Apologies for the late reply.
The capture was done using some electronics (a controller) and a bit of imagination. Not as cheap or as convenient as Kim's method. I posted the image as it shows the coverage of a lsu 4.9 and the values that occur in a mechanical diesel engine compared to what Kim is seeing on a modern diesel. I would like to do the same test on a modern diesel along side Kim's method. Time and a donor vehicle required though.
Do you have the psdata file as I am intrigued by the units?
Did you utilise the custom probe settings for O2% level and PSI etc?
I did use the custom probe settings. The psi values are the standard ones (almost
The Lambda (L) values are taken from an output on the controller which is required to control and monitor pump current, as you know. The controller also has a data output so software can be used to monitor L or A/F ratio's. The output I used is separate from the data output and is user programmable. So I just programmed certain outputs for certain L values calculated from pump current (the values are then interpolated). This becomes an output from the controller. I then put the same value's in the custom probe interpolated again. Done. Sounds easier than what it was to figure out.
The Oxygen values are taken from a table that gives the info for given L values.
As for accuracy, that's hard to assess and part the reason I didn't follow through. In theory it should be accurate and it does seem to match the data output.
Also the output is a differential (noise reduction/accuracy) output and because I used all four channels I couldn't do a normal two channel/maths calculation for true output. So possible errors here. Of coarse the latest picoscope https://www.picoauto.com/automotive-oscilloscope.html
can get round this problem I think. As I said, proving or disproving the figures became a hurdle.
Temp was calculated using a basic calculation plus an additive allowing for cold junction compensation, but didn't allow for the Curie point. In this application it appears to be to large an error which again (at that time), was hard to prove or disprove the accuracy. But..... in the last couple of months may have found a way to check accuracy by using an Ad595(?) www.me.psu.edu/rahn/me462/AD594_5_c.pdf
chip. This multiplies thermocouple output by about 250 and applies cold junction compensation. The output can then be converted to temp by using the formulae and applying to K type nist tables. So will probably have another play at improving that (have had some thoughts about it), although this is unlikely to happen in the next few months.
I hope this answers your questions.
On a side note, I did run the wide band beside a narrow band during closed loop. That did record just over and under L1
Still not convinced about diagnostic values but tuning,