Reduction of linear timebase dataset to logarithmic

[liboiron]

I will be acquiring datasets with a large range in time (0.1 ns to 0.1 s) and my oscilloscope will only acquire linear timebases, so I will have to acquire 500,000 to 1,000,000 time points per experiment to get the time resolution I need in the nano and microsecond time range.  I would like to know if there's a way to use DIAdem to reduce the dataset to a logarithmic timebase of 2000-5000 points without averaging, i.e. by generating a logarithmic timebase (X values) by only pulling the X and Y points out of the dataset that follow a logarithmic timebase.  Simply, I'm trying to avoid using splines or n averaging that might distort my data, instead pulling the actual determined data points to generate a smaller logarithmic dataset from a much larger linear "mother" file.

Any assistance very much appreciated.
Regards,
Barry.

[Brad_Turpin]

Hi Barry,

Why do you specifically resist the idea of averaging the values together that belong inside one logarithmic interval?  You must have great confidence that each and every data point that you acquire is noise free.  If you want to look at a y vs. Log(x) or Log(y) vs. Log(x) graph of your data, then there would be no time distortion in averaging together the values that fit inside each logarithmic interval-- expressed as a single point on the log graph, but you would most likely eliminate any noise by doing so.

In direct answer to your question, if you want to pluck out only the lucky points that most closely align with the logarithmic instant in time, then you will need to do this with an explicit VBScript loop and the PNo() function.  If instead you are willing to average together all the values that are centered around that logarithmic instant in time, then you can use the built-in "Linear Mapping" function in DIAdem ANALYSIS.  Either way you will first need to generate a new channel which contains the logarithmic progression of x values-- this you can do with the built-in "Generate Numeric Channel" function in DIAdem ANALYSIS.

My recommendation is to try the 2 above dialogs in ANALYSIS and review the result yourself.  If you feel that the averaging is distorting the displayed results, then you can sit down and program the VBS loop, but my guess is that you won't need to.

Regards,
Brad Turpin
DIAdem Product Support Engineer
National Instruments