Bored?

I've recently done an engine swap and in the process rewired just about everything, including putting an NB instrument cluster into my na. In order to drive + calibrate the speedometer I also bought and installed an Apexi RSM.

The speedometer signal from the gearbox (a nissan unit) is successfully driving the Adaptronic ECU and Apexi RSM, and I have the right speed showing there.

The RSM output (pink wire) is connected to pins 2L and 2M on the cluster, and the tacho is wired to 2K. I have double checked that there is continuity on those connections, and that the dash has a good ground.

It is quite perplexing, because I saw working when I had the car in gear on stands before I put the dash back together. Now the speedo reads zero all the time (also the odometer doesn't increment). The tacho and other gauges are working fine.

If anyone can shed some light on the pinout of the dash WRT the speedometer signal that would be extremely helpful - I'm wondering if I had accidentally shorted onto a nearby pin or something in order to make it work previously. Someone must have done this before!

fmowry

iTrader: (8)

A guy sent me this when I was troubleshooting my '01 wiring for my LSX swap. Might help.

Bored?

Thanks for the link and fast response.

That spreadsheet matches the colors of wires in the 2000 wiring diagram for speedo and tacho.

I had seen elsewhere the suggestion that the yellow wire (pin 1M I believe, but I don't have an actual nb wiring loom...) should be connected to the speed sender, but I think the track on the back of the PCB is connected to ground.

The diagram I have shows the VSS connected between 1M and 2M, and some unknown signals on 2L (from PCM and Cruise control).

Did you have any luck wiring it up with the speed signal just coming in on the yellow wire?

Joe Perez

iTrader: (8)

On a '99-'00, the speed sensor comes into the instrument cluster on 1M (yellow) and 2M (orange).



There's no specific documentation which describes this signal, though the symbol used in the part illustration suggests a VR sensor. There's also nothing which clearly states that one side or the other goes to ground, however we can infer from the fact that on vehicles equipped with an A/T, the orange wire also goes to the transmission controller, the orange wire is clearly not grounded and must contain a signal.

Since the RSM's output is single-ended, I think you're on the right track connecting its pink wire to the vehicle's orange wire. I would further connect the yellow wire to the RSM's ground.

If that doesn't work, you might need to install a 1:1 transformer between the RSM and the speedo, to achieve an isolated, differential signal.

Do you have access to a readable copy of the RSM wiring manual? The only one I can find (here) only shows dots where most of the text should be when I view it.

fmowry

iTrader: (8)

I'm pretty sure I used the orange wire only from the VSS from the GM PCM/ECU for my swap. I know I did not use both. I though one was to control speed, and the other was to send to cruise.

Frank

Joe Perez

iTrader: (8)

Could very well be that the speedo will function with a single-ended input. If the speedo input circuit is genuinely differential, then it would want the unused side to be referenced to the same ground as the signal feeding the other side.

If the wiring diagram is accurate, then the sensor is clearly outputting a differential signal, similar to the output of a two-wire VR crank sensor. The fact that half of it is being picked off and also sent to the transmission controller is interesting, but not likely relevant.

I have no idea is the speedo input circuit truly is differential, but the signal going to it appears to be.


The cruise control computer does not directly attach to the transmission's speed sensor. It gets its speed signal from an output of the instrument cluster- specifically, a white/black wire at position 2L, which also runs to the ECU.


I'll try to remember to scope the sensor on my '99 tranny this evening.

Bored?

Interesting ideas Joe, especially WRT the differential signal - I had assumed it would be the same as the nissan sensor I'm using that is just pulled to ground as it rotates, emitting a square wave. Would be very helpful if you can confirm what kind of sensor the nb miata has.

The link to the Apexi documentation you posted is the same one I have been using. The relevant excerpts are attached. As you can see it only shows being connected to the computer, which may be because they're only using it to remove the speed limiter in that case (as opposed to calibrating the speedo). They don't mention VR sensors anywhere in the RSM documentation.

It makes sense that 2L might be a speed output to the ECU and that the conversion from VR to square wave is happening inside the instrument cluster. I hadn't considered that.

I wonder if my driving 2L and 2M together and seeing the speedo move was good luck or bad... hopefully nothing has been damaged or debugging this will be even harder.

Joe Perez

iTrader: (8)

That's how the speed sensor on the NA works- there's a magnetic reed switch inside the speedometer itself, which is actuated by the rotating disc that's driven by the cable.


We'll have an answer later today.



That would make sense. I've confirmed from the factory documentation that the pulse on 2L is a 5v square wave. The documentation does not tell me the absolute direction of that signal, however it logically makes sense that it would be generated by the speedometer and received by the ECU and the cruise controller.

So 2L would be a perfect line to intercept if you want to eliminate a top-speed governor, as that's the line that'd be doing it.

I'm willing to bet money, however, that the speedo input (1M / 2M) is expecting a differential sine-wave signal from the sensor.

Joe Perez

iTrader: (8)

Well, here we are:



That's definitely a VR sensor. Probably built like a stepper motor internally (judging by the notchiness of it) but it's generating an AC voltage when it turns.

In this capture, I'm turning it just by running the gear up and down my arm, so it's not necessarily representative of the speed it'd normally be turning at, however the waveform its producing is going to look like this pretty much no matter what.

p-p amplitude here is about 5 volts, which is perfect for our purposes. It varies (as with all VR sensors) with speed, but that's probably not important. Get the 1:1 transformer I posted above. Wire the primary between your pink wire and ground, and wire the secondary to 1M and 2M at the instrument cluster.

Bored?

Thanks so much for the help, Joe, it has been invaluable.
I can't get exactly the transformer you posted, but I can get hold of this one:

600/600 Ohm Isolation Transformer - Jaycar Electronics

Not quite as convenient since it doesn't have the flying leads, but I think it will do the job.

Joe Perez

iTrader: (8)

Well, that's quite a costly little piece of kit, but it ought to work. The important characteristics for our purposes are that the windings ratio be 1:1 (or thereabouts), and the impedance high enough that it's not going to drag down the signal on the pink wire.

Bored?

Here is a bit of an update:
I spent some time on the weekend messing about with it. The Apexi RSM has stopped working reading the speed sensor in the gearbox reliably - I'm not sure what the problem was exactly. The Adaptronic has a pull-up to 3V on its speedo output, while the RSM appears to have a pull-up to 5V (measures 5V on the speed input), so the first thing I did was disconnect the RSM.

Still couldn't get it to work reliably, though it would measure speed when I manually pulsed the input to ground. Getting desperate I also tried connecting an external pull-up to 12V, and also to ~6v using a voltage divider. Using the latter kinda worked, but not until the speed was up quite high. I don't know exactly what to make of these results - perhaps the internal pull-up of the RSM isn't working correctly, or perhaps the switch output isn't clean enough - I suspect it is designed to be driven from the clean square wave output of an ECU or speedo.

Anyway, I gave up after quite a lot of trial and error on the weekend, but did find this device on the interwebs:
http://autospeed.com.au/cms/title_Sp...5/article.html

I have actually made an earlier version of this kit for a previous car, but didn't pay much attention to the details at the time. Turns out it supports a bunch of different input and output configurations, including being able to generate a 5V pp AC output, just like the NB's VR sensor output you posted above.

I picked the kit up today, so hopefully all will go smoothly this time!

Bored?

I built and installed that kit, and now I have a working speedo.
In my particular set-up I have it set to 99, and that gets me within about 10% of my iPhone's GPS indicated speed.

Hopefully the info in this thread will make it easier for the next person who finds themselves needing to drive an NB speedo from an odd source.

Stein

iTrader: (46)

Good info, thanks for the update. I may need this to drive the NB speedo on the 5.0 swap car.

aidandj

iTrader: (5)

<p>Necro bump.</p><p>Were putting an NB cluster in yank's maita because he can't fit a speed sensor in his transmission (exhaust)</p><p>ABS wheel gives us VR signal which then is conditioned and set to the MS.</p><p>Now we need an output from the MS to drive the NB speedo. Standard output would give us a square wave, we need to get this into a VR signal right?</p>

aidandj

iTrader: (5)

<p>Orrr, will the nb speedo take a square wave just fine? Maybe I need to borrow a function generator and try it out.</p>

aidandj

iTrader: (5)

<p>So cool fact.</p><p>NB speedo can be run with a square wave.</p><p>Video coming soon.</p>

aidandj

iTrader: (5)

<p> </p>