c******k
发帖数: 1140 | 1 比如10nm的spectral width的 laser pulse, 经过一段光纤色散后，脉冲宽度展宽了，
请问这时的它的spectral width有变化吗？ |
c******k
发帖数: 1140 | 2 我自己也用光谱分析仪测了色散前和色散后的脉冲的spectral width. 结论也是变窄了。
有什么公式或者理论支持这一现象吗? 多谢 |
c******k
发帖数: 1140 | |
A*u
发帖数: 111 | |
A*u
发帖数: 111 | 5 fiber的transmission在10nm之内应该可以看作是平的，
所以我觉得应该不会有spectral width的变化
但如果有非线性效应， 如supercontinuum
spectral width会变宽
【在 c******k 的大作中提到】
: 比如10nm的spectral width的 laser pulse, 经过一段光纤色散后，脉冲宽度展宽了，
: 请问这时的它的spectral width有变化吗？
|
A*u
发帖数: 111 | |
c******k
发帖数: 1140 | 7 我举个例子是10nm的，我们实际上用的是30nm 的脉冲，经过四米的光纤传播后用光谱
分析仪测脉冲频谱， 发现 Delta_Lambda变成15nm了？
【在 A*u 的大作中提到】
: fiber的transmission在10nm之内应该可以看作是平的，
: 所以我觉得应该不会有spectral width的变化
: 但如果有非线性效应， 如supercontinuum
: spectral width会变宽
|
A*u
发帖数: 111 | 8 建议用两条同一类型fiber一长一短再试试
【在 c******k 的大作中提到】
: 我举个例子是10nm的，我们实际上用的是30nm 的脉冲，经过四米的光纤传播后用光谱
: 分析仪测脉冲频谱， 发现 Delta_Lambda变成15nm了？
|
z*****n
发帖数: 7639 | 9 yes, its highest frequency part is cutoff.
【在 c******k 的大作中提到】
: 比如10nm的spectral width的 laser pulse, 经过一段光纤色散后，脉冲宽度展宽了，
: 请问这时的它的spectral width有变化吗？
|
z*****n
发帖数: 7639 | 10 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
发帖数: 111 | 11 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.
|
A*u
发帖数: 111 | 12 这个高频能被反映在optical transmission里面吗？
如果不能，它为什么能被光谱仪探测到呢？
【在 z*****n 的大作中提到】
: yes, its highest frequency part is cutoff.
|
z*****n
发帖数: 7639 | 13 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 A*u 的大作中提到】
: 这个高频能被反映在optical transmission里面吗？
: 如果不能，它为什么能被光谱仪探测到呢？
|
b****r
发帖数: 2555 | 14 大哥，你不是学光学的吧。
首先光纤的色散不影响频谱分布，这里不能简单的用傅立叶变换来套。光学信号有个
相关性的概念，它的时域和频域分布很可能不是傅立叶变换对。简单的例子，蜡烛光
在时域上是cw信号，但是频域上也是一个很宽的信号，这里就不是傅立叶变换对的关
系——因为蜡烛光没有相关性。
光纤色散从物理上来说就是各个频谱分量在光纤中的传输速度不一样，所以时域上表
现为脉冲会展宽，但是频域上可能不会受影响。比如说现在最新提出来的光纤传输概
念，就是有意识的增强光纤色散来压制非线性，最后在利用负色散来恢复信号在时域
上的分布。这个过程中信号的频谱就不会变化——roughly speaking。
光纤本身的传输带宽和色散没有什么关系，而是被光纤的材料以及设计所决定。也就
是所谓的cut-off wavelength。
回到原来的问题，脉冲展宽不说明脉冲在频谱上会变窄，没有必然联系。但是脉冲在
光纤中传输可能会变窄，原因是光纤本身的带宽，或者非线性。1.55微米30nm的信号，
大概是5THz的带宽，光纤的带宽可能够，也可能不够。可以设计一个简单的实验。
把激光输出光强降低10倍，再看通过光纤后的 |
b****r
发帖数: 2555 | 15 只有色散的话，频谱形状也不会变啊。色散就是大家都过去了，只不过到达的时间
不一样。但是这个到达时间的差别在ps量级，对光谱仪使用的探测器来说是忽略不
计，因为反正是一个积分的过程。
光纤的传输函数和色散没有关系。
【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.
|
b****r
发帖数: 2555 | 16 光谱仪测到的不是简单的频谱，而是信号自相关函数的傅立叶变换（？）。我回头
去查一查wolf那本书去。
thin
【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
发帖数: 2555 | 17 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 c******k 的大作中提到】
: 我举个例子是10nm的，我们实际上用的是30nm 的脉冲，经过四米的光纤传播后用光谱
: 分析仪测脉冲频谱， 发现 Delta_Lambda变成15nm了？
|
A*u
发帖数: 111 | 18 还可能根laser和fiber fiber和光谱仪的耦合有关
【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.
|
A*u
发帖数: 111 | 19 optical spectrum 就只有optical wavelength 的信息
你说的spectrum是optical spectrum吗？能被光谱仪探测到吗？
thin
【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
发帖数: 916 | 20 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
发帖数: 916 | 21 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
【在 b****r 的大作中提到】
: 大哥，你不是学光学的吧。
: 首先光纤的色散不影响频谱分布，这里不能简单的用傅立叶变换来套。光学信号有个
: 相关性的概念，它的时域和频域分布很可能不是傅立叶变换对。简单的例子，蜡烛光
: 在时域上是cw信号，但是频域上也是一个很宽的信号，这里就不是傅立叶变换对的关
: 系——因为蜡烛光没有相关性。
: 光纤色散从物理上来说就是各个频谱分量在光纤中的传输速度不一样，所以时域上表
: 现为脉冲会展宽，但是频域上可能不会受影响。比如说现在最新提出来的光纤传输概
: 念，就是有意识的增强光纤色散来压制非线性，最后在利用负色散来恢复信号在时域
: 上的分布。这个过程中信号的频谱就不会变化——roughly speaking。
: 光纤本身的传输带宽和色散没有什么关系，而是被光纤的材料以及设计所决定。也就
|
A*u
发帖数: 111 | 22 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
发帖数: 111 | 23 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
发帖数: 2555 | 24 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
发帖数: 2555 | 25 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
发帖数: 2555 | 26 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
发帖数: 111 | 27 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
发帖数: 2555 | 28 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
发帖数: 590 | |
b****r
发帖数: 2555 | 30 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
发帖数: 7639 | 31 no, I mean the spectrum of pulses in time domain
signal.
【在 A*u 的大作中提到】
: optical spectrum 就只有optical wavelength 的信息
: 你说的spectrum是optical spectrum吗？能被光谱仪探测到吗？
:
: thin
|
A*u
发帖数: 111 | 32 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
发帖数: 134 | 33 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
发帖数: 1140 | 34
~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
发帖数: 111 | 35 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
发帖数: 1140 | 36 35Mhz
【在 A*u 的大作中提到】
: what about repetition rate?
|
c******k
发帖数: 1140 | 37 I tried to measure the spectral width of the pulse using OSA when
propagating a 1m fiber and a 3m fiber,respectively.
The spectral width@ FWHM is decreased to be 20nm with the 3m fiber compared
to the spectral width 32nm with the 1m fiber.
【在 c******k 的大作中提到】
: 比如10nm的spectral width的 laser pulse, 经过一段光纤色散后，脉冲宽度展宽了，
: 请问这时的它的spectral width有变化吗？
|
A*u
发帖数: 111 | 38 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
发帖数: 111 | 39 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
发帖数: 134 | 40 30 mw 100fs 35m sure they do have nonlinear effects. Check your pulse chirp.
Free space laser or fiber laser? |
|
|
m****n
发帖数: 68 | 41 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
发帖数: 1140 | 42 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
发帖数: 2555 | 43 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.
|
