Anybody know how the vehicle speed sensor works?

[corrado33 10]

So as most of us probably know, the rado came with a few different types of instrument clusters. Some of them had a max speed of 140 mph... some of them (likely from the UK) had a max speed of 160 mph. While I have no intention of switching the cluster in my car, my brother who has built a VRT rado with 700 HP wants to use the 160 mph version (for obvious reasons), but obviously it's not plug and play with the normal 140 mph cluster cars. (At least I don't think so, even if it is, I still want an answer to my question.)

 

So being the electronics nerd I am, I've been tasked with building a little converter to make the 160 mph version accurate in a 140 mph cluster car... with no extra steps (other than plugging/soldering my adapter box in.)

 

At first I thought, "Meh, it's probably just varying the voltage which translates to a certain mph using a map in a ROM somewhere." Yeah, I was wrong. I stuck a voltmeter on the speed sensor wire on the back of the speedo (white wire according to the bentley) and connected the black voltmeter wire to the the grounding block behind and to the left of the cluster, and while the voltage DID change... it was very erratic. It seemingly went to 2.6-2.8 V while you were moving... and with no real pattern. It didn't increase when I went faster, and it didn't decrease when I went slower. Even more interestingly, the white wire would either go to 0V OR it would go to 5.3V when stopped. And it would do so interchangeably. So if I stopped at a stop sign, it'd drop to 0V, then I'd start moving again, then stop at the next stop sign, and it'd go to 5.3V, next stop sign, 0V, next stop sign, 5.3V.

 

So... obviously this isn't voltage. It can't be current or resistance either as that would still have shown to be relatively normal behavior in relation with Ohm's law. Although it COULD still be these just with the tiniest of ranges, but I doubt it. I suppose it could also act like a PWM signal and change the duty cycle, but that'd show up in a cursory voltage measurement.

 

Anyway, so if it isn't voltage, current, or resistance... then I suppose it could be either one of two things.... frequency.... which I find likely considering the look of the speed sensor itself, or actual digital data communication. I find the latter unlikely as it's too complicated for cars this old, although the 0V or 5V behavior has me worried that it may be.

 

So if anybody actually knows what kind of data the speed sensor outputs, then I'd be happy to learn. If anybody has a speed sensor handy (aka not installed in the car) and wants to attach a multimeter to the terminals while spinning the sensor with a drill or something, I'd be very grateful.

 

Also, while the manual transmission kind seems to be... well... physically activated, the auto trans types seem to be a hall sensor (since they're completely sealed and don't physically interact with the transmission at all.) I would imagine these would HAVE to work by outputting a pulse for very pass (and therefore a variable frequency for the speed.)

 

Anyway, any help is appreciated.

 

 

EDIT: On second thought, it's almost definitely frequency. Since when moving the voltage is ~2.68V and the maxes are either 0V or 5.3V, it's almost definitely just frequency and I just got lucky with it stopping on 0V and 5.3V interchangeably. With it just being frequency the voltage when moving will ALWAYS stay at 2.65V or so since the on period and off period are likely roughly equal and both are reduced when frequency goes up.

 

Oh and for how I plan to adapt it? Well, while I am an electronics nerd, I'm not an electrical engineer. I'd probably just program a little atmel microcontroller to input the frequency or whatever from the sensor, then output the "corrected" signal on another pin according to a "table" stored in memory. If I'm lucky.... the speedo will work on a linear relationship with the speed sensor, so I could just use some math and an equation. But if it's not linear, then I'd probably have to use a table as performing complex mathematical equations on the microcontroller every 100th or 1000th of a second isn't the easiest thing to do. Regardless, I'd also make it have a few calibration pots, and maybe even make it fancy and give it a little USB interface so it could be programmed from a computer with putty or something. If I were to get fancy with it, then it could become universal and not just for corrados... which is attractive. Oh and if I made it adjustable like this then we could print out own speedo backing plates and just use this to make the needle accurate to it.

Edited December 13, 2019 by corrado33

[fendervg 32]

It just sends pulses (i.e. frequency) corresponding to the rotation in the gearbox internals as far as I know. The clocks are then calibrated to know that so many pulses - x mph.

 

The same signal is used to control the rear spoiler via the spoiler control model and to provide a feed to the radio if it supports what's known as GALA, speed sensitive volume control. There is a little junction box under the dash where this signal is split out into three wires. If the one for the radio gets grounded or connected by accident it can interfere with the spoiler.

[Dox 23]

If the VRT is FWD why not use a signal from a rear wheel ABS sensor to get true road speed rather than wheel spinning gearbox speeds?

[fendervg 32]

Assuming your brother will still be using an 02A gearbox with the speed sensor fitted, I don't see any reason why the signal would not work with the 160mph clock - the pulses would just be more frequent for the higher speeds, and the needle would reflect that accordingly.

 

A bigger issue is that he probably won't be running 15" wheels and stock tyres, so unless they are sized correctly that might mean that the speedo won't be accurate any more...

[corrado33 10]

Assuming your brother will still be using an 02A gearbox with the speed sensor fitted, I don't see any reason why the signal would not work with the 160mph clock - the pulses would just be more frequent for the higher speeds, and the needle would reflect that accordingly.

 

A bigger issue is that he probably won't be running 15" wheels and stock tyres, so unless they are sized correctly that might mean that the speedo won't be accurate any more...

 

You are correct. He said that the speedo was only 4-5 mph off. He REALLY wanted me to take a look at the tach, cause apparently that's off as well? But the tach signal comes from the ECU so if I had to guess, I'd say it's likely just a varied voltage as that's much easier to generate than a specific frequency.

[corrado33 10]

If the VRT is FWD why not use a signal from a rear wheel ABS sensor to get true road speed rather than wheel spinning gearbox speeds?

 

That.... is a good idea. ha I'll ask him.

[dr_mat 0]

I'd say "because then you're messing with the ABS sensors" and you're more likely to have brake system problems.

It doesn't matter which feed you use, you still need to calibrate it. The one on top of the gearbox is on the output shaft so it reflects road speed just as well as an ABS wheel sensor would.

[Dox 23]

I'd say "because then you're messing with the ABS sensors" and you're more likely to have brake system problems.

It doesn't matter which feed you use, you still need to calibrate it. The one on top of the gearbox is on the output shaft so it reflects road speed just as well as an ABS wheel sensor would.

 

I'd assumed at 700 odd horses he'd got some rather impressive brakes and the ABS would be long gone? The gearbox sensor would read wheel speed which with 700BHP wouldn't necessarily be road speed

Edited December 20, 2019 by Dox

[dr_mat 0]

Dunno, but the wheel speed sensor still depends on wheel and tyre size so you need to calibrate for that. Gearbox feed has the exact same problem, you just need a different calibration because it also goes through final drive and the diff. Personally I'd keep the ABS working with 700 bhp in a Corrado, given a choice!

[dr_mat 0]

So as most of us probably know, the rado came with a few different types of instrument clusters. Some of them had a max speed of 140 mph... some of them (likely from the UK) had a max speed of 160 mph. While I have no intention of switching the cluster in my car, my brother who has built a VRT rado with 700 HP wants to use the 160 mph version (for obvious reasons), but obviously it's not plug and play with the normal 140 mph cluster cars. (At least I don't think so, even if it is, I still want an answer to my question.)

 

EDIT: On second thought, it's almost definitely frequency. Since when moving the voltage is ~2.68V and the maxes are either 0V or 5.3V, it's almost definitely just frequency and I just got lucky with it stopping on 0V and 5.3V interchangeably. With it just being frequency the voltage when moving will ALWAYS stay at 2.65V or so since the on period and off period are likely roughly equal and both are reduced when frequency goes up.

 

Oh and for how I plan to adapt it? Well, while I am an electronics nerd, I'm not an electrical engineer. I'd probably just program a little atmel microcontroller to input the frequency or whatever from the sensor, then output the "corrected" signal on another pin according to a "table" stored in memory. If I'm lucky.... the speedo will work on a linear relationship with the speed sensor, so I could just use some math and an equation. But if it's not linear, then I'd probably have to use a table as performing complex mathematical equations on the microcontroller every 100th or 1000th of a second isn't the easiest thing to do. Regardless, I'd also make it have a few calibration pots, and maybe even make it fancy and give it a little USB interface so it could be programmed from a computer with putty or something. If I were to get fancy with it, then it could become universal and not just for corrados... which is attractive. Oh and if I made it adjustable like this then we could print out own speedo backing plates and just use this to make the needle accurate to it.

 

It would be linear, so IF the two speedos are differently calibrated (I can't see why they would be), you would be able to linearly scale the pulses from input to output side.

 

However, I would only expect the two speedo binnacles to require different calibration if they were intended for cars with different final drive, but AFAIK that's not the case, AFAIK all Corrado VR6/SLC had the same final drive ratio, and the same wheel sizes, so the speed signal should be the same, and the dash calibration should be the same.

 

And 4-5mph out of actually pretty close.. factory calibrated gear is rarely any closer than that, they always over-read by a comfortable 5-8%.