I like to watch!

 

I say keep it simple, and just go with changing the ECT only. Look how the opinions differed over just the temp bars.

 

Interesting. I might be interested in this product.

 

I broke down and bought a ECT sensor and did some more testing. I put it in a pot of oil (M1) on the stove and recorded the temperature and resistance as it heated up. This took about 15 minutes.

Then I watched it as it cooled down and noticed the temps did not match what I saw on the way up. I did some experimenting and determned that the sensor changes temperature much more slowly than my thermocouple temperature probe, so what was happening is the ECT sensor was always 15 to 30 seconds behind the temp probe, resulting in incorrect readings.

I did a cool-down test (over an hour) while constantly stirring the pot and got much more consistent results. Just stirring would change the resistance by 5 or 10 ohms, as the ECT sensor, being halfway out of the oil and in cooler room temp air, would read cooler than the oil unless you kept it stirred.

I did several "constant" temp tests by tweaking the flame and moving the pot to keep the temp and resistance readings fairly constant over several minutes. Once I achieved this "equilibrium" I figured the sensor had stabilized, so I used those measurements to verify the cool-down readings were accurate. And they were, but the heat-up temps were no where close, sometimes being off by 3 or 4 ohms, sometimes by 30 ohms.

I then plotted the cool-down measurements on my original chart of the OBD temperatures, and they are reasonably close. I think the original OBD measurements are a little high, since I took the data as soon as the OBD showed a temperature change, instead of finding the "center" of the temperature range. If I corrected for that it would move the OBD curve to the left slightly, putting it even closer to the boiling oil measurements.

I'll probably use an average of the two for building the gauge project.

.

edited to correct for screwed up Excel chart first posted 6PM

 

Hi Modifry, this is VERY interesting and nicely done work on your part. Can you draw any conclusions about the various EMS ECT calibrations that are floating around out there (per our email exchange)? Is the calibration supplied by AEM anywhere close to being right?

CBender

 

And a big thank you to Mrs Modifry for putting up with pans full of oil on the stove.

 

[QUOTE=cbender,Sep 22 2004, 06:45 PM]Can you draw any conclusions about the various EMS ECT calibrations floating about (per our email exchange)?

 

I'm not sure she knows where the stove is.

 

LOL I take it she doesn't read this forum.

 

Mr. Bender - see, nagging does help.

Measured Ohms - cool-down test with constant stirring
Stabilized ohms - random measurements taken after several minutes of constant oil temperature, achieved with much fiddling of the stove settings. Some error here because fiddling with the knobs meant I couldn't stir the pot as much.
GG & GR - Tests done by others
AEM .94 - The AEM calibration data I got from you (I assume SW version .94)
AEM 1.03 - The AEM calibration data I got from you (I assume SW version 1.03)

It appears that the first AEM calibration was a perfectly straight line. Too bad the sensor output isn't. I think this is called "it made the software easy".

On the second attempt they got a little bit of a curve, but it's still nearly-straight. The only good news is they got one data point right and it's where it's most important. Other than that, both of the AEM calibrations look way off to me.



.