darksky

Some of you might remember that my 2008 TL has an error in its Speedo (it reads too slow) and that I was trying to figure out if the Speedo and Trip Computer get their speed data from the same sensor, or from two unique sensors. This would affect MPG calculations, average speed calculations, ABS, etc. etc. Additionally, if the error was present in the odometer, that would affect the true miles racked up on my car. Yeah, I can be a little OCD about things. Read these two threads for more background if you are interested:

Speedo and odometer ... where do they get their data?
needle Speedo is off by 4 MPH... can it be adjusted?

Anyway, we went on a 4 h road trip (300+ miles each way) last week and I decided to use the opportunity to answer these questions. In broad strokes, I believe that the Speedo and the Trip Computer get their speed data from different sensors. The odometer gets its data from a third unique sensor. I base these conclusions on the different errors I measured in each of these systems. Read on for the details.

Speedometer
Method 1 – Compare the cruise control to a calculated average
I set my cruise control and used highway mile markers as my measure of distance and a stopwatch for time, and then calculated my average speed.

If speed is indeed constant (with the cruise control it should be), the average speed should be the same as the cruise set point. I timed each run with the stopwatch at different mile markers in the trip. Simple visualization of these data in your favorite spreadsheet app via a graph of distance vs. time will result in a line whose slope is the average speed.

I did this @ four different speeds (60, 65, 71, and 76 MPH) and found that after driving at least 10 miles and collecting at least 5 data points per speed, the average error in the Speedo for these four speeds was 2.6 %.

Example with the cruise set to 76 MPH:



Method 2 – Compare the cruise control to a GPS
Just as a sanity check, I set the cruise @ 70 MPH and compared the speed displayed on a Magellan 4000-something GPS that was tracking nine satellites at the time. After traveling 6 or 7 miles, the average speed reported on the Megellan bounced between 68-69 MPH. This method gave an error in the Speedo of about 3 %.

Trip Computer
Again basing the distance traveled on highway mile markers, I reset the Trip Computer and started a stopwatch at the same point. At the end of 217 miles, the Trip Computer reported an average speed of 68 MPH. The calculated average speed based on the 3 hours, 1 minute, and 21 seconds it took was 71.93 MPH. This comes out to an error of 4.2 %.

If you do this same exercise with less distance, the measured error is lower and maxes out at 4.2 %. I have data for 17, 23, 51, 101, and 217 miles. The errors are 3.4 %, 3.6 %, 3.8 %, 4.2 % and 4.2 % respectively. In other words, it took at least 100 miles for the error to stabilize indicating that you can’t do this experiment in a shorter trip.

Odometer
We did 220 miles on the same highway using consecutively numbered mile markers as the measure of distance. I reset the trip odometer to coincide with the passing of a mile marker at the beginning of the trip. About 3 hours and 220 mile markers later, the odometer read 220.1. I checked it along the way too and found that it was dead-on until about 100 miles (it read 100.1) and interestingly, the error didn't double at 200 miles as I would have expected.

This is telling me that the mile markers might not be perfectly calibrated and that probably a 0.1-0.2 mile error after 200 miles would be acceptable and within the error of the mile markers themselves. My assumption was confirmed on the return trip, when we traveled 235 miles on the same highway. This time, the trip odometer was dead-on hitting 235.0 within about 10 yards of the mile marker.

Therefore, my conclusion is that my odometer is very accurate. This isn’t so unexpected given that class action lawsuit against Honda/Acura a few years ago for this very reason!

Conclusion
My Speedo displays an error of 2.6 %. If you give the Trip Computer enough data, its average speed gives a slightly higher error of 4.2 %, which is more than 1-1/2 times larger than the Speedo error. To me, these data are suggesting that there are indeed two unique sensor inputs (each with their own unique errors) for speed. The odometer gets its data from an independent sensor since it doesn’t display any error that I can measure.

csmeance

In my opinion, I believe that the odometer and gauge get the information from the same signal, but the speedometer displays something different. Alot of member has reported going 152, and in the videos that are shot, the speedometer shows a difference of 8 MPH. This would mean at 80MPH, you are going 76MPH, at 40 you are going 38 and so on.

darksky

@csmeance - If this has taught me anything, it's that you need to have a reasonable data set to make the measurement meaningful. In other words, hitting 152 MPH for 2 or 3 seconds isn't gonna cut it. You would have to maintain that speed for 20 or 30 miles. Aside from that, most of us tend to drive about 1/2 that fast day-to-day

I'd argue you want to sample speeds that reflect your average use of the vehicle. Also, although the error may have a linear range, there is no evidence to say that it's linear up to 152 MPH. It may very well stay linear to say 110 MPH, but then increase in a non-linear fashion.

Anyway, in the absence of any experimental data, or technical documentation from Acura, I still believe we're seeing the results of independent sensors, and I'll challenge folks to provide proof otherwise.

csmeance

One member here does 160 on a daily basis and has shot videos and film of it as he drives on the autobahn. Search for him. I believe his speedometer showed around 155 when his MID showed 147.

zeezz

good detective work.

darksky

Dude, I doubt the TL can hit 160 period. Beyond that, like I showed, you need a pretty large dataset (100 miles) for the Trip Computer to approach a stable error. To do the speedo/milemarker thing, you'd need a good 10 miles of constant travel speed. So maybe this person can set his cruise for 160 and do the experiment.

Again, most of us here in the 'States drive about 1/2 that speed so I think folks here would be interested in the error up to about 80 MPH

6spMike

Firstly - I'm not trying to bash - great work on taking the data.
Was your wife/GF along? My wife would have been really annoyed, but that's a different story.
Just because the errors are different does not mean the sensors are different.
Why would acura use 3 sensors when they could get by with one?
...The wear and inflation of your tires also has an effect.
I would guess that your tires are about 1/2 worn, since the odo was dead on.
I would also guess that the speedo is calibrated to read a bit low on purpose, possibly so with brand new tires it would not show you going slower than you actually are.
Maybe there is some sort of bug in the trip computer algorithm. Did you happen to notice if the trip computer duration timer was accurate?

darksky

@6spmike - Yeah, my wife was with me and yes, she bitched about it several times! The car is a 2008 and only has 2,600 miles on it. The tires are brand new as well.

As to your question about using redundant sensors, I have no answer. I called Acura's customer care line and they could/would not give me a straight answer. I even asked to speak to a tech and again, I was told this wasn't possible. It's not like this is some sort of trade secret and I'm sure this is documented in some manual for the car. I also called a few dealerships asking the same question but found that they too didn't know. One of their "techs" mentioned he thought it used a "multiplex system" of several different speed inputs which would be consistent with my observations, but again, nothing official unfortunately.

The speedo might be purposely calibrated to real a touch slow; according to the dealership guys, it can be +/-5 % so mine is technically "within specs."

To answer you question about the trip computer's duration timer: it was within 1 minute of my stopwatch after 3 hours and change so it isn't that (I thought of this too since rate = distance/time). I really wish there was some way I could adjust the speedo to correct the display error as well as the trip computer's error since I went through the trouble of actually measuring them and because I can be a bit OCD about these sorts of things.

I can understand why legally the odometer can't be messed with, but the speedo and trip computer should have some adjustable feature just like the compass and the outside thermometer do. If for no other reason than as a means to correct an error introduced by using non-OEM sized wheels/tires.

darksky

5 minute rule prevented my from adding: If for no other reason than to provice a means to correct the speedo (not odometer) for an error introduced by using non-OEM sized wheels/tires which many of us are doing.

geekybiker

IIRC its US law ( and most 1st world international locations) That speedos must read high. The acceptable error varies, but in no case is a speedo reading low acceptable.

warjat

I noticed differences recently as well.
I was on the highway, and I set my cruise control to 120km/h.
Once up to that speed, I reset the trip computer.
My GPS was showing me a speed of 119km/h, but because I have brand new tires on my car, that's a forgivable difference.
However, even after driving for over 100km, my trip computer was showing me an average speed of 109km/h.
I haven't tried this at different speeds, so I don't know whether it's a difference of ~11, or 9%. Either way, that's a BIG difference!

paperboy42190

iTrader: (7)

+1, its been like this for every car I've ridden in. speedo always reads higher than actual speed(based on gps). its erred on the high side so people cant blame the manufacturer for speeding unknowingly.

LaCostaRacer

^ The trip computer is showing an average and is subject to round-off error. Even on a perfectly level road, I would think the cruise will very a little because of wind and other conditions.

If you really want to test your speedometer, use a handheld GPS that shows speed and see what that says.

Renescent

Does the MID get it's data from the ECU versus mechanical for the speedo (I don't know if speedo is mechanical/electrical/unicorn)? Maybe it's just an odd algorithm?

4drviper

when i go 70, and GPS agrees it's 70,

trip comp says 68... (when reset'd in motion)

always 2 less. (when over 50)