| |
 |
|
|
Science Forum Index » Optics Forum » Double transmission peak narrow band filter?
Page 1 of 1
|
| Author |
Message |
| Salmon Egg |
 Posted: Tue Apr 29, 2008 12:02 am |
|
|
|
Guest
|
I know how to design a narrow band transmission filter. Off-hand I do
not see to extend the design to two or more transmission peaks. For
example, how can you design a thin-film filter with sharp transmission
at 450nm and 550nm.
For a single transmission line, you put a half-wave layer between two
stacks of repeated quarter-wave HL.
Bill |
|
|
| Back to top |
|
| Guest |
| Posted: Tue Apr 29, 2008 12:21 pm |
|
|
|
|
On Apr 28, 10:02 pm, Salmon Egg <Salmon...@sbcglobal.net> wrote:
Quote: I know how to design a narrow band transmission filter. Off-hand I do
not see to extend the design to two or more transmission peaks. For
example, how can you design a thin-film filter with sharp transmission
at 450nm and 550nm.
For a single transmission line, you put a half-wave layer between two
stacks of repeated quarter-wave HL.
Bill
Wild Bill
For the single you should have the air | HLHLHHLHLH | glass, as you
stated.
For a 2 cavity you will have air | HLHLHHLHLH L HLHLHHLHLH | glass.
The L layer in between is called a coupling layer.
If you want the band narrower just as more HL and LH to the outside.
Michael
www.oscintl.com |
|
|
| Back to top |
|
| Salmon Egg |
| Posted: Tue Apr 29, 2008 7:06 pm |
|
|
|
Guest
|
In article
<ffb7dab2-083d-4beb-abc6-334371b87f0e@l17g2000pri.googlegroups.com>,
mpate@oscintl.com wrote:
Quote: Wild Bill
For the single you should have the air | HLHLHHLHLH | glass, as you
stated.
For a 2 cavity you will have air | HLHLHHLHLH L HLHLHHLHLH | glass.
The L layer in between is called a coupling layer.
If you want the band narrower just as more HL and LH to the outside.
I am going to look at this in more detail. Off hand, I am guessing this
amounts to what in electrical circuits is called coupling two resonant
circuits together. As the coupling coefficient gets increased, you go
from a single peak to critical coupling and into a double peak. This is
often used to shape pass bands. I do not see how you go on to separated
sharp transmission peaks.
The fundamental problem, as I see it, is that a single peak filter of
this kind works because there is only one wavelength at which you get
the equivalent of absentee layers formed by pairing of the quarter-wave
layers in the stacks around the true half wave absentee layer. I was
hoping that there is a clever way around this limitation.
Bill |
|
|
| Back to top |
|
| Coater |
| Posted: Wed Apr 30, 2008 4:52 pm |
|
|
|
Guest
|
One way is to design minus filters that block the required zones, and have
their edges on each side of the bandpass you are interested in. Lots of
layers, and lots of precision required.
Another way is to spring for a good thin-film program, design a bandpass,
then optimize for the shape you want. The results will also be lots of
layers, and the precision will be just as demanding.
Just out of curiosity, I did this, and it is possible to pass 450nm & 550nm
(roughly 10nm FWHM) while blocking elsewhere.
Good Luck
"Salmon Egg" <SalmonEgg@sbcglobal.net> wrote in message
news:SalmonEgg-CF46B3.22021028042008@news.la.sbcglobal.net...
Quote: I know how to design a narrow band transmission filter. Off-hand I do
not see to extend the design to two or more transmission peaks. For
example, how can you design a thin-film filter with sharp transmission
at 450nm and 550nm.
For a single transmission line, you put a half-wave layer between two
stacks of repeated quarter-wave HL.
Bill |
|
|
| Back to top |
|
| Salmon Egg |
| Posted: Thu May 01, 2008 12:20 am |
|
|
|
Guest
|
In article <SS5Sj.65$oJ4.3135863@petpeeve.ziplink.net>,
"Coater" <na@nothere.com> wrote:
Quote: One way is to design minus filters that block the required zones, and have
their edges on each side of the bandpass you are interested in. Lots of
layers, and lots of precision required.
Another way is to spring for a good thin-film program, design a bandpass,
then optimize for the shape you want. The results will also be lots of
layers, and the precision will be just as demanding.
Just out of curiosity, I did this, and it is possible to pass 450nm & 550nm
(roughly 10nm FWHM) while blocking elsewhere.
Lots of layers appear necessary for this kind of thing--and that is the
problem.
Using electrical engineering terminology, (That gterminology should be
used more in optics than it is now.) an antireflection coating is an
impedance (index) matching network. At a single frequency, that is easy
to do on paper. It becomes more difficult if you try to match over a
large bandwidth and incident angle range.
Almost any coating you add on will tend to lead toward mismatch and
increased reflectivity unless designed not to do so. It is no trick to
deposit multiple (HL)^n multilayers with different normalized
wavelengths to end up with broadband high reflectivity. The narrow band
filter design is a good trick using the properties of absentee layers to
index match at the peak transmission wavelength while staying mismatched
at other wavelengths. I do not know how to take advantage of absentee
layers to provide a second transmission peak.
Bill |
|
|
| Back to top |
|
| Coater |
| Posted: Thu May 01, 2008 9:12 am |
|
|
|
Guest
|
"Salmon Egg" <SalmonEgg@sbcglobal.net> wrote in message
news:SalmonEgg-A2B595.22204230042008@news.la.sbcglobal.net...
Quote: In article <SS5Sj.65$oJ4.3135863@petpeeve.ziplink.net>,
"Coater" <na@nothere.com> wrote:
One way is to design minus filters that block the required zones, and
have
their edges on each side of the bandpass you are interested in. Lots of
layers, and lots of precision required.
Another way is to spring for a good thin-film program, design a bandpass,
then optimize for the shape you want. The results will also be lots of
layers, and the precision will be just as demanding.
Just out of curiosity, I did this, and it is possible to pass 450nm &
550nm
(roughly 10nm FWHM) while blocking elsewhere.
Lots of layers appear necessary for this kind of thing--and that is the
problem.
That's the nature of the beast. We can't change physics, so we have to live
with that fact.
Quote: Using electrical engineering terminology, (That gterminology should be
used more in optics than it is now.)
You say tomahto, I say tomato.
Most coating designers are aware of the electronic vs optical analogies,
since the math is identical. But many, if not most of us are quite
comfortable with the terminology we already use.
<snip>
Quote: impedance (index) matching network. At a single frequency, that is easy
to do on paper. It becomes more difficult if you try to match over a
large bandwidth and incident angle range.
Almost any coating you add on will tend to lead toward mismatch and
<snip>
Quote: I do not know how to take advantage of absentee
layers to provide a second transmission peak.
Bill
I gave you two techniques that work. Just slipping in an extra half-wave
layer where you want it, is not the way to get the job done.
And thanks for the dissertation on coating design. I'm assuming you meant it
for the rest of the readers, and not for me.
Coater |
|
|
| Back to top |
|
| Salmon Egg |
| Posted: Thu May 01, 2008 2:03 pm |
|
|
|
Guest
|
In article <GdkSj.70$oJ4.3155993@petpeeve.ziplink.net>,
"Coater" <na@nothere.com> wrote:
Quote: And thanks for the dissertation on coating design. I'm assuming you meant it
for the rest of the readers, and not for me.
True.
In regard to using circuit techniques for thin-film designs, I did not
much use by coaters of the concepts of characteristic or iterative index
(impedance) for repeating layer combinations. The same for propagation
constants. I found those concepts very useful.
Bill |
|
|
| Back to top |
|
| |
|
Page 1 of 1
All times are GMT - 5 Hours
The time now is Mon Oct 13, 2008 3:14 pm
|
|