Page 1 of 1

Need help diagnosing speed sensor (DPI1)

Posted: Fri Oct 12, 2018 3:25 am
by infernosg
About a week ago my speedometer started reading slow. I'm using a Hall Effect sensor connected to the stock mechanical takeoff on my transmission as the signal for the vehicle speed source. The sensor is fed switched 12V and ground from the ECU and outputs a 0-12V square wave to DPI1. My first thought was I killed the sensor but before I drop $100 on a replacement I wanted to make sure. However, in my testing I've found everything appears to be working:

1. All the mechanical connections in the transmission are intact. The OEM speedometer drive and driven gears turn with the driveshaft and the Hall Effect sensor input shaft spins with them with no slipping.

2. I tested the Hall Effect sensor with a multimeter and confirmed there are 16 pulses/rev and approximately 12V out during each pulse. The only thing I noticed is there is a very brief moment of discontinuity going from minimum to maximum voltage and back during each pulse.

3. I put the Hall Effect sensor input shaft in a drill chuck and spun it at approximately 400 and 1600 RPM. DPI1 is calibrated for 16,000 pulses/mile so 400 RPM should correspond to 24 mi/hr and 1600 RPM to 96 mi/hr. In both cases my speedometer was showing within 1-2 mi/hr of expected. I confirmed this is the case with both the engine off and running.

Based on 1-3 everything should be working, but when I put everything together it doesn't or if it does, not for long. Here are some more observations:

4. The indicated speed and its variation seem to be dependent on engine speed. Cruising in 5th at 2500 RPM should indicate about 50 mi/hr but instead my speedometer shows around 30 mi/hr and it varies by as much as 10 mi/hr. If I downshift to 3rd and engine speed increases to 4000 RPM the indicated speed will be 40 mi/hr and vary by 5 mi/hr. This trend increases until error and variation are nearly zero at redline.

5. I was able to get the speedometer working briefly after "resetting" the ECU. I reloaded the latest firmware, uploaded a base VE map, then imported my latest map. The car sat for 4 days and during my next drive the speedometer was working for about 15 minutes. After 4 WOT pulls to redline in 2nd gear it started reading slow again and has ever since.

6. The speedometer has been working consistently for over a year but I've only recently started doing pulls to redline and bouncing off the rev limiter. I'd bounce off the rev limiter for about a 1/2 second at most. My rev limiter is a hard ignition cut at 9000 RPM.

7. While testing the connections to the Hall Effect sensor the output was changing from 0-12V to 0-6V when it was connected to the ECU. For some reason I'm getting 5V on the ECU side of DPI1. I didn't make my harness and I'm trying to confirm with its builder if there's some kind of pull up resistor or something on that circuit. I did find a couple other posts mentioning something similar but in both cases their DPI issue "resolved itself."

8. Today I was doing some WOT pulls in 3rd gear to around 7000 RPM. In the log I saw a consistent engine speed at which the speedometer would "catch" up to the actual vehicle speed. At around 4500 RPM in each of the 5 pulls the slope of the indicated speed would suddenly change. Nothing else I logged displayed a similar behavior.

At this point I've exhausted all my ideas. I'm about to bite the bullet and send the Hall Effect sensor out for a replacement for $100. I'm less concerned about the extra expense than I am that something is messed up on the car be it mechanical or electrical. Is there something else I can test that I'm missing?

Re: Need help diagnosing speed sensor (DPI1)

Posted: Fri Oct 12, 2018 10:21 am
by JimKden
On the Diagnostic page, is the DPI reading as it should? That will tell if the hardware and wiring is working ok.

Re: Need help diagnosing speed sensor (DPI1)

Posted: Sat Oct 13, 2018 5:33 am
by infernosg
I turned on the diagnostic page and took a quick log as I drove around town looking for fuel. I think the attached screenshot shows best what I'm working with. It shows the variation in speed even though engine speed is nearly constant. It also shows how the vehicle speed varies with engine speed at higher RPM. The indicated DPI frequency matches the indicated speed (frequency x 3600 sec/hr / 16,000 pulses/mi). I don't understand what is going on with DPI period. Period is supposed to be 1/frequency. In some cases that's true but in other cases the period is alternating between positive and negative (?!) time.

At this point I think either the Hall effect sensor is bad or I'm getting some kind of interference. It seems easy to conclude the sensor is bad. It's in a pretty inhospitable area subjected to high vibration and temperature for over a year now. However, that doesn't explain why the indicated speed varies with engine speed or why it worked briefly when I reset the ECU. Then there's the fact this all only started once after I started doing WOT pulls and hitting the rev limiter.