modulation, bit rate, lasers and diodes sources

I would, maybe naively, think that carrier recombination would matter only
in the case of direct, internal modulation. If the modulator was external
and the two sources, laser diode and non-laser diode were turned on to
operate in continuous mode, the pulses should look more or less the
same...
or not?
Will the coherence properties on the laser light give better quality pulses
once we reach a certain on-off keying speed?

I think external amplitude modulators (choppers) are faster than any
internal modulator....









>On 01/17/2011 05:22 PM, glen herrmannsfeldt wrote:
>> Tim Wescott<tim@seemywebsite.com>  wrote:
>> (snip)
>>
>>> I gather from your artificial distinction between 'laser' and 'diode'
>>> that you do not realize that a laser diode and a light emitting diode
are
>>> both diodes.  One emits non-coherent light, the other (with similar
>>> active material) exhibits laser action.  Both are semiconductor
diodes.
>>
>> Yes, so a laser diode is a subset of LED, but it is not usual
>> to describe it that way.
>>
>>> When a diode conducts it does so because one or both sides of the
>>> junction are flooded with minority carriers (electrons in P-type
>>> semiconductor, holes in N-type semiconductor).  These carriers
recombine
>>> with the local majority carriers.  The amount of time that it takes
for
>>> these carriers recombine is a limit to the speed of the diode.
>>
>> I hadn't thought of the confusion from carrier modulation to
>> current carriers when I wrote that.
>
>Until you posted your revelation of the difference between (charge) 
>carrier and (radio) carrier, I didn't either!
>
>-- 
>
>Tim Wescott
>Wescott Design Services
>http://www.wescottdesign.com
>
>Do you need to implement control loops in software?
>"Applied Control Theory for Embedded Systems" was written for you.
>See details at http://www.wescottdesign.com/actfes/actfes.html
>

On 01/18/2011 06:40 AM, fisico32 wrote:
(top posting fixed)
>
>> On 01/17/2011 05:22 PM, glen herrmannsfeldt wrote:
>>> Tim Wescott<tim@seemywebsite.com>   wrote:
>>> (snip)
>>>
>>>> I gather from your artificial distinction between 'laser' and 'diode'
>>>> that you do not realize that a laser diode and a light emitting diode
> are
>>>> both diodes.  One emits non-coherent light, the other (with similar
>>>> active material) exhibits laser action.  Both are semiconductor
> diodes.
>>>
>>> Yes, so a laser diode is a subset of LED, but it is not usual
>>> to describe it that way.
>>>
>>>> When a diode conducts it does so because one or both sides of the
>>>> junction are flooded with minority carriers (electrons in P-type
>>>> semiconductor, holes in N-type semiconductor).  These carriers
> recombine
>>>> with the local majority carriers.  The amount of time that it takes
> for
>>>> these carriers recombine is a limit to the speed of the diode.
>>>
>>> I hadn't thought of the confusion from carrier modulation to
>>> current carriers when I wrote that.
>>
>> Until you posted your revelation of the difference between (charge)
>> carrier and (radio) carrier, I didn't either!
 > I would, maybe naively, think that carrier recombination would matter
 > only in the case of direct, internal modulation. If the modulator was
 > external and the two sources, laser diode and non-laser diode were
 > turned on to operate in continuous mode, the pulses should look more
 > or less the same...
 > or not?

Yes, they would.

 > Will the coherence properties on the laser light give better quality
 > pulses once we reach a certain on-off keying speed?

Yes.  I think that we're all ignoring the dispersive property of fiber, 
which is going to tend to spread out an LED signal more than a laser 
signal -- but it's been a long long time since I needed to know this 
information, and I never needed to know more than the basics.

 > I think external amplitude modulators (choppers) are faster than any
 > internal modulator....

I believe that you are right.

-- 

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Do you need to implement control loops in software?
"Applied Control Theory for Embedded Systems" was written for you.
See details at http://www.wescottdesign.com/actfes/actfes.html

On Jan 17, 8:19&#4294967295;pm, glen herrmannsfeldt <g...@ugcs.caltech.edu> wrote:
> fisico32 <marcoscipioni1@n_o_s_p_a_m.gmail.com> wrote:
> > if distance and power density were not a problem for either a laser and a
> > diode, what does "how long it takes the carriers to recombine" exactly
> > mean? I am not clear on that...
>
> Oops. &#4294967295;In semiconductor physics, carriers are electrons and holes.
> After you turn off the current, it takes some time for them to be
> used up in light generation. &#4294967295;That is faster in lasers.
>
> > In any communication system there is a carrier signal of
> > central frequency f and the modulating signal (message signal)
> > that has a certain bandwidth.
> > A diode has a larger linewidth (bandwith) than a laser. Is there some
> > relation between the source linewidth and the bandwidth of the modulating
> > signal?
>
> Theoretically yes, but we are still a looong way from that.
>
> Optical frequencies are in the 1e14Hz range, so over four orders
> of magnitude from gigabit ethernet.
>
> -- glen

I recall a Lucent article a few years back (perhaps around a decade)
where they achieved a terabit per second data rate over fiber. They
employed both wavelength and polarization diversity.

Clay

fisico32 <marcoscipioni1@n_o_s_p_a_m.gmail.com> wrote:
(snip)

> I would, maybe naively, think that carrier recombination would matter only
> in the case of direct, internal modulation. If the modulator was external
> and the two sources, laser diode and non-laser diode were turned on to
> operate in continuous mode, the pulses should look more or less the
> same...
> or not?

Yes.  But all networking products that I know of, such as ethernet
(which comes in both LED and laser forms) modulate the current.

Because of the short cavity length, the linewidth of semiconductor
lasers is much larger than that of more traditional lasers
with long resonant cavities.   One can add an external resonator
to a semiconductor laser, and even add a diffraction grating to
control the feedback frequency.

-- glen

Tim Wescott <tim@seemywebsite.com> wrote:
(snip)

> Yes.  I think that we're all ignoring the dispersive property of fiber, 
> which is going to tend to spread out an LED signal more than a laser 
> signal -- but it's been a long long time since I needed to know this 
> information, and I never needed to know more than the basics.

Well, traditionally single mode fiber is used with laser
sources, and multimode with LED sources.  As LEDs aren't as
bright as lasers, the larger diameter multimode allows for more
light to get through.  (And multimode is cheaper.)  

With gigabit ethernet there was interest in using multimode
fiber for shorter distances with laser sources, and it was found
that it didn't work very well at all!

If you put 

"controlled launch" gigabit

into google, you should find the papers describing the problems
of using gigabit ethernet laser sources with multimode fibers.

The LED source excites most of the modes of a multimode fiber, and
things work just fine.  A laser source only excites a small number
of modes and, as it turns out, going right down the center 
hits the worst combination of modes.  The solution is "controlled
launch", which, using a short single mode fiber, goes into the
mutlimode fiber slightly off center.  That hits a better mode
distribution in the fiber, with much better results.

-- glen

Hello,
thanks for the info. I guess you guys are getting stuck with fiber (single
mode and multimode), with its own propagation and attenuation properties.
Surely a laser does better than a LED and can carry information without
degradation for a longer distance with less dispersion of any type...

But I guess my question is more of a physics question.
If we talk about free space, no fiber involved, with a transmitter and a
receiver separated by a distance of only 5 meters (very close so the power
density, higher in lasers, is not a factor) and used an external modulator,
will we be able to modulate a laser and a LED at the same speed or will the
LED have a poor performance once we start chopping its beam too fast?

I think so, but I have not yet found a place that would give me a very
clear and technical answer for that.....
thanks!












>Tim Wescott <tim@seemywebsite.com> wrote:
>(snip)
>
>> Yes.  I think that we're all ignoring the dispersive property of fiber,

>> which is going to tend to spread out an LED signal more than a laser 
>> signal -- but it's been a long long time since I needed to know this 
>> information, and I never needed to know more than the basics.
>
>Well, traditionally single mode fiber is used with laser
>sources, and multimode with LED sources.  As LEDs aren't as
>bright as lasers, the larger diameter multimode allows for more
>light to get through.  (And multimode is cheaper.)  
>
>With gigabit ethernet there was interest in using multimode
>fiber for shorter distances with laser sources, and it was found
>that it didn't work very well at all!
>
>If you put 
>
>"controlled launch" gigabit
>
>into google, you should find the papers describing the problems
>of using gigabit ethernet laser sources with multimode fibers.
>
>The LED source excites most of the modes of a multimode fiber, and
>things work just fine.  A laser source only excites a small number
>of modes and, as it turns out, going right down the center 
>hits the worst combination of modes.  The solution is "controlled
>launch", which, using a short single mode fiber, goes into the
>mutlimode fiber slightly off center.  That hits a better mode
>distribution in the fiber, with much better results.
>
>-- glen
>

On Jan 18, 5:29&#4294967295;pm, "fisico32" <marcoscipioni1@n_o_s_p_a_m.gmail.com>
wrote:
> Hello,
> thanks for the info. I guess you guys are getting stuck with fiber (single
> mode and multimode), with its own propagation and attenuation properties.
> Surely a laser does better than a LED and can carry information without
> degradation for a longer distance with less dispersion of any type...
>
> But I guess my question is more of a physics question.
> If we talk about free space, no fiber involved, with a transmitter and a
> receiver separated by a distance of only 5 meters (very close so the power
> density, higher in lasers, is not a factor) and used an external modulator,
> will we be able to modulate a laser and a LED at the same speed or will the
> LED have a poor performance once we start chopping its beam too fast?
>
> I think so, but I have not yet found a place that would give me a very
> clear and technical answer for that.....
> thanks!
>
>
>
> >Tim Wescott <t...@seemywebsite.com> wrote:
> >(snip)
>
> >> Yes. &#4294967295;I think that we're all ignoring the dispersive property of fiber,
> >> which is going to tend to spread out an LED signal more than a laser
> >> signal -- but it's been a long long time since I needed to know this
> >> information, and I never needed to know more than the basics.
>
> >Well, traditionally single mode fiber is used with laser
> >sources, and multimode with LED sources. &#4294967295;As LEDs aren't as
> >bright as lasers, the larger diameter multimode allows for more
> >light to get through. &#4294967295;(And multimode is cheaper.) &#4294967295;
>
> >With gigabit ethernet there was interest in using multimode
> >fiber for shorter distances with laser sources, and it was found
> >that it didn't work very well at all!
>
> >If you put
>
> >"controlled launch" gigabit
>
> >into google, you should find the papers describing the problems
> >of using gigabit ethernet laser sources with multimode fibers.
>
> >The LED source excites most of the modes of a multimode fiber, and
> >things work just fine. &#4294967295;A laser source only excites a small number
> >of modes and, as it turns out, going right down the center
> >hits the worst combination of modes. &#4294967295;The solution is "controlled
> >launch", which, using a short single mode fiber, goes into the
> >mutlimode fiber slightly off center. &#4294967295;That hits a better mode
> >distribution in the fiber, with much better results.
>
> >-- glen- Hide quoted text -
>
> - Show quoted text -


For your free space propagation example, use passive mode locked
lasers (these create trains of ultra short light pulses). These can
generate pulses with femto second durations. You won't get there with
just an LED. With multiple sources of such pulses you can gate each
source with either a pockels or kerr cell  and then combine the
mutliple outputs into a single beam to create a ridiculously high rate
of data on that light beam.

Does that help?

Clay

p.s. a femtosecond is 10^-15 second!!

On Jan 18, 5:29&#4294967295;pm, "fisico32" <marcoscipioni1@n_o_s_p_a_m.gmail.com>
wrote:
> Hello,
> thanks for the info. I guess you guys are getting stuck with fiber (single
> mode and multimode), with its own propagation and attenuation properties.
> Surely a laser does better than a LED and can carry information without
> degradation for a longer distance with less dispersion of any type...
>
> But I guess my question is more of a physics question.
> If we talk about free space, no fiber involved, with a transmitter and a
> receiver separated by a distance of only 5 meters (very close so the power
> density, higher in lasers, is not a factor) and used an external modulator,
> will we be able to modulate a laser and a LED at the same speed or will the
> LED have a poor performance once we start chopping its beam too fast?
>
> I think so, but I have not yet found a place that would give me a very
> clear and technical answer for that.....
> thanks!
>
>
>
> >Tim Wescott <t...@seemywebsite.com> wrote:
> >(snip)
>
> >> Yes. &#4294967295;I think that we're all ignoring the dispersive property of fiber,
> >> which is going to tend to spread out an LED signal more than a laser
> >> signal -- but it's been a long long time since I needed to know this
> >> information, and I never needed to know more than the basics.
>
> >Well, traditionally single mode fiber is used with laser
> >sources, and multimode with LED sources. &#4294967295;As LEDs aren't as
> >bright as lasers, the larger diameter multimode allows for more
> >light to get through. &#4294967295;(And multimode is cheaper.) &#4294967295;
>
> >With gigabit ethernet there was interest in using multimode
> >fiber for shorter distances with laser sources, and it was found
> >that it didn't work very well at all!
>
> >If you put
>
> >"controlled launch" gigabit
>
> >into google, you should find the papers describing the problems
> >of using gigabit ethernet laser sources with multimode fibers.
>
> >The LED source excites most of the modes of a multimode fiber, and
> >things work just fine. &#4294967295;A laser source only excites a small number
> >of modes and, as it turns out, going right down the center
> >hits the worst combination of modes. &#4294967295;The solution is "controlled
> >launch", which, using a short single mode fiber, goes into the
> >mutlimode fiber slightly off center. &#4294967295;That hits a better mode
> >distribution in the fiber, with much better results.
>
> >-- glen- Hide quoted text -
>
> - Show quoted text -

In another post I gave an example of how to create very short temporal
pulses.

To address what I think is your basic question is why a laser can be
faster than an LED. The answer to this lies with how the light is
produced.

Before Einstein came along we only talked about two ways light
interacted with matter. Simply an atom could either absorb light that
hit it and an energized atom could release light after an random
amount of time. But Einstein postulated a third type of interaction.
This third way is called stimulated emission. Essentially you have an
energized atom and a light particle (photon) passes by the atom and
without this particle being absorbed by the atom, the atom is coerced
into emiting its own photon that has the same frequency and phase as
the stimulating photon. Thus you can now guess at how lasers work via
Light Amplification via Stimulated Emission of Radiation.

Some lasing materials have low gain so the light must be passed back
and forth through a number of times to build the photon flux up to the
point of lasing (the number of stimulated phontons exceeds the number
of photons lost from the system). Some materials do have a high enough
gain that one pass of the light through the material is enough for the
pulse of light to stimulate photons from most of the excited atoms.
Recall polarization and doppler shifts will keep some of the atoms
from responding to the stimulative wave.

To learn more about these very short pulses, look up "passive mode
locked lasers" and "ultra short pulses." This should give you plenty
to read.

Clay