A friend discovered that the raw AN voltages are logged incorrectly starting with v1.100 of FCEdit.

The datalogit box uses an 8-bit A/D converter that can read values from 0 - 5V for a resolution of 5/255. In versions previous to 1.100, the full 14 character string equivalent was stored in the log, but for some reason beginning with v1.100, the values are simply truncated to 3 decimals, which loses the exponential part. Therefore, you end up with this situation:

So in other words, any voltage values logged that are less than 0.1V are erroneous. Maybe not such a big deal, except when looking at averages, max, or min.</v1.100></v1.100>

 

I imagine truncating the values marginally improved performance. The effect this would have in terms of accuracy depends on the voltage scale of whatever you are logging.

For example, my innovate LC-1 by default will output a range from 7:1 to I think 24:1 AFR. With that kind of scale it would be more sensitive to truncated voltage readings. Innovate recommends though that you rescale your 0-5v output from 9:1 to 16:1 , meaning you wouldn't need accuracy down to as many decimal places.

 

yeah this might simply be a way to improve logging speeds with the datalogit. i doubt it will make a significant impact in logging, especially with AFR's. since 0.1v would be pretty much full rich. and full rich is usually between 8:1- 10:1 afr for the sensor, depending what you or the manufacturer programs it to be. Unless you really really tune rich these values would never be hit.

On the other hand some people log other 0-5v things that might be affected. though i cant think of anything precise enough to matter.

 

I have posted how to fix the DL AUX setup so that DL logs record the same values as the O2 WB's display. This insures accuracy do to any software or hardware differences.

I called it calibrating the DL to the WB.

 

I think you guys are missing the point.

The point is 0.1E-02 = 0.001 does NOT equal 0.1. Versions 1.100 and later chop off the exponential part. Has nothing to do with calibration because it's a math error caused by FC-Edit's internal conversion to a string and then truncating the length. Like I said, it probably only matters for average/min/max, or like Dudemaaan said, if you have a sensor that is less than 0.1V.

 

True 1.96 E-02 is .0196 not 1.96. But you are missing the REAL point.
That is why they made multiple upgrades to the software to fix problems.
Who gives a **** about the early bad software! The current is 1.120. Check it out.

I have a FD3S beige box. I use the FD3S software for updating my maps, but use the UNIVERSAL software for logging as it is faster. It logs the AN RAW data to 3 decimal points without any exponential and is accurate.

 

1.120 > 1.100

 

Here are my three tests with only the ignition turned on and no O2 WB hooked up so it defaults to setup voltage.

FD3S v 1.082; 5.88E-02 = .0588
FD3S v 1.120; .059
UNIV v 2.125; .059

My tests shows that the last version of FD3S is accurate and so is the latest UNIV.
They went from 4 to 3 decimal digits and rounded the 3rd decimal digit.
Viewing the 1.082 data and 1.120 data with 1.120 software gives the same AFR if viewed with 3 decimal places.