Comparing data with different recording frequencies/times

[Don'tTellMe]

I have been through a number of posts and still can't crack this one. Hopefully someone will find this an easy question.

I am trying to compare two power measuring devices. Annoyingly, both devices record at slightly different intervals. One at 0.55 sec intervals the other at 0.50

I'm thinking the best thing to do is to create two sets of data averaged by the second.

I found an old thread that seemed to be trying to do the same thing averaging data over 30s.

DonkeyOte suggested an =1+Countif function, but I can't get it to work (I think that there have been a number of changes to the countif function since then).

Any help with the attached would be gratefully recieved.

[Kaper]

I think the best idea would be to approximate changes between points by some function. The second easiest (and probably in most cases good enough) will be just linear approximation.
I am skipping the easiest, but much less precise possibility - just using Vlookup and finding Watts at given time or if no value at this exact time, the last value measured before that time.

So I'd prepare common moments in time in one column (say G) spaced every half second.
Then in H2 (and copy all the way down):

Please Login or Register  to view this content.

as a helper value for next calculations. I am using helper column, to keep formulas shorter, but of course this hepler column formula could be nested into main formula.
in I2 (and down) the main formula:

Please Login or Register  to view this content.

and similar formulas in J and K columns for the second set of measurements (see attachment).
And then you have reasonable estimates of both wattages in the same moment.
Of course, instead of preparing column with common time one could use one time axis as a reference, and calculate second wattages at the same time as were measured first.

See attached file with the graph of differences Watts1-Watts2 every half second.
The second graph is just to show how the approximation works (in a short timeslot 00:04-00:10)

[Don'tTellMe]

Brilliant. Thanks for the quick response.

I'll work those formulas into my data and see whether I can make it work.