Speed Sensor #1

[b3lha]

Here's a question for the gearbox gurus: What does speed sensor #1 actually measure? It must be counting the number of rotations of some shaft over a period of time.

I know that notionally VSS1 reports the rear wheel speed. But it can't count rear axle rotations because it is located in the gearbox, not the rear diff. To report rear wheel speed, the TCU would have to count rotations of the prop shaft and divide this by the diff ratio to work out the rear wheel speed. In this case, the TCU must know the diff ratio.

If this is how it works then VSS1 would read incorrectly on cars with a swapped diff ratio while VSS2 in the front diff would read correctly. So there would be a discrepancy between the front and rear speed sensors. Can anyone confirm this?

Looking at the program in the TCU, it seems to look at what gear you are in and multiply by the gear ratio. Does this mean that VSS1 is on the input side of the gearbox? Or is the program possibly multiplying rpm by gear ratio in order to fake a VSS1 signal in the event of a sensor failure?

Any ideas?

[oab_au]

Originally Posted by b3lha

Here's a question for the gearbox gurus: What does speed sensor #1 actually measure? It must be counting the number of rotations of some shaft over a period of time.

I know that notionally VSS1 reports the rear wheel speed. But it can't count rear axle rotations because it is located in the gearbox, not the rear diff. To report rear wheel speed, the TCU would have to count rotations of the prop shaft and divide this by the diff ratio to work out the rear wheel speed. In this case, the TCU must know the diff ratio.

If this is how it works then VSS1 would read incorrectly on cars with a swapped diff ratio while VSS2 in the front diff would read correctly. So there would be a discrepancy between the front and rear speed sensors. Can anyone confirm this?

Looking at the program in the TCU, it seems to look at what gear you are in and multiply by the gear ratio. Does this mean that VSS1 is on the input side of the gearbox? Or is the program possibly multiplying rpm by gear ratio in order to fake a VSS1 signal in the event of a sensor failure?

Any ideas?

Hi Phil that's the way I see also. I mentioned this before as the source of transfer clutch binding in 4.44 conversions. I have always believed that any rotational difference above 20% applied the transfer clutch, so with the front sensor reading wheel speed and the rear reading wheel speed /diff ratio, a 4.44 would cause binding.

That is in the US/JDM models only, the VTD has both sensors on the gearbox output shafts.

Harvey.

[b3lha]

Thanks Harv,
So VSS1 is counting rotations of the prop shaft? Equivalent to (rpm * gear ratio) when the TC is locked? And then that gets factored with the diff ratio (and wheel size) to determine the speed?
Phil.

[oab_au]

Originally Posted by b3lha

Thanks Harv,
So VSS1 is counting rotations of the prop shaft? Equivalent to (rpm * gear ratio) when the TC is locked? And then that gets factored with the diff ratio (and wheel size) to determine the speed?
Phil.

Yes that would be right. That would make the difference in the way the two types of TCU, VTD and transfer, treat the speed sensor signal.

Harvey.

[b3lha]

Originally Posted by b3lha

I know that notionally VSS1 reports the rear wheel speed. But it can't count rear axle rotations because it is located in the gearbox, not the rear diff. To report rear wheel speed, the TCU would have to count rotations of the prop shaft and divide this by the diff ratio to work out the rear wheel speed. In this case, the TCU must know the diff ratio.

If this is how it works then VSS1 would read incorrectly on cars with a swapped diff ratio while VSS2 in the front diff would read correctly. So there would be a discrepancy between the front and rear speed sensors.

I figured out another piece of the puzzle. The TCU counts pulses/time from VSS1 and multiplies them by a constant to work out the speed in km/h.

On a USDM TCU, the constant is located at address 0xC019.

I *think* the formula for the constant is 470 x 1/ratio.

For a stock 3.545 rear diff it is set to 133
For a 3.700 rear diff it should be 127
For a 3.900 it should be 121
For a 4.111 it should be 114
For a 4.444 it should be 106

Changing this value in the TCU should correct the VSS1 reading after a diff swap and may cure the solenoid C problem that Harvey mentioned.

[svxistentialist]

Originally Posted by oab_au

That is in the US/JDM models only, the VTD has both sensors on the gearbox output shafts.

Harvey.

What you both are deducing here sounds about right to me.

So explain the following for me please.

If I want to change the final drive in my VTD, I am presuming I have to fit a rear 4.444 diff say. To match this in the front I'm presuming I will have to change one of the pinions in the front diff to give the same 4.444 output ratio, so that the front and back axles are matched.

Assuming I am correct so far, and knowing the front and rear sensors are both inside the gearbox, then I would deduce that both sensors are "seeing" equivalent revolutions, and therefore would send matched speed signals to the TCU.

However I would expect the clock to be reading fast by a big percentage in this case. To make the clock [speedo] read correctly I imagine it would be necessary to give the TCU a new constant for multiplying out the speed.

Do you think this would be correct for the VTD?

Joe

[b3lha]

Originally Posted by svxistentialist

If I want to change the final drive in my VTD, I am presuming I have to fit a rear 4.444 diff say. To match this in the front I'm presuming I will have to change one of the pinions in the front diff to give the same 4.444 output ratio, so that the front and back axles are matched.

Yes.

Originally Posted by svxistentialist

Assuming I am correct so far, and knowing the front and rear sensors are both inside the gearbox, then I would deduce that both sensors are "seeing" equivalent revolutions, and therefore would send matched speed signals to the TCU.

Yes. On a Euro/Aussie VTD box with both sensors internal.

Originally Posted by svxistentialist

However I would expect the clock to be reading fast by a big percentage in this case. To make the clock [speedo] read correctly I imagine it would be necessary to give the TCU a new constant for multiplying out the speed.

Probably. My description above relates to sensor 1 on a USDM ACT4 box. But you are asking about sensor 2 on a UK VTD box. I will investigate how it works and get back to you.

[TomsSVX]

VSS1 monitors the rotational speed of the transfer clutch hub. VSS2 monitors rotational speed of the front differential carrier IE front axles.

VSS1 monitors the transmissions output to the rear wheels. If this rotation is slower or faster than VSS2 detects, it engages the AWD. VSS2's main function is to supply the ECU,TCU, and driver of the vehicles actual speed. Only in the case of a VSS2 failure will the VSS1 be used as anything more than a reference on how much drive to delivery to the rear wheels. Now given that the VSS1 monitors pinion revolutions while VSS2 monitors ring revoltions a larger gap in differences will cause the TCU to think it needs to deliver more power to the transfer clutches. BUT there still needs to be throttle applied for this to take place. Without significant throttle voltage, there is still a high duty rate for the Sol. C which prevents binding. In fact, the difference in AWD distribution will likely be greeted with open arms as it will likely improve the car's handling

Tom

[b3lha]

Originally Posted by svxistentialist

However I would expect the clock to be reading fast by a big percentage in this case. To make the clock [speedo] read correctly I imagine it would be necessary to give the TCU a new constant for multiplying out the speed.

I had a look at the dump of your UK TCU and it does have a similar constant for the speed calibration.

The same constant is used for VSS1 and VSS2. It's address is 0xC01A and the formula appears to be (940 x 1/ratio).

For a stock 3.700 it is set to 254.
For a 3.900 you would set it to 241.
For a 4.111 you would set it to 229.
For a 4.444 you would set it to 212.

That would correct the TCU's internal speed readings. I haven't been able to figure out whether it would also correct the speed signal that the TCU sends to the ECU and Speedo. But it wouldn't hurt anything to give it a try.

Note that the maximum value for this constant is limited to 255, so you couldn't run a 3.545 on this TCU without some extra work.

[svxistentialist]

Thanks Phil.

Excellent information. Upgrade path is now very clear.

Joe

[b3lha]

Here's another question for the gearbox gurus:

Does the US Front Wheel Drive transmission have VSS1 inside the transmission, or only VSS2 in the front diff?