Date: Wed, 25 Apr 2001 16:11:01 -0700
From: Alan Weinstein <ajw@hep.caltech.edu>
To: jay@ligo.caltech.edu, kstrain@physics.gla.ac.uk, coyne@ligo.caltech.edu,
   fritschel_p@ligo.mit.edu, dhs@ligo.mit.edu, janeen@ligo.caltech.edu,
   gari@ligo.caltech.edu, kells_b@ligo.caltech.edu, smith_m@ligo.caltech.edu,
   jordan@ligo.caltech.edu
Subject: 40m ROC question

Hi all,

I would be grateful if you could spend a few minutes
thinking about and giving me your opinions
about the following issue;
we can't move forward with 40m optics
until it's resolved.

I've asked about this in a previous email,
but only got a few responses, insufficient
to make a decision; so I'm trying again,
with more numbers.

The AdvLIGO big thinkers (you know who you are!)
are considering blowing up the beams in the arms,
to reduce thermal noise, which scales inversely
with some power of beam size.

Shall the 40m do the same?
We can't specify ROC for any of the core optics
until we decide that.

Of course, we won't have thermal problems at the 40m.
The only reason I can think of to emulate this
AdvL innovation is so that the Guoy phase
between the ITM and ETM is the same
between 40m and AdvL, so that the WFS alignment matrix,
cavity stability, HOMs, etc, are identical;
there will of course remain big differences
due to the factor 100 in the arm length.

Here's some numbers:


            g   ROC_ITM  ROC_ETM  w_ITM   w_ETM  Guoy Phase
                  (m)      (m)     (mm)    (mm)    (radians)

LIGO I     1/3    7400    14250    36.3    45.6    0.955

AdvL FSi  0.785  35000    35000    54.0    54.0    0.482
AdvL SAP  0.860  54000    54000    60.0    60.0    0.386

40m, flat-curved, current baseline design:
40m-1      1/3     inf     57.4    3.03    5.24    0.955

40m, flat-curved, blown up:
40m-2     0.785    inf      178    4.98    5.62    0.482

40m, symmetric, 1/3:
40m-3      1/3    90.5     90.5    3.98    3.98    0.955

40m, symmetric, 0.785:
40m-4     0.785    336      336    5.29    5.29    0.482


where:
g = (1-L/R)_ITM * (1-L/R)_ETM

w's are field amplitude 1/e radii, power 1/e^2 radii;
    I *think* that the diameter of the beam
    with (1 - 1ppm) power = 5.257 * w (for TEM00)

And here are some questions:

- AdvLIGO experts: what do you say to this question?

- ASC experts: given the factor 100 difference in arm length,
  does this effort to match Guoy phase make the 40m-2
  a better test of alignment control than 40m-1?

- Janeen, does that 1ppm beam diameter run into aperture
  problems with the 3" beamsplitter? (w_BS ~= w_ITM).

- Gari, are there any restrictions associated with the
  specification of the ROC's?


- Mike, do you forsee any complications associated
  with a potential change of baseline to 40m-2?
  I presume that all that changes, besides for the
  CO ROC, is that last MMT. How about scattering, baffling, etc?

Any other comments or advice?

Thanks!  AJW

_____________________________________________________________
From k.strain@physics.gla.ac.uk Thu Apr 26 01:53:40 2001

Dear Alan,

I can't see a strong need to emulate the Adv LIGO arm cavity guoy phases
on the 40m (when it won't be right around the beamsplitter). So unless
someone else pushes you in that direction I would suggest you go for the
40m-1 case, which gives you more tolerance.

BTW you have the Adv LIGO mirrors a bit too curved, we were thinking more
of 54km roc to reduce thermoelastic noise effect with sapphire. (See
Primary requirements & design parameters in
http://www.ligo.caltech.edu/~ligo2/scripts/l2refdes.htm)

That is as helpful as I can be at the moment.

Cheers,
        Ken

_____________________________________________________________
From pf@ligo.mit.edu Thu Apr 26 08:11:28 2001

I agree with Ken. The alignment controls block diagram--the testing of
which (including the length controls) is the 40m's main goal--isn't going
to change with the different g parameters.

Peter

_____________________________________________________________
From jordan@ligo.caltech.edu Wed Apr 25 16:18:47 2001

Alan,

I think my initial response to this question was misguided. A g-factor of 0.8
shouldn't cause any major difficulties, if you and the experts decide this is
worth implementing. As we discussed, the ACIGA high power facility should be
able to explore the (complementary) stability vs. power questions.

Jordan

_____________________________________________________________
From romie_j@ligo.caltech.edu Thu Apr 26 10:55:36 2001

Dear Alan,
Might be moot by this point but my 2 cents worth is:

40m-1, 1ppm beam is not clipped
40m-2, 1ppm beam is clipped by 6.0 mm
40m-3, 1ppm beam is clipped by 0.6mm
40m-4, 1ppm beam is clipped by 7.7mm

The leg of the structure clips the incoming beam.
If you need me to look into this further (what portion of the beam is
clipped going to the SRM and the ETMx), just tell me.

Janeen

Yes, the beamsplitter is the worst, due to the 45 deg. incidence. When you
choose the beam size, tell me, so I can double check all incoming and
outgoing beams for clipping.

_____________________________________________________________
From kells@ligo.caltech.edu Thu Apr 26 10:37:55 2001

Hi Alan,

Directlly there would seem to be no important reason to mimic theAdv. Ligo
g factors. However I have two indirect consequence comments:

1. my guess is that be most significant "practical" consequence ofthe Adv ligo
arm cavity design is that the beam ~maximally fills the TM apertures (somewhat
more so than LIGO I ? I havent gone through the latest numbers !). Living with
a similar situation at the 40m might reveal a host of reality problems which will not
be seen with the tentative design parameters. For instance the "old" 40m had very
noticilbe HTM effects which we do not apparently see at LHO. Now I know that
you are not considering expanding the 40m beam THAT much. However one
could think in terms of "apodizing" the TM coatings (as was done with the 40m
recycling ITMs) to mimic the LIGO situation.


2. I have been doing extensive modeling of SBs in a LIGO-like ifo subject to
large distortions (ITM-thermal lens llike): will report on this in detail at the
upcoming MIT meeting. I do not fully understand the detailed behaviour (so I
cannot make any categorical statement WRT your question). OUr standard
view of the SB behaviour, within a regime where the SBs are "far" from res.
in both arms, is that the SBs are severely degraded by circulating in the ~degenerate,
and highly distorted RC, but that the arms play an insignificant role. However I
see ,TYPICALLY, in full FFT simulations that this is not so. The SB recycling gain,
and the imbalance of SB+ vs SB-, are strongly influenced by fine tuning of the
arms (eg micro arm length, or equivalently fSB by  <few kHz, and presumably
by the arm g factors). Since the FFT code conveniently allows fSB to be
adjusted arbitrarily, a "good" fine adjustment can always be invoked. However,
with a MC configuration fSB is fixed, and I therefore imagine that a careful
choice of g factors could be an important decision. Of course a critical parameter
in all this is what the RC distortions actually are. It is not so clear that a "good"
g factor can be rationally determined (at least on a finite time scale), but fully
understanding the phenomenon may indicate "bands" to be avoided, etc.

See you in May,

Bill

__________
From ajw@heppcs0.cithep.caltech.edu Thu Apr 26 14:41:06 2001

Hi all,

Thanks to Ken and Peter for weighing in
that blowing up the beams at the 40m is not an
important test for the 40m, even for the
alignment system.

I guess I agree, but I haven't
thought deeply enough about the differences
in the alignment of 40m and 4km IFOs
to really know what I'm talking about.

Thanks to Bill for his comments.

Response to Bill's point 1:
the blown up AdvL beams
are not huge blow-ups; and the mirrors are much bigger
than ligo 1, already; I don't think they're
maximally filling the TM apertures;
I thought that the thin BS (at 45 degrees) aperture
is the limiting factor, and the BS aperture is
limited by the drumhead mode.

For the 40m, we're nowhere close to filling
the TM aperture; but for the BS, see below!
Also, we will definitely be "apodizing"
the 40m TM HR coatings in order to reduce
back-scattered light, as recommended by Mike Smith.

Bill's point 2:
I will be surprised if you tell me that
thermal effects are relevant at the 40m,
even with 1w of input power.
The idea is to stay clear of all thermal effects,
because the 40m can't be a good test of AdvL's thermal
load, anyway. It's not in the list of goals
of the 40m upgrade to investigate thermal distortions.
Save that for GinGin and LASTI.

___
Janeen points out that with a blown-up beam at the 40m,
the 1ppm beam is clipped by 6.0 mm at the BS.
This seems unacceptable to me.
Short of making a bigger BS, I take this as the best reason
to NOT blow up the beams at the 40m.


Please let me know if you think I'm not making
the right decision, or for the wrong reasons...

Thanks!  AJW