Speedometer off from GPS and ECM

That's funny, I just noticed that just the other day and have been meaning to ask. Seems odd to me that the %error would be on the faster side

Sent from my HTC One using Tapatalk

stock size wheels and tires?

Speedometer is almost always off, seems more or less normal to me. For legal/safety reasons, the speedo should read slightly faster than you are going however, which is what I find abnormal in your case.

Comparing a speedometer and a GPS is like comparing an Uzi with a .50 BMG. At the upper end of the speedo readings it will be as much as 10-15% off. GPS is always more accurate.

Speed-o's shouldn't ever read off (by that I mean be intentionally set to read high from the factory)...that's a lawsuit waiting to happen and mileage fraud.

Also, you guys are missing the point. The signal from my ECM is dead on with GPS. What is DISPLAYED is incorrect. The gauge cluster is being TOLD to show 75 mph but shows 72....

awenthol said:

Speed-o's shouldn't ever read off (by that I mean be intentionally set to read high from the factory)...that's a lawsuit waiting to happen and mileage fraud.

Also, you guys are missing the point. The signal from my ECM is dead on with GPS. What is DISPLAYED is incorrect. The gauge cluster is being TOLD to show 75 mph but shows 72....

Click to expand...

it is very weird that the ECM and gauge cluster are not reading correctly, there is most likely an issue there, you COULD open up the cluster and slightly adust the needle, be very careful though as you could do damage to the electronics behind when moving the needle.

As for the speedo reading off, it is definitely off by a certain percentage as your speed increases (compared to the more accurate GPS reading). Sometimes as high as 4-5% in higher speeds (top speed for example).

awenthol said:

Speed-o's shouldn't ever read off (by that I mean be intentionally set to read high from the factory)...that's a lawsuit waiting to happen and mileage fraud

Click to expand...

Only if it's by a lot. At legal speeds it should be close. Tire wear alone will throw off mileage and speed readings.

Same for me 2mph off tested this am. Cluster reads 70 when ECM (accessport) and GPS read 72. Stock wheels and tires. Again... both ECM and GPS read the same and cluster is off

5k miles on car and tires

My speedo had been off since I bought the car as a third owner at 94k mine reads 70mph at an actual speed 65 mph. I've had threes sets of tires and two sets if wheels. I always have the same %error linear speed error. At residential speeds it is high by a couple of mph. I keep a GPS on the dash in speedo mode at all times. Jeepers! Now I'll have to put this complaint in the Are you happy with your STI thread!

OR . . . Perhaps I should go on the assumption that my car is faster than others and worth more! Yes the butt dyno says its appreciating already!

You do realize that the ECM gets the raw hall effect or magnetic pickup signal and that it's going to lose some accuracy when transferred to the cluster that has an accuracy generally +/-10% of scale. There is an allowable error and 2 mph is well within that. Not sure why a 2 mph difference is a big deal to anyone.

heavyD said:

You do realize that the ECM gets the raw hall effect or magnetic pickup signal and that it's going to lose some accuracy when transferred to the cluster that has an accuracy generally +/-10% of scale. There is an allowable error and 2 mph is well within that. Not sure why a 2 mph difference is a big deal to anyone.

Click to expand...

From the source to the ECM, I can see (but there isn't)...not from the ECM to the cluster--or are you referring to the stepper motor in the cluster. Again, the ECM is reading and transmitting correctly.

My theory is that the dial did not get put on the stepper in the proper orientation.

On 2008+ STIs, the signal comes from the VDC module. The VDC module synthesises a signal from the four ABS sensors and transmits that to the combination meter. The ECU receives it via CAN from the VDC module.

I'm not sure why they don't use CAN to transmit the speedometer info to the combination meter, but maybe they wanted it to be more direct and fail proof in case of a CANbus failure. Whatever the case, my speedometer is spot on to the ECU and GPS values up to the 80MPH area (never bothered with more). I wonder if it's a calibration issue with later model GR/GVs...

Sabrewings said:

"but maybe they wanted it to be more direct and fail proof in case of a CANbus failure."

Click to expand...

good guess

Sabrewings said:

Whatever the case, my speedometer is spot on to the ECU and GPS values up to the 80MPH area (never bothered with more). I wonder if it's a calibration issue with later model GR/GVs...

Click to expand...

No, I'm GD and off by as much as anyone else.

awenthol said:

My theory is that the dial did not get put on the stepper in the proper orientation.

Click to expand...

Well that's always possible but doubtful as the stepper motor itself is not a perfectly accurate instrument. The error is completely normal.

mheyman said:

No, I'm GD and off by as much as anyone else.

Click to expand...

GDs had a known issue where the speedometer read fast. There was actually a class action about it since it affected the odometer and raised their mileage prematurely. GRs/GVs have not suffered from anything like that until now, and seeing as the systems are completely different it's likely no the same issue.

The biggest candidate at the moment is the needle on the stepper motor. I wonder if it points directly at zero (to a high precision, so don't just glance at it) when you're stopped? Is the error a constant value (3MPH) or is proportional to speed?

awenthol said:

Nevermind that theory...how about bad stepper? I winter prep'd the car this weekend, so it's in hibernation. Not going to be able to test other speeds until spring.

Image Link

Click to expand...

there's your first mistake...storing the car for the winter