Go mba. Frankly, little value for you to go phd. Let alone cost/benefit trad e off. Economic policies refer to mostly two fields, macro theory and macro develop ment. You can have a look at publications in these two fields and compare th em to your goal.
You background should be OKey. In the coming years, doing PHD does not sound a good choice unless one has visa or job issues. But in the long run, curiosity and motivation can make huge differences in many areas including Economic research or practice. If you has the money to support your living, the patience to improve your crafts, and the skills to communicate your ideas, I don't see why you cannot make a good transition. Just my 2 cents.
You are right on this sense. But talking about spectrum you are wrong. Yes we don't loss any wavelength component of the light in the pulse, but spectrum of wave is a concept of the shape of the wave, not exactly the wavelength inside it.
【在 A*u 的大作中提到】 : 波形展宽不是因为高频被滤掉了 : : 的高
A*u
24 楼
shape of the pulse depends on optical spectrum and the phases of all the components 测一个optical spectrum难道还能得到不同optical frequency component的相位的信 息? 那测一个optical spectrum就可以知道shape of the optical pulse了?
【在 z*****n 的大作中提到】 : You are right on this sense. But talking about spectrum : you are wrong. : Yes we don't loss any wavelength component of the light : in the pulse, but spectrum of wave is a concept of the : shape of the wave, not exactly the wavelength inside it.
【在 z*****n 的大作中提到】 : yes, its highest frequency part is cutoff.
z*n
26 楼
These are two different concepts. The wavelength components combination is not the same as spectrum. Just thin king of the pulse sent by the Tx, it also contains all the wavelengths, but the pulse shape is more rectangular.
【在 z*****n 的大作中提到】 : You are right on this sense. But talking about spectrum : you are wrong. : Yes we don't loss any wavelength component of the light : in the pulse, but spectrum of wave is a concept of the : shape of the wave, not exactly the wavelength inside it.
【在 z*****n 的大作中提到】 : These are two different concepts. : The wavelength components combination is not the same as spectrum. Just thin : king of the pulse sent by the Tx, it : also contains all the wavelengths, but the pulse shape : is more rectangular.
b*r
30 楼
Both dispersion and nonlinear effect should be negligible for 4m of fiber transmission. From the data sheet of your fiber, you should be able to calculate L_d(dispersion length) and L_n(nonlinear length) for your fiber. And they should be much large than 4 meters. So I would say it is the bandwidth of your fiber limits.
【在 b****r 的大作中提到】 : Both dispersion and nonlinear effect should be negligible for 4m of : fiber transmission. : From the data sheet of your fiber, you should be able to calculate : L_d(dispersion length) and L_n(nonlinear length) for your fiber. And : they should be much large than 4 meters. : So I would say it is the bandwidth of your fiber limits.
【在 z*****n 的大作中提到】 : These are two different concepts. : The wavelength components combination is not the same as spectrum. Just thin : king of the pulse sent by the Tx, it : also contains all the wavelengths, but the pulse shape : is more rectangular.
c*u
33 楼
Of course dispersion will affect a 30 nm bandwidth pulse in 4 m of fiber, especially if the center wavelength is not 1310 nm or it's specified zero-dispersion wavelength. 30 nm is very wide bandwidth for most applications. Let's say dispersion 5 ps/nm/km, then for 4 m and 30 nm, it is almost 0.6 ps. The bandwidth is narrowed to almost 6 nm. On the other hand, it's very rare that a fiber can't support 30 nm bandwidth.
【在 b****r 的大作中提到】 : Both dispersion and nonlinear effect should be negligible for 4m of : fiber transmission. : From the data sheet of your fiber, you should be able to calculate : L_d(dispersion length) and L_n(nonlinear length) for your fiber. And : they should be much large than 4 meters. : So I would say it is the bandwidth of your fiber limits.
c*u
34 楼
hehe, you read book too much. The light from candle and the spectra of the candle still follows the Fourier transform. The light from candle can be approximated as a white noise-type random process due to the central-limit theorem (many sources are emitting independently). The Fourier transform of the white noise is of course wide-band in the frequency domain. I am confused why people here got so confused about the "wavelength" and "spectrum". Wavelength is just the x-axis of spectrum. That's
i don't agree with the last couple of paragraphs If all the wavelengths are in phase, it is the fourier transform limited pulse. see http://en.wikipedia.org/wiki/Bandwidth-limited_pulse and delta frequency * delta t =constant If they are not in phase, say due to dispersion, the above equation is not valid any more. In a fiber, if there is no absorption and no nonlinear effect, all the wavelength components should be preserved and no new one should be created.The reason why it is broadend is as
【在 c*u 的大作中提到】 : hehe, you read book too much. The light from candle and the spectra of : the candle still follows the Fourier transform. The light from candle : can be approximated as a white noise-type random process due to the : central-limit theorem (many sources are emitting independently). The : Fourier transform of the white noise is of course wide-band in the : frequency domain. : I am confused why people here got so confused about the "wavelength" and : "spectrum". Wavelength is just the x-axis of spectrum. That's
A*u
36 楼
how do you get the 6nm bandwidth?
【在 c*u 的大作中提到】 : Of course dispersion will affect a 30 nm bandwidth pulse in 4 m of fiber, : especially if the center wavelength is not 1310 nm or it's specified : zero-dispersion wavelength. 30 nm is very wide bandwidth for most : applications. Let's say dispersion 5 ps/nm/km, then for 4 m and 30 nm, it : is almost 0.6 ps. The bandwidth is narrowed to almost 6 nm. : On the other hand, it's very rare that a fiber can't support 30 nm : bandwidth.
b*r
37 楼
I would say you need to read or any Coherency optics book, the appendix of will do also. Talking about the spectrum of optical signal without considering the temporal coherency of light is simply wrong. Anyone who learned physical optics should understand the importance of temporal coherency and spatical coherency, saying that temporal coherency has no effect on spectrum is just wrong. Simply put, the spectrum detected by any OSA is NOT the fourier transfor
【在 c*u 的大作中提到】 : hehe, you read book too much. The light from candle and the spectra of : the candle still follows the Fourier transform. The light from candle : can be approximated as a white noise-type random process due to the : central-limit theorem (many sources are emitting independently). The : Fourier transform of the white noise is of course wide-band in the : frequency domain. : I am confused why people here got so confused about the "wavelength" and : "spectrum". Wavelength is just the x-axis of spectrum. That's
b*r
38 楼
From the temporal broadening I guess, but still this is not a FT transform relation.
【在 A*u 的大作中提到】 : how do you get the 6nm bandwidth?
b*r
39 楼
Yes, dispertion might impact on a 30nm signal. That is why I suggest him to calculate L_d. But follow your logic, if we have a 1000km fiber, the signal will disperse like crazy, you might get a much broader signal. And FT of the signal gives you a delta function in frequency domain - do you think that is physically possible? Why we need something called filter? Any highly dispersive structure should do the work, right?
【在 c*u 的大作中提到】 : Of course dispersion will affect a 30 nm bandwidth pulse in 4 m of fiber, : especially if the center wavelength is not 1310 nm or it's specified : zero-dispersion wavelength. 30 nm is very wide bandwidth for most : applications. Let's say dispersion 5 ps/nm/km, then for 4 m and 30 nm, it : is almost 0.6 ps. The bandwidth is narrowed to almost 6 nm. : On the other hand, it's very rare that a fiber can't support 30 nm : bandwidth.
A*u
40 楼
i guess this problem should not be discussed in EE
gives structure
【在 b****r 的大作中提到】 : Yes, dispertion might impact on a 30nm signal. That is why I suggest him : to calculate L_d. : But follow your logic, if we have a 1000km fiber, the signal will disperse : like crazy, you might get a much broader signal. And FT of the signal gives : you a delta function in frequency domain - do you think that is physically : possible? Why we need something called filter? Any highly dispersive structure : should do the work, right?
b*r
41 楼
Since every signal should follow FT, can you explain to me why sun light is cw-like in time domain, but also has a very broad spectral? Cw in time domain gives you delta function in frequency domain, right? You simply can not explain this without coherency theory.
【在 c*u 的大作中提到】 : hehe, you read book too much. The light from candle and the spectra of : the candle still follows the Fourier transform. The light from candle : can be approximated as a white noise-type random process due to the : central-limit theorem (many sources are emitting independently). The : Fourier transform of the white noise is of course wide-band in the : frequency domain. : I am confused why people here got so confused about the "wavelength" and : "spectrum". Wavelength is just the x-axis of spectrum. That's
r*s
42 楼
加个精吧 我怕找不到了
b*r
43 楼
Yes, I forgot there is something called transform-limited laser pulse. I would say this concept is not widely understood outside of the ultra-fast people.
the
【在 A*u 的大作中提到】 : i don't agree with the last couple of paragraphs : If all the wavelengths are in phase, it is the fourier transform limited : pulse. : see http://en.wikipedia.org/wiki/Bandwidth-limited_pulse : and delta frequency * delta t =constant : If they are not in phase, say due to dispersion, the above equation is not : valid any more. In a fiber, if there is no absorption and no nonlinear : effect, all the wavelength components should be preserved and no new one : should be created.The reason why it is broadend is as
z*n
44 楼
no, I mean the spectrum of pulses in time domain signal.
ok. then we may be talking about different things. but i think lz is aksing about optical bandwidth, which should correspond to power spectral density
【在 z*****n 的大作中提到】 : no, I mean the spectrum of pulses in time domain : signal.
h*w
46 楼
1. You should state what kin d of pulses you are using? fs, ps or ns 2. what is the peak power? so we can see if there are nonlinear effects. 3. Assuming there are no nonlinear effects, for a typical fiber either 1300nm or 1500nm or visible, there are no problems for conducting 30 nm light source. 4. Never confuse with electrical signal transmission. 5. If there are nonlinear effects and assuming your fiber and the connection are damn good. Yes, pulse spectrum can be narrowed due to self phase m
c*k
47 楼
~100fs pulse width I don't know the peak power, but the average power is 30mW. I used a SMF fiber @ 1550nm connection be
【在 h*w 的大作中提到】 : 1. You should state what kin d of pulses you are using? fs, ps or ns : 2. what is the peak power? so we can see if there are nonlinear effects. : 3. Assuming there are no nonlinear effects, for a typical fiber either : 1300nm or 1500nm or visible, there are no problems for conducting 30 nm : light source. : 4. Never confuse with electrical signal transmission. : 5. If there are nonlinear effects and assuming your fiber and the connection : are damn good. Yes, pulse spectrum can be narrowed due to self phase : m
A*u
48 楼
what about repetition rate?
【在 c******k 的大作中提到】 : : ~100fs pulse width : I don't know the peak power, but the average power is 30mW. : I used a SMF fiber @ 1550nm : connection : be
c*k
49 楼
35Mhz
【在 A*u 的大作中提到】 : what about repetition rate?
c*k
50 楼
I tried to measure the spectral width of the pulse using OSA when propagating a 1m fiber and a 3m fiber,respectively. The spectral [email protected] FWHM is decreased to be 20nm with the 3m fiber compared to the spectral width 32nm with the 1m fiber.
since it is a fs pulse, the shape is only dependent on wavelength components and their phases.
【在 z*****n 的大作中提到】 : You are right on this sense. But talking about spectrum : you are wrong. : Yes we don't loss any wavelength component of the light : in the pulse, but spectrum of wave is a concept of the : shape of the wave, not exactly the wavelength inside it.
A*u
52 楼
you can estimate the peak power assuming a Gaussian or simply rectangular shape with the averager power and the rep rate. im not familiar with the threshold for SPM though
【在 c******k 的大作中提到】 : 35Mhz
h*w
53 楼
30 mw 100fs 35m sure they do have nonlinear effects. Check your pulse chirp. Free space laser or fiber laser?
m*n
54 楼
这个讨论好呀, 其实siegmann 的书的chapter9 写的很好, 只有没有chirp的pulse才 是transform limited, short pulse after fiber is highly chirped but the spectrum should not change unless bandwidth limited or non linear effect for example self phase modulation etc. That can be easely acumulated by fs pulse in fiber.
c*k
55 楼
请提供书名,谢过
for example self phase modulation etc. That can be easely acumulated by fs pulse in fiber.
【在 m****n 的大作中提到】 : 这个讨论好呀, 其实siegmann 的书的chapter9 写的很好, 只有没有chirp的pulse才 : 是transform limited, short pulse after fiber is highly chirped but the : spectrum should not change unless bandwidth limited or non linear effect for example self phase modulation etc. That can be easely acumulated by fs pulse in fiber.
b*r
56 楼
You can check out: by G.P. Agrawal, should be enough. Apparently you are in the application side, why bother knowing the bandwidth of your laser?
【在 c******k 的大作中提到】 : 请提供书名,谢过 : : for example self phase modulation etc. That can be easely acumulated by fs : pulse in fiber.