Decaying signal resonator power

[lukacsmw]

I am measuring the reflected power of a resonator for potential wireless applications. I'm pulsing the resonator with a 10.64GHz (resonate freq) pulsed waveform (10ns pulse width, 200ns period). Use an oscilloscope, it's easy to see the reflected pulse - with a maximum voltage just after the 10ns incident pulse. I've gated out the incident pulse and performed an FFT in Matlab to determine the power of the return pulse. The problem is that the power very much depends on the windowing of my FFT. As I enlarge the window, the power decreases by as much as 3-5dBm. From a practical perspective, what would be the best way to determine the reflected power of the signal?

I thought about taking an FFT using ever decreasing windowing, plotting the data and determining an intercept which would be the theoretical maximum power output - but I don't see that being very practical.

Thanks!

 

[waht]

From a practical perspective, what would be the best way to determine the reflected power of the signal?

Connect a directional coupler in reverse, and monitor the reflected power with a power meter, or a spectrum analyzer.

Or do a sweep with a network analyzer.

 

[lukacsmw]

I thought about the directional coupler in reverse. I'm actually using a circulator - but it accomplishes the same thing. The problem is that my reflected pulse contains both the resonator response and the reflected interrogation pulse. The resonator only absorbs about 4-5dB at the resonate frequency, so I'm still receiving a return interrogation pulse. The spectrum Analyzers I'm familiar with do not have the capability to remove that reflected pulse and determine the power of just the resonate response.

When I try to do so with an oscilloscope using the gating function to isolate only the resonate response, then I have the issue describe above where the power level drops as I enlarge the gating window to encompass more of the response.

 

[f95toli]

What kind of window are you using?
Somewhere at work I have a long review article about various windowing functions, when to use which window etc, if you want I can try to find it on Monday.

 

[lukacsmw]

The article on windowing functions sounds great - if you can find that article that would be greatly appreciated. I've implemented the window two ways, first using the mathematical gating function of my DPO7000 series oscilloscope, second just dumping the data into MATLAB and deleted the data points outside the window.

Both result in the same problem that the magnitude of the spectral response decreases as the windowing increases - meaning that I don't have a reliable, repeatable measure of the reflected power. I'm going to be taking several dozen measurements of some new resonator designs, and I need a repeatable methodology for reflected power.

Thanks!