I know other folks on the forum have reported their temperature sensors being off by several degrees! Has anyone gotten this fixed to their total satisfaction yet?

If so, what did your dealer do to resolve the issue?

Thanks!

 

Bump. No one has any ideas on this one?

 

How, precisely, did you determine that your temp sensor was off? What was the procedure you used?

 

I initially thought my 04TL outdoor thermometer was reading close to 10 degrees high. To test it I held a digital thermometer out the open sunroof and took measurements on three occasions. The readings were within one degree of what the 04TL thermometer was reading.

-r

 

Using a travel clock that has a digital temp gauge built into it that I've had for several years, I tried two different overnight tests ...

1) I put in the garage next to the front bumper, and checked it in the AM. Clock said 50. Car said 54.

2) I left the car outside overnight and put the clock in the same place by the front bumper, and checked in the AM. Clock said 28. Car said 36.

 

my car is also off by 4 sometimes 5 degrees

 

Ever thought your clock is off???

 

Was the clock going over 19 mph?

You might want to read the manual.
Sometimes, you can learn wonders. . .

Eneg

 

I drive by a TV station that has weather posted on a big display...I gotta believe it is accurate, otherwise shame on the weatherman...and my Audi is spot on and the TL....sorry don't know yet! Based on what my home outside thermometer says (which is always within on degree of the Audi) the TL is within a couple of degrees.

 

The scientific way to test your TL's thermometer is to place a known-accurate thermometer within an inch or so of the TL's temp sensor (anyone know exactly where that is?) and compare readings. Driving by a bank termometer or sticking a thermometer out the sunroof is not a valid test since there's no guarantee the air temperature in those areas is the same as the air surrounding the TL's temp sensor, especially if it's located under the hood next to the (hot) engine.

 

I did have the same problems, and went to my dealer for fixing.

They hooked up PGM to adjust my system, and lowered it four degree down. I have to say I prefer to seeing low temp instead of higher one.

I got my conclusions after comapring to home thermometer and other vehicles' temp readings. The dealer said to me I can have any standard I like, just tell them how many degree up/down you'd like to have.

 

I also had my dealer adjust mine. They told me they can do it in 3 degree increments. Driving by several banks one day, my TL was 4-8 degrees higher than they were.

I just got the car returned with no chance to check it. As another poster suggested, I would prefer to be wrong on the cool side than the hot side. Other than vanity or fun, I would like to know when the temperature is around 32 degrees and the roads start freezing. That is why I would like to get as close as possibly to reality. I drove a loaner 2004 TL and it was also 4-5 degrees high.

 

That's what I tried to make sure before the dealer adjusted my temp. They cannot make wrong.

Also, the report they gave to me show:
"Cust states outside temp is running 5 degrees too high-advise"
"112 via PGM test-adjust temp resading down4"

I don't understand why your dealer said it's only do it by 3 degree increments, then my dealer said 4 instead.

 

I just visited my dealer today... He told me there was no adjustment that can be made. He said the best test was taking the sensor, dipping in ice water...(32 degrees) and measuribg electrical resistance. He said this is a replace module. I saw it right in their tech manual. How did I know it was off?
I was driving in a blinding snowstorm, and it read 42 degrees!
Mine is generally 6-10 degrees high...
I would like to know what procedure was used to "adjust" the temp instead of a "remove and replace" of the sensor.
I too, would really like to know when the roads start frrezibg!
Thanks

 

If you go to the MID display, you can call up the outside temp... Hold the reset button for about 10 seconds... and an adjustment will scroll through... +3 +2 +1 0 -1 -2 -3 ... When you want to lower it 3 degrees, release the reset button when it hits -3.
Any more than that and you have to change out the sensor as stated above.

 

Calibration

It can be adjusted by plus or minus 3 degrees!!!

1) Turn ignition switch to ON position (II).

2) Select the outside temp on the MID display. Press and hold the SELECT/RESET button for 10 seconds. When you continue to hold the button, the display will scroll through the temp settings from +3 to -3 degrees (celcius).

3) When the desired correction value is displayed, release the button and the recalibrated air temp will be displayed. Each time a correction value is entered, it replaces the previous value.

 

Didn't someone say you can adj. again, another + or- 3 degrees? This adjust it yourself proceedure outlined above is not in the manual and apparently some dealers don't know about it based on the posts.

I adjusted mine down 3 and now it reads (rolling into the garage) within a degree of my digital wireless Radio Shack thermometer...which has two sensors that agree when both are in the shade. I have one in back of the house and one in front...so that when the sun is on one I can look at the other to get the "real" temp.

 

Yes, someone speculated that it could continue to be bumped down after having adjusted it, but this turned out not to be the case for the 04 TL as it apparently might be for other model years.

 

Ravenhwk,

As I responded in your earlier posting about how to calibrate the sensor (here), the offset adjustment only gets you so far.

Note:For the following example, I'm going to use Maxim's DS60 chip, because it has convenient packaging, an it's inexpensive (and you can bet that car manufacturers take that into account).

Temperature conversion requires a temperature transducing sensor (Say, Degrees F into Volts), and then something that converts the output of the sensor into your temperature display.

Ideally, the sensor would have some sort of linear equation. An ideal linear (straight-line) equation looks something like this:

Volts Output = (Gain) * (Temperature - [Offset Value]).

For the DS60 sensor the Gain is (1/288) V/Deg F, and the Offset Value is 90.112 Deg F (I converted everything to Deg F for everybody. If you're over the pond, I hope you don't mind. ). So, if we stick those values in, our DS60 sensor's voltage equation now looks like:

Volts = (1/288) * (Temperature + 90.112)

Playing with a little algebra, we can find out the outside temperature from the equation above:

Temperature = (288 * Volts) - 90.112

The car's computer wil use an equation like this to convert the voltage read from the sensor into your temperature display (in Deg F). Note that the "Offset Value" in this equation is -90.112.

With me so far? If so, let's try out the equation. Let's say that the sensor is generating 0.53 Volts. Plugging that into the equation, we'll get:

Temperature = (288 * 0.53) - 90.112

This would result in a value of 62.528 Deg F, which would probably display as 63 Deg F on your display, because the car's computer will probably round up the value since it is over 62.5 Deg (this is a common practice).

How does the Acura Temperature Calibration Adjustment come into play? Basically, you will be modifying the "-90.112" value. If you adjust the Temperature Calibration by "-1", you change the "-90.112" into "-91.112". If you adjust the Temperature Calibration by "+3", you change the "-90.112" into "-87.112". Do you see how this works? The calibration could be very handy, BUT ...

OK ... now for the real world. What I described above is a theoretical ideal. There's two big monkey wrenches that I'm going to throw in.

1) There is no such thing as a perfectly linear commercially available temperature transducer - at least not yet. All existing temperature transducers have a certain amount of error in them, some more than others. The most accurate temperature transducers are thermocouples (used by the National Institute of Standards and Technology for highly calibrated temperature references), and they are moderately expensive. On top of that, they are very non-linear. Most thermocouple characteristics can be approximated by 8th-order polynomial equations. Using some decent algorithms, a thermocouple sensor can accurately produce temperature readings with +/- 0.1 Deg of error.

Now, here's the important thing: Your car isn't going to be using a thermocouple sensor. It'll use some commercially available, off-the shelf sensor, like the DS60. So, I just did a quick check on "www.maxim-ic.com" to see what the published error is. A typical DS60 sensor will have +/- 2.0 Deg C of error, which will correspond to +/- 1.8 to +/- 3.6 Deg F of error. "Hey, that's great!" you think, "Acura's adjustment will allow me to fix this, because I can adjust up to +/- 3 Deg F." Unfortunately, that's not true.

Remember how I said that there aren't any linear sensors out there? Well, the +/- 2.0 Deg C (+/- 3.6 Deg F) of error is available across the entire range. So, when the outside temperature is 110 Deg F, you could be off by +3.4 Deg F, and read 113 Deg F. When the outside temperature is 74 Deg F, you could be off by -3.2 Deg F, and read 71 Deg F. When the outside temperature is 20 Deg F, you could be off by +2.8 Deg F, and read 23 Deg F. The sensor is working as designed, and it is staying within specification.

As long as the sensor is accurate to +/- 3.6 Deg F across the entire range, it's working to specification.

2) To truly get the accuracy described for Maxim's DS60 chip (or any other chip, for that matter), high precision voltage sources and circuit calibration is required for each device being produced. It may even include hand-picking of parts to be installed, depending on the design. In our case, each "device being produced" is an Acura TL. This is an expensive process, that usually requires a human to run testing across multiple temperature ranges with a precision temperature source. This is very expensive. Do you think that Acura does hand calibration of the temperature sensor circuitry for each car as it is being produced? I sure don't.

Now that I've described these two issues, I'll just state that there are other items that can make the temperature display inaccurate. I've only described the characteristics of the sensor. There are other electronic parts involved, such as Analog/Digital converters.

The end result is, you won't be able to get much satisfaction from Acura (or any other car manufacturer, for that matter) if your diplay is off. You could have a sensor that is out of specification. Your sensor could be within specification, but the voltage going to the circuit could be out of spec. The software that analyzes the data can induce errors. There's lots of things that can go wrong, and Acura won't be able to reasonably respond to all of them.

I hope I've given you a little more insight into what's going on behind the scenes. I wish I could have given you an answer you would have liked, though.

żGotJazz?

PS: Get the impression that I've done some temperature sensing design work in the past?

 

apparently more acurate: it agrees with
AM radios current temps , and gives no major gllitches like showing 46 degree in snow. When I test drove a new 6mt it showed 52 degree, mine showed a more accurate 46 degrees, more accurate with my butt feel termometer,

 

was exactly same as on my bmw !!!1

 

:lol2: I should say so

 

Z Factor -

I have to pay for this stoopid car somehow!

 

I'm not saying that anybody's display has to be inaccurate. What I am saying is that there are a lot of reasons why your display may be inaccurate, and some of those reasons are difficult to control.

The sensor described above is a typical cost-effective sensor. There are more expensive sensors that can provide more accurate readings. One of the best sensors that Maxim produces has about half the errors that I described above. I don't know what sensors are actually being used by Acura or BMW. The reason that I selected the DS60 was because it could be the sensor Acura uses, and it showed the problems better because of it's error levels.

When a manufacturer makes any kind of sensor, they publish the "Nominal" values for the sensor, and they also publish the "minimum" and "maximum" values as well. If you get a part whose characteristics land exactly on the "nominal" values, you're going to see more accurate readings. However, if you get a part whose characteristics land exactly on the "maximum" or "minimum" values, you will be seeing the greatest amount of errors allowed by the sensor. Most of the parts land on the "nominal" values, but some don't. If you are unlucky enough to fall on the "maximum" or "mnimum" ranges ... oh, well .

Other things that can affect your display are where the sensor is mounted, and what calculations are used to determine a "stable" temperature. You can read more about that: here.

 

do you really believe those displays outside building are correct? temp readings are relative to where the display's sensors are. likely they arent where you are.

 

No, I almost never trust the temperature displays on buildings. Getting a consistantly accurate temperature reading is tricky. You need to put the sensor:

(1) where there is no shade,
(2) with constant low-level air movement,
(3) high enough off the ground to avoid surface effects.

Why would I know? Check out this interesting site!

(OK, I may be a little biased about how interesting that site is!)

If you want to get a more accurate temperature reading for your area, you might want to try: Weather Underground. They will give you the temperature readings from your nearest airport. Airport temperatures are very important (they determine landing characteristics), so they tend to be more accurate. Depending on your airport's data update rate, you might find this useful.