sensors for speedo in trans

[TMoreken]

If I have my 915 trans out of the car, how difficult is it to locate the sensors for the speedometer and can they be replaced? My speedometer has been checked for accuracy and is spot-on to 60MPH. After that, its seems to slow as the car continues to accelerate. It's been suggested some of the magnets have fallen off inside the trans. Any thoughts?

[Oh Haha]

The sensor,(only 1) is located on the passenger's side of the trans, about halfway up the body.
It is a disc that mounts to the trans with one small bolt and bracket.

Have you taken your speedo out and had it checked OUT of the car? Mine was doing some crazy things when I first bought my car and I had it rebuilt and checked.

How slow is it after 60MPH? Is it just the car not going any faster? Not being a smarta$$ but just trying to help.

[TMoreken]

The "sensor" you are talking about works by picking-up a signal from a disc spinning inside the trans. As the disk spins, a certain amount of "magnets" spin giving a "signal" per revolution. My sensor is new. The speedometer was checked for accuracy while the odo gear was being replaced.

In theory, if one or more of the magnets "fell off", the sensor would only get 6 pulses-per-revolution instead of 8 or 10 etc. The faster it spun (higher MPH), the fewer pulses (per given speed). Again, this is a theory.

Anyone else have experience with this? My tachometer will continue to climb in 5th gear, but w/o a correct corresponding increase in MPH. For example, My speedo shoes roughly 100 MPH @ 5000 RPM which I know is way off. At 60, it's around 2700. The math just doesn't compute.

[Oh Haha]

I see now. At least you can rule out the speedo.

I'm going to have to let others take this one as I am unsure, short of disassembly, you would check for the magnets in the trans.
Sorry I can't help more.

[ChrisBennet]

Here's a picture of the speedo wheel with the magnets on it from a 930 (same as a your's I think).


If a magnet was missing, I would think you would register a slower speed all the time. In other words, if the speedo is looking for 8 pulses per revoltion and it "sees" 7 wouldn't it register 7/8 the speed? Just a theory.

The "sensor" is just a reed switch so the signal you see would be a square wave. I bet you could see a missing magnet on an oscilloscope. If you want to throw a scope on it, the speedo sensor has a connector in the tunnel under the cover just in front of the rear seats. I think one of the 2 wires is green and I'm not sure what the other color is (brown?).


-Chris

[ischmitz]

Your speedo may be spot on at 60mph but not linear. Have it checked by a spedometer shop such as North Holywood. Or simply hook up a function generator and measure the frequency of the input and the indicated speed. There should be 6480 pulses per mile. So to generate 60mph signal you need 6480/60/60=108Hz. For 120mph it is 216Hz

As others said, a missing magnet would cause a scale error: The indicated speed would be lower by a factor given by the number of missing magnets and affect all speeds. New speed = x/8 * real speed

The only other explanation I could see is that the reed sensor on the outside of the tranny is somehow failing in a weired way and is not able to follow the higher frequency of passing magnets. However, that is a far-fetched idea.

Ingo

[TMoreken]

Thank you to everyone for your responses. Where is the magnet wheel in the trans? I will try to test it w/o removing it. Answers to follow.

[ischmitz]

The wheel is on the differential opposite to the diff cover plate. It becomes obvious, once you locate the reed speed sensor clamped to the outside of the transmission "listening" to it. However, getting to it would require to drain the oil and take the entire differential out. I don't think this is waranted - there are easier methogds to test this with an an oscillioscope.

[Porsche_monkey]

Put an ohmeter on the two wires and see how many times the reed switch closes in one revolution of the transmission output.

[TMoreken]

Great! I'd like to eliminate the obvious before tearing my diff apart. I'm assuming on a 915 gearbox, I should read continuity 8 times in one revoution. I wouldn't mind getting my speedometer recalibrated. I'd just like to know if the input signal to the reed sensor is correct.

Thanks guys.

[scarceller]

Ingo,

I am trying to figure out how many speedo pulses per mile in the 915 Trans and saw your post.

You mention 6,480 per mile, how did you figure this out? I ask because I'm trying to calculate this for my WideBandO2 control system to log Speed MPH.

My calculatuions are as follows:

First, I figured the circumference of my rear wheel like this:
Diameter of wheel measured with tape measure is 25"
PI * 25 = 78.54" circumference.

So one turn of the rear wheel should move the car forward about 78.54"

Next I calculated how many rear wheel turns required to cover 1 mile which is also 63,360" per mile

63,360 / 78.54 = 807 wheel turns to a mile

Now this is where I'm not 100% certain but here goes:
The final drive ratio for the 915 is 3.875:1 so I want to calculate how many final drive turns per mile like this:

807 wheel turns * 3.875 = 3,127 final drive turns per mile.

Then if the final drive has 8 magnets it will generate 8 pulses per final drive turn.
3,127 * 8 = 25,016 pulses per mile

This calculation is strictly based on math and I could have something wrong.

Any input would be helpfull.

Originally Posted by ischmitz

Your speedo may be spot on at 60mph but not linear. Have it checked by a spedometer shop such as North Holywood. Or simply hook up a function generator and measure the frequency of the input and the indicated speed. There should be 6480 pulses per mile. So to generate 60mph signal you need 6480/60/60=108Hz. For 120mph it is 216Hz

As others said, a missing magnet would cause a scale error: The indicated speed would be lower by a factor given by the number of missing magnets and affect all speeds. New speed = x/8 * real speed

The only other explanation I could see is that the reed sensor on the outside of the tranny is somehow failing in a weired way and is not able to follow the higher frequency of passing magnets. However, that is a far-fetched idea.

Ingo

[ChrisBennet]

Originally Posted by scarceller

Ingo,

I am trying to figure out how many speedo pulses per mile in the 915 Trans and saw your post.

You mention 6,480 per mile, how did you figure this out? I ask because I'm trying to calculate this for my WideBandO2 control system to log Speed MPH.

My calculatuions are as follows:

First, I figured the circumference of my rear wheel like this:
Diameter of wheel measured with tape measure is 25"
PI * 25 = 78.54" circumference.

So one turn of the rear wheel should move the car forward about 78.54"

Next I calculated how many rear wheel turns required to cover 1 mile which is also 63,360" per mile

63,360 / 78.54 = 807 wheel turns to a mile

Now this is where I'm not 100% certain but here goes:
The final drive ratio for the 915 is 3.875:1 so I want to calculate how many final drive turns per mile like this:

807 wheel turns * 3.875 = 3,127 final drive turns per mile.

Then if the final drive has 8 magnets it will generate 8 pulses per final drive turn.
3,127 * 8 = 25,016 pulses per mile

This calculation is strictly based on math and I could have something wrong.

Any input would be helpfull.

I'm too busy to check your math but as for as the rolling circumference of you tire goes, I would measure it with a tape.
Make a chaulk mark on the pavement and on the tire and roll the car forward until the tire chalk mark is in the same position, then mark the street. Then measure the distance between the chaulk marks.
-Chris

[Matt Monson]

You say you have the trans out of the car? Just pull side cover. Skip the math. Skip the guess work. On a 915 there are no paper shims that adjust any of your ring and pinion settings so all you will need to do is order a replacement seal for the side cover from our host. You won't mess anything up by pulling out the differential.

Also check the drain plug to see if it's got any of the magnets stuck to it.

I would hate to think what one of those magnets would do if it ran between the gears on your ring and pinion. If you think you have an issue here, fix it now before it turns into a complete tranny rebuild...

[scarceller]

Originally Posted by ChrisBennet

I'm too busy to check your math but as for as the rolling circumference of you tire goes, I would measure it with a tape.
Make a chaulk mark on the pavement and on the tire and roll the car forward until the tire chalk mark is in the same position, then mark the street. Then measure the distance between the chaulk marks.
-Chris

Chris, great idea I'll do this next to double check my math.

[scarceller]

Originally Posted by ChrisBennet

I'm too busy to check your math but as for as the rolling circumference of you tire goes, I would measure it with a tape.
Make a chaulk mark on the pavement and on the tire and roll the car forward until the tire chalk mark is in the same position, then mark the street. Then measure the distance between the chaulk marks.
-Chris

Chris, I just marked the road with tape and the math is spot on at 78" thanks for the tip to double check the math.

[scarceller]

Originally Posted by ischmitz

Your speedo may be spot on at 60mph but not linear. Have it checked by a spedometer shop such as North Holywood. Or simply hook up a function generator and measure the frequency of the input and the indicated speed. There should be 6480 pulses per mile. So to generate 60mph signal you need 6480/60/60=108Hz. For 120mph it is 216Hz

As others said, a missing magnet would cause a scale error: The indicated speed would be lower by a factor given by the number of missing magnets and affect all speeds. New speed = x/8 * real speed

The only other explanation I could see is that the reed sensor on the outside of the tranny is somehow failing in a weired way and is not able to follow the higher frequency of passing magnets. However, that is a far-fetched idea.

Ingo

UPDATE: Ingo, your 6456 number is correct, I did not realize that the wheel with the 8 magnets that signals the sensor spins at rear wheel speed. I had assumed it spin at final drive shaft input speed wich is 3.875 times faster than wheel speed.

So the updated calculation show this:

63,360" per mile / 78.54" tire circumference = 807 tire turns per mile

Then:

807 * 8 magnets = 6,456 pulses per mile.

Then you can also express this:

100MPH will turn wheels 645,600 per hour
or
100MPH will turn wheels 10,760 per minute
100MPH will turn wheels 180 per second

So, 180hz signal will produce 100MPH on the speedo.

And you stated 6,480/mile close enough in my book.

Mystery solved.

[ischmitz]

Scal, glad you figured it out. I was traveling for two weeks and haven't had a chance to read the board.

Now, will you test the speedo to check if there is something wrong with it?
Did you check that you get 8 pulses per wheel rotation on your tranny?

Cheers,
Ingo

[scarceller]

Originally Posted by ischmitz

Scal, glad you figured it out. I was traveling for two weeks and haven't had a chance to read the board.

Now, will you test the speedo to check if there is something wrong with it?
Did you check that you get 8 pulses per wheel rotation on your tranny?

Cheers,
Ingo

Ingo,

My speedo is working just great and I do have the 8 pulses per wheel rotation.

The reason I'm looking at this is that I'm going to record speed in MPH via my WideBandO2 controller. Right now I have everything ready except for one small issue:

My controller needs 0-5 volt square wave (TTL) and the Carrera speedo has a nice output terminal 'A' that has a very clean square wave like signal but it's only 1Volt peak to peak (-0.5 to +0.5 volt signal). Now I'm working on a simple Op-Amp circuit that I will simply stick (glue) to the back of the speedo, the circuit will be powered right from the +12V at the speedo as well as the ground, then the 'A' terminal goes into the circuit - magic happens in the circuit - and 0-5V square wave comes out of the circuit and is then feed to my WBO2 controller.

I will continue posting details to my Wideband O2 post:
Wide Band O2 + data logger for 84-89 Carrera

Here is what I'm recording:
AirFuelRatio
RPM
AirFlowSensor voltage (0-5v)
WOT switch position (open or closed)
Speed in MPH (working on this now)

But thanks a bunch to this thread for the details needed to help figure out the number of speedo pulses per mile.