Measuring power with analog oscilloscope

[ayeaye]

An example of edge detect with gimp, difference of gaussians, filters -> edge-detect -> difference of gaussians, radius 1 was 0, radius 2 was 40, how to do that paint job now?

Ah, the cheapest analog multiplier i found in ebay was ad633, also by Analog Devices, $1.76 with shipping  http://www.analog.com/media/en/technical-documentation/data-sheets/AD633.pdf , would that do? One purpose i see for such integrated circuits, is that they enable to measure power even with a single channel oscilloscope with no external triggering. Also by using analog multiplication or addition, it's possible to calculate even a phase shift with these scopes, and get two traces the same as with a two channel scope. In an analog way, or then again for example by a python script doing calculations on traces drawn with gschem.

[ayeaye]

Ok, try to count the pixels of one trace. This is not power, as it is only a current, just an example to get an idea how it can be done.

Soooooooo, after a quite formidable painting job, i got the image below. I painted with ff0000 and 0000ff, but for some reason in the image it appeared as 254,0,0 and 0,0,254 . convert is a part of the imagemagick package.  convert capfill.jpg capfill.txt  converts it all to a text file, so that all can be seen. It also enables to count, what first comes to mind is doing replacement with gvim, but the text file is huge and this is not a good method. So the following is how to count red and blue pixels.

$ convert capfill.jpg -fill black +opaque "rgb(254,0,0)" -format %c histogram:info:
    271098: (  0,  0,  0) #000000 black
     23942: (254,  0,  0) #FE0000 srgb(254,0,0)
$ convert capfill.jpg -fill black +opaque "rgb(0,0,254)" -format %c histogram:info:
    273383: (  0,  0,  0) #000000 black
     21657: (  0,  0,254) #0000FE srgb(0,0,254)

So we got 23942 red pixels and 21657 blue pixels. For power we have to subtract one from the other, that is  23942 - 21657 = 2285 . The image is known to be 640 pixels wide, so we have to divide that by the width of the image in pixels  2285 / 640 = 3.570 . Now this is the average "power" in pixels. If we did calculate power, then this had to be multiplied by the scale of the oscilloscope trace, and divided by how many vertical pixels are in one scale, which is how much power is one vertical pixel.

The precision of that is not very great, because the trace was wide and all that was lost, plus any other inaccuracies in conversion. I say certainly this method is much less precise than doing it by drawing the traces with gschem, in spite of the primitivity of that method, unless one can say a method that is more accurate. And at least for me it took much more tome to do all that image manipulation, than to draw it with gschem. Both on the image of the oscilloscope screen, and by hand just drawing in gschem, the latter was the fastest, in spite of figuring where all these points should be. But one should decide by oneself what method fits one the best.

So hope you got some idea of measuring power with analog oscilloscope, with any ways to do that.

[ayeaye]

This, drawing traces in gschem, and then doing a processing using a python script, enables to do many other things. Like say we hypothetically have the most simple oscilloscope that has only one channel, and no external triggering or external sync. For example in the circuit above, we measure voltage on the resistor and capacitor (like the voltage in channel 2 above), voltage on the resistor, and also voltage on the capacitor, and draw all these three traces in gschem the same as we did draw two traces above. Measure just one after another, at different times, not simultaneously.

Then it is possible to write a python script, that adds the traces of voltages on the resistor and on the capacitor, and compares the result to the trace of the voltage on the resistor and capacitor, with a various phase shift, finding the phase shift with which the resulting trace matches that trace the most, thus finding the phase shift between the traces. Then that same python script can write an sch file with traces of the current and the voltage with the correct phase shift, as if they were measured with a two channel oscilloscope. And from that the python script above can calculate power.

So with the method like this, the same measurements can be made with the simplest one channel oscilloscope without external sync or triggering, as with a two channel oscilloscope, without any additional equipment or circuit, including no analog multiplier. In case one uses a very old oscilloscope with only a single channel, or such. Everything can still be done.

gschem and a python script can also be used for the simplest graphing of measured values, without using anything more fancy for drawing graphs.

Hope it was useful for someone for some purpose.