repeating the procedure on the other side produced the exact same results. Therefore, if both wheels rolled forward simultaneously, two full "axle" rotations would produce exactly 2 speedometer cable revolutions. Logically, this makes sense, but you never know until you try. Therefore, demonstrating that the rear drive ratio really doesn't matter when recalibrating the speedometer on these cars. The only thing required is to know the tire's number of revolutions per mile rating. Something most tire manufacturers generally now provide.

Thank you Kevin. This tidbit saved me a lot of trouble. If anybody knows differently, please chime in. Thanks.

- MR

[304065]

For a 901 box, the Speedo drive comes off the end of the lay shaft. That is the little brass lollipop below.



This is also the shaft that drives the ring and pinion, so there is a gear reduction in the bevel drive to match the ring and pinion ratio, to get to 1:1 to match your observation, I suppose.

[M_deJong]

Same with 915, speedo drive is on the end of the pinion shaft. So final drive ratio does matter.

Thanks for the 1:1 finding, that is very useful.

[MountainRoads]

Same with 915, speedo drive is on the end of the pinion shaft. So final drive ratio does matter.

Yes, I ran this all past my "go-to" mechanic for a final say-so and he concurs. He said he'd have to check his book to be certain, but he thought there were only two different ring/pinion ratios available in the early trannies and it would be extremely unlikely that anyone would change the ring/pinion gears. 1-5 gear ratios yes, but not the ring/pinion. Anyone changing ring and pinion would also need to change the "lollipop" referenced above to keep the 1:1 axle revolution to speedometer cable revolution ratio. In truth, the speedometer drive *effectively* comes off the ring/pinion. Not literally. Apologies for my misinformation on that part.

- MR