How to measure the losses in microstrip using spectrum analyzer?

[prakashbg04]

hey guys, i need your help...i am doing a project regarding losses in microstrip. Do any of you knows how to measure the losses using spectrum analyzer(gsp 810 gwinstek->it is the spectrum analyzer model's name). I need manual for this. Thanks in advance:smile:

 

[enjunear]

Re: losses in microstrip

You would be much better off using a VNA (Vector Network Analyzer). Those devices are specifically made to measure all of your loss effects (return loss and insertion loss).

 

[prakashbg04]

Re: losses in microstrip

You would be much better off using a VNA (Vector Network Analyzer). Those devices are specifically made to measure all of your loss effects (return loss and insertion loss).

hello...but i cant afford to buy VNA...i am university student and i got a very limited budget to accomplish my project..my university only offers spectrum analyzer, so i gotta do this project using spectrum analyzer....any suggestion on how to measure the loss?thanks in advance

 

[albbg]

Re: losses in microstrip

As an option you can use the internal tracking generator (if it is installed).

1) connect the tracking output with the spectrum input usign the same cables you will use to connect the device. Probably you will have one cable from tracking to the microstrip and one cable from the microstrip to the spectrum. So to connect them toghether you can use an adapter. Measure the reading of the spectrum at the frequency you want.
2) remove the adapter and insert you microstrip and measure the reading of the spectrum at the same frequency of before
3) calculate the loss making the difference between the two readings

This measurement doesn't take into account the adapter loss (on the order of 0.1 dB or less at a frequency of 1 GHz) so this will be an error you will introduce.

 

[enjunear]

Re: losses in microstrip

As an option you can use the internal tracking generator (if it is installed).

1) connect the tracking output with the spectrum input usign the same cables you will use to connect the device. Probably you will have one cable from tracking to the microstrip and one cable from the microstrip to the spectrum. So to connect them toghether you can use an adapter. Measure the reading of the spectrum at the frequency you want.
2) remove the adapter and insert you microstrip and measure the reading of the spectrum at the same frequency of before
3) calculate the loss making the difference between the two readings

This measurement doesn't take into account the adapter loss (on the order of 0.1 dB or less at a frequency of 1 GHz) so this will be an error you will introduce.

This is about the only way to make your loss measurements using a spec analyzer. Beware that your resolution will not be very good, and you have no way of determining how much of the insertion loss is due to mismatch, and how much is due to true resistive losses in the PCB.

What kind of comparisons are you making with your insertion loss measurements? What are you trying to prove/disprove with your project?

 

[prakashbg04]

Re: losses in microstrip

This is about the only way to make your loss measurements using a spec analyzer. Beware that your resolution will not be very good, and you have no way of determining how much of the insertion loss is due to mismatch, and how much is due to true resistive losses in the PCB.

What kind of comparisons are you making with your insertion loss measurements? What are you trying to prove/disprove with your project?

basically my project is about verifying the existing formula. Let's say there is 3 formulas by different authors to calculate losses. So, I need to simulate the formula and observe how does the losses varies with the increase of frequencies(in my case I am varying from 100MHz to 500MHz). Later on I need to do the experiment part and measure the losses with the change of frequencies and finally I need to determine which formula is the best/closest to the experiment's result that I have carried out. So,basically this is what my project all about

 

[albbg]

I think if the losses or your microstrips are quite high (some dB) then the mentioned unaccuracy (adapter losses, substarte losses, etc.) could be neglected and you can roughly compare the experimental with theoretical data. Otherwise a much more accurate calibration is needed.

 

[prakashbg04]

hey guys..how do i calculate the losses from the spectrum analyzer?any formula??

 

[albbg]

Please refer to my previous post in which I explained you the procedure to measure the loss using the spectrum+tracking generator

 

[jiripolivka]

As you attempt to measure relatively short microstrip line, a low loss can be expected. Spectrum analyzers are calibrated in terms of relative power level, but the typical step size is ~10 dB, some instruments can be set to 1 dB resolution.
The best way to measure such low-loss is to use a power meter. Typically, a good power meter can have 0.1..0.2 dB resolution, so it is better suited for your task. If you only have a spectrum analyzer, you can build a simple RF detector with a 50-Ohm input impedance adjusted by a 50-Ohm resistor (use a SMD version of resistors and diodes to use it above 2 GHz). You can easily calibrate such detector with a good signal generator, and use it as a power meter with a good resolution. Its response will be temperature dependent, so make sure it will not change during measurements. The RF power level applied to it should not exceed 0 dBm (1 mV) for best results.