FJO Analog Input AFR PID

[heavyfoot]

I have an older FJO WBO unit. Model CWC0002. I want to bring the AFR Analog signal into the backbox as an input. Does anyone have the calc pid for it?

[emarkay]

What are the output voltages of the FJO unit?

[heavyfoot]

Their table in the manual only went to 15.0, but their graph went to 20.0 so I had to interpolate.
AFR Volts
10.00 0.81
10.25 0.91
10.50 1.01
10.75 1.11
11.00 1.21
11.25 1.31
11.50 1.42
11.75 1.52
12.00 1.61
12.25 1.71
12.50 1.80
12.75 1.88
13.00 1.97
13.25 2.06
13.50 2.14
13.75 2.22
14.00 2.30
14.25 2.38
14.50 2.45
14.75 2.51
15.00 2.54
15.25 2.59
15.50 2.62
15.75 2.65
16.00 2.68
16.25 2.71
16.50 2.74
16.75 2.77
17.00 2.80
17.25 2.82
17.50 2.85
17.75 2.88
18.00 2.89
18.25 2.92
18.50 2.95
18.75 2.98
19.00 3.01
19.25 3.03
19.50 3.05
19.75 3.06
20.00 3.08

thx

[heavyfoot]

:nixweiss: I got in touch with them for a table. Here is what they sent. I have not tried to calibrate their controller to the black box yet. So the connected and not connected voltages are not correct.

I tried to use the manage attachments button to upload the excel file but it failed. So I copied and pased the data into a text file. Hopefully you can see it. 725 Please note after I get a DVM and perform the connected and not connected calibration the correction factor numbers change. Do I need to do that first before I can come up with the equation for Black box logging? I was thinking I should be able to figure out the equation fromt the spreadsheet but it looks like to be it's going to be some type of a calculus function. Been too many years since I had that.

:nixweiss:

[heavyfoot]

Let me update all of you. My FJO Wideband is what they now call a Legacy Unit. I went to the FJO web site and downloaded an excel spreadsheet and guide for calibrating the Wideband itself. Basically its a table That has a column with AFR and a column with corrected voltage. You hook up the wideband, removing the sensor from the exhaust pipe. Let the sensor warm. Keep analog output of FJO unit disconnected from EFI Live black box. Use Digital Multimetere to read and record the voltage. Now since car is not running and sensor is sitting in free air you will see 20.00 AFR. Which gave me a disconnected value of 4.82 volts. Now simply plug the FJO analog input into the black box and record the connected voltage. I saw 4.83 volts. Since the two values are close that means there is a large amount of impedance in the black box. This is good because more impedance means better isolation for the signal. Once I entered the connected and disconnected values into the spreadsheet it gave me the corrected voltage for a given AFR. That part was simple. Here is where I struggled. I then needed to create a FJO Pid in the calc_pid.txt file. In order to figure things out I used the expression from the LC1 pid. I saved the file and added the PID to one of the charts to evaluate the signal. I was seeing a signal but of course the AFR number was wrong. So how I do I come up with the correct AFR? Here is what I did. I copied the two columns of data (AFR and corrected voltage) onto a new worksheet. I then Plotted the data using an excel chart. Be careful this is where it gets tricky. When placing the chart select the XY (Scatter Plot) Scatter with data points connected by smoothed lines without markers. This will plot the data on a point by point basis without having fixed increments on the x or y axis. Origionally I used the wrong chart so I kept coming up with the wrong expression. What tipped me off was the fact my chart did not look like their plot of the same data. Ok you now have a chart of AFR versus corrected voltage. So how did I figure out the equation? Lot's of calculus and differential equations. Nope. I used a function in excel. Left click on the plotted line (it will highlight). Right click then select "add trendline." On the "Type" tab Select "Polynominal". Initially set "order" to a value of "2." Click on the "options" tab select (check the boxes) "display equation on chart" and "display R-squared value on chart." Click "ok." On the cart you will now see an equation and a R squared value. You should see the trend line basically overlapping the origional plot. The R squared value is a measure of how accurate the equation is. A R squared value of 1.0 is perfect. I got a value of 0.9846. This means my equation will be correct 98.46% of the time. If your R squared value is lower you can make it more correct by increasing the order. As you do the trend line will get to the point where it will completely overlap the origional plot. Click on the trendline you added. Right click then select "format trendline." Increase the "order" number. I increased my "order" to a value of 4. Doing so gave me a new R squred value of 0.9995. Now my trendline entirely overlaps the origional plot. The equation is what you enter into the efilive calc_pids.txt file. Here is mine.

There is another way to do it. By calculating the span of AFR and span of volts you can figure out volts/afr. This however is a linear equation and will only be accurate at two points on the curve. If you use 10. and 20.0 afr as the limits it will be way off around afrs of 14.67. if however you use 2.5 and 15.0 as the limits it will be more accurate. But still not be totally correct. Thats why I ended up abandoning the equation and went with the polynominal equation. I entered equation into calc pid file, re-installed sensor and fired car up. the AFR recorded by EFI live is exactly what I see with my FJO software. Sweet! I put in this blurb cause I figure the same approach for finding the equation can be done with anyones wideband. All you need is to enter data into excel for AFR and voltage.

[joecar]

You can also use the lookup() function in your calculated pid...
see page 93 in Scan Tool Manual pdf.