| Author |
 Topic  |
|
|
waveguidethz
7 Posts |
 Posted - 08/12/2012 : 3:56:55 PM
|
Origin Ver. and Service Release (Select Help-->About Origin):
Operating System:
1. i have some polymers characterized by transmission terahertz time domain spectroscopy (THz-TDS).
2. i have saved the time domain data into Excel, with column A being (Delay ps), column B (Amplitude nA) as following shows.
3. now i copy column A being (Delay ps) and column B (Amplitude nA) directly onto OriginLab (units of time ps and amplitude nA have to be change or not???)
4. do a FFT on this data
5. conresponding output options in plot checked. ( someone told me to change factor to +1 (science), and i did, but dun know why)
6. the frequency, amplitube, phase, and power etc are outputted to new worksheets in the data workbook.
was i doing the correct steps?
coz i found the calculation of refractive index of the polymer wrong by using the data of phase in data workbook of originlab.
|
|
|
greg
USA
1378 Posts |
Posted - 08/13/2012 : 3:03:43 PM
|
The Units header row is for reference only and is not taken into consideration for FFT or any other analysis.
Since your time data is in pico seconds, the output frequencies will be in terahertz.
The only settings in FFT which affect Phase output are
Factor - Which only affects the sign of the phase
and
Unwrap phase - which can have an effect on magnitude
Any calculations based on Phase result should take the above into consideration. |
 |
|
|
Drbobshepherd
USA
Posts |
Posted - 08/14/2012 : 6:09:48 PM
|
Did you run a reference transmission measurement? You may get the correct result if you ratio your sample data to that of an empty transmission line. Then the phase shift will be proportional to the product of the refractive index and the thickness of your sample. Also, don't forget to adjust for off-axis incidence if you are using a waveguide.
DrBobShepherd |
|
|
|
waveguidethz
7 Posts |
Posted - 08/15/2012 : 04:59:52 AM
|
quote:
Originally posted by greg
The Units header row is for reference only and is not taken into consideration for FFT or any other analysis.
Since your time data is in pico seconds, the output frequencies will be in terahertz.
The only settings in FFT which affect Phase output are
Factor - Which only affects the sign of the phase
and
Unwrap phase - which can have an effect on magnitude
Any calculations based on Phase result should take the above into consideration.
thank you for telling frequencies outout in terahertz. that is what I want exactly, otherwise tedious calculations have to done.
atctually, i wanna to get phase difference between two signals (one is Reference and other being Sample). by doing this, i have first to:
1. get the phase data of Reference (as above shows)
2. then get the phase data of Sample
3. finally subtract phase of Reference from Sample and i have the phase difference.
i just wonder whether using the data of phase outputted to new worksheets in the data workbook for my calculations is proper?
sorry it might be so confusing, as i am a complete beginner with Originlab....
|
|
|
|
waveguidethz
7 Posts |
Posted - 08/15/2012 : 05:34:42 AM
|
quote:
Originally posted by Drbobshepherd
Did you run a reference transmission measurement? You may get the correct result if you ratio your sample data to that of an empty transmission line. Then the phase shift will be proportional to the product of the refractive index and the thickness of your sample. Also, don't forget to adjust for off-axis incidence if you are using a waveguide.
DrBobShepherd
thx for answering. it seems uve done the similar measurements?
1. i already run a reference transmission measurement.
2. what does it mean by " phase shift will be proportional to the product of the refractive index and the thickness of your sample"
the equation i am using for refractive index is
where d(cm) is thickness of the sample, Co is the speed of light in vacuum, w is the angular frequecy, and O is the relative phase
my understanding of relative phase is subtracting phase of Reference from Sample. and of angular frequecy is 2*pi*f*10^12
where f is frequency in column Freq(X) from data workbook, being multiplied by 10^12 since it is in terahertz.
so now i just wonder whether the phase data in data workbook can be used drectly for my calculations???
 |
|
|
|
Drbobshepherd
USA
Posts |
Posted - 08/15/2012 : 11:48:30 AM
|
Yes, I've done this type of measurement many times, and I can tell you from experience, it isn't easy.
By "phase shift", I mean the same thing as "phase difference" or "relative phase".
Please do not take offense, but I need to know if you realize that in order to use your equation, phase must be converted to radians. The phase in yout FFT results is obviously in degrees. Could this be your problem?
I noticed that the phase of your 1st data point(f=0) is 180 degrees. That doesn't make sense to me. I think you have forced a 180-deg shift in your FFT results; this would be correct for a relectance measurement but not a transmission measurement. However, if you did this for both reference and sample measurments, I do not think this should affect your final result.
Your equation looks correct to me. Theoretically, you should simply subtract reference phase from your sample phase to obtain relative phase. Remember, Phi must be in radians, and d and Co and must employ the same units of length (i.e. d in cm and Co in cm/s), and f must be in Hz.
Good luck,
DrBobShepherd
Measured data is band-limited so your FFT results will not make sense outside the band limits of your setup. |
|
|
|
waveguidethz
7 Posts |
Posted - 08/21/2012 : 04:06:56 AM
|
quote:
Originally posted by Drbobshepherd
Yes, I've done this type of measurement many times, and I can tell you from experience, it isn't easy.
By "phase shift", I mean the same thing as "phase difference" or "relative phase".
Please do not take offense, but I need to know if you realize that in order to use your equation, phase must be converted to radians. The phase in yout FFT results is obviously in degrees. Could this be your problem?
I noticed that the phase of your 1st data point(f=0) is 180 degrees. That doesn't make sense to me. I think you have forced a 180-deg shift in your FFT results; this would be correct for a relectance measurement but not a transmission measurement. However, if you did this for both reference and sample measurments, I do not think this should affect your final result.
Your equation looks correct to me. Theoretically, you should simply subtract reference phase from your sample phase to obtain relative phase. Remember, Phi must be in radians, and d and Co and must employ the same units of length (i.e. d in cm and Co in cm/s), and f must be in Hz.
Good luck,
DrBobShepherd
Measured data is band-limited so your FFT results will not make sense outside the band limits of your setup.
Dear DrBobShepherd,
I would like to express my sincere gratitude to you for your detailed explanation.
I was originally wondering whether the phase should be in degrees or radians for calculationg refractive index untill your teachings, same for definition of relative phase and units of length used.
As for the 180-deg shift, though it does not make sense, correct results still can be found if it is done to both reference and sample measurments.
Now the final results come out all right.
Once again, thank you!
|
|
|
| |
Topic |
|
|
|
correct steps to perform FFT?