head spinning on stripchart relative time

[ElectroLund]

I have worked with CVI for many years and by far the hardest thing to wrap my head around is strip chart x-axis time formatting.  Killing hours here trying to get this right.

I have some data that dumps into a serial port from an instrument at a rate of 5 milliseconds per frame.  I have a multithreaded function that receives the data.  That function has a while loop to read from the serial port and then add it to the strip chart.  Easy.

Except for the x-axis.  I want to display relative time based on the known 5ms frame rate.  Not so easy:

SetCtrlAttribute(panel, chart, ATTR_XFORMAT, VAL_RELATIVE_TIME_FORMAT);
SetAxisTimeFormat(panel, chart, VAL_BOTTOM_XAXIS, VAL_RELATIVE_TIME_FORMAT, "%#M:%S.%1f");

SetCtrlAttribute(xrfGraphPanel[GPH_LOG], GPH_LOG_CHART, ATTR_XAXIS_GAIN, 5.0); 
SetCtrlAttribute(xrfGraphPanel[GPH_LOG], GPH_LOG_CHART, ATTR_XAXIS_OFFSET, 0.0);		

I have read many threads here, each with conflicting advise about which should be gain/offset.  None of this combinations plot accurate timescale for me.  Here's a view of the data:

I have a Timer() call before and after the data dump, giving me 1.391 seconds, which I believe.  Yet, my time scale on the strip chart never formats correctly.

Thoughts?

[CVI-User]

Never did this before with a strip chart but I got curious. Does this help: Specifying Dates and Times for Axis Values ?

Excerpt:

If you use the relative time format, LabWindows/CVI expects that you plot data that contains the number of seconds since some arbitrary point in time, commonly denoted as t₀.

When the axes are in absolute time or relative time format, the value you specify for ATTR_XAXIS_OFFSET or ATTR_YAXIS_OFFSET is the t₀ value. The value you specify for ATTR_XAXIS_GAIN or ATTR_YAXIS_GAIN is the Δt value.

[gdargaud]

I remember getting pissed at the time axis of strip charts long ago also. So I wrote my own: https://gitlab.com/dargaud/TimeStripLabel