Re: Laser ranging to moon begs questions


John C. Polasek <jpolasek@xxxxxxxxxx> writes:
On 27 Oct 2007 15:38:31 -0400, Craig Markwardt
<craigmnet@xxxxxxxxxxxxxxxxxxxxxxxxxx> wrote:


John C. Polasek <jpolasek@xxxxxxxxxx> writes:

On 26 Oct 2007 09:53:01 -0400, Craig Markwardt
<craigmnet@xxxxxxxxxxxxxxxxxxxxxxxxxx> wrote:



John C. Polasek <jpolasek@xxxxxxxxxx> writes:
...

Phorate = N/Pwide = 2e28 ph/sec 1 pulse during pulse.
This is photons per second during the pulse.

True. But, as noted previously, you failed to multiply by pulse width
when you computed total number of photons received per second (which
is averaged over pulse and non-pulse time intervals).
I used photon rate while the pulse is on. That's when the detectins
must be done. When the pulse is over, the experiment is over in a
sense. Why would anyone care about averaging over all time? As you
note below, the narrower the better.

The experimenter cares about averaging over the total experiment time,
because that determines how many total photons his detector will
receive in a given wall clock time. The experimenter does not live
his life in 200 ps pulses, and the observatory is not available in 200
ps pulses.

I think you know that's not right. ...

Wait, so you think the experimenter lives his live in 200 ps pulses?
or that the observatory is available in 200 ps pulses? That is
ludicrous.
What kind of talk is that? See below. Electronic returns are easily
and routinely, and in this case, mandatorily gated; there is no other
way that you can assert that a return is from the original pulse. That
talk about "observatory available bla bla" sounds like babble. Pulse
gating was just ordinary radar electronics even back in the 60's.

You are confusing the "gating" issue with the number of photons
actually returned during the experiment. "Observatory availability"
is hardly babble. Consider that most observatories have multiple
projects, so a single project will be assigned a limited amount of
observing time.

Furthermore, since returned photons are grouped/averaged into "normal
points" of duration ~10-15 minutes (wall clock time) before delivery
to archives for science analysis, the number of photons per unit wall
clock time is relevant. While this discussion has focussed on the
McDonald Observatory, which typically receives ~10 returned phtons per
normal point, the OCA observatory typically receives ~100 photons per
point, and Apache Point many thousands (Shelus et al., Murphy et al.).

The experiment is ultimately interested in how many return
pulses are received per unit of wall clock time.

... It is mandatory to include a
(sliding) gate that is only differentially wider than 200 ps to accept
none but valid returns, and quite simply, since the timing of the
return is known a' priori. The experiment would fail immediately if
the receiver were open to receive noise for the 10^11 picoseconds
between pulses, (prf = 10).

That is true. "Averaging over the total experiment time" means
multiplying your (ph/s/pulse) by the pulse width (s/pulse) and by the
pulse rate (pulse/s), to ultimately get returned pulses per second.

Why would I, with a duty cycle of 2e-9, or two parts per billion
would I want to tally all the idle time, which amounts to 500,000,000
empty pulses. The "vigor" of the reception improves with pulses per
second of ON time.

What "you" would want to do is irrelevant. What the experimenter
would want to do is be sure a significant number of photons is
returned within an observation of, say, ten minutes of wall clock
time, which is the typical "normal point" observation duration.

picoseconds per pulse. The result is an absolutely miniscule number
(3e-8 ph/sec).
Your "absolutely miniscule number" is also your number. We are not
counting cosmic ray events which would be random.
We have an ideal experiment where the next event is known to within
picoseconds and all we need to measure is a small amount of "stagger"
and a means for gating out the noise, is it any wonder you get a
miniscule number if you compute results in EST?

The 3e-8 number is found by taking *your* numbers, and estimating the
number of returns per unit wall clock time. That is relevant because
a typical observation is ~tens of minutes of wall clock time.

However, your numbers underlying the 3e-8 figure were incorrect, as
detailed before.

I note that you did not respond.
I repeat, such a calculation would be of only obscure statistical
significance. Pulse over, game over.

Being incorrect by a factor of ~10^{-8} is of obscure statistical
significance?
You're not always wrong, just this time.

I note your non-sequitur.

Perhaps you should recalibrate your significance scale.
Repeat: you need a 200+ps gate that invalidates your e-8 argument. The
interpulse noise would devour your single photon.

Actually, I did account for that gate. Unfortunately I erased the
full context, which was,

: Your value (and units) are non-sensical. If you *really* mean 15.4
: photons per pulse *per second of laser on-time* -- per above, you
: described as "ph/sec ... during pulse" -- then to find photons per
: second, you should multiply by 10 pulses per second, times 200
: picoseconds per pulse. The result is an absolutely miniscule number
: (3e-8 ph/sec).

It was (a) your arithmetic error (leading to 15 ph/s/pulse), and (b)
*your* forgetting to multiply by the 200 ps pulse width or the pulse
repitition rate which lead to the ludicrously small number.

I note your lack of response.

...
Well you're right. I used 6 cm in an earlier thread so I don't know
how I used 60m this time.
I am getting (above) 111sec to the next pulse, so would it be fair to
say that by timing samples out of a 6cm pulse span one could narrow
the timing to .6cm with 100 samples, or to .06cm in 10,000 samples?

Probably not.
I'm trying to help here. It's not unreasonable to wait 10,000 samples
and thus cope with the paucity of information. There's no doubt that
something like that does happen, but we just don't have a verbato,
recital. This whole scenario has been fraught with approximations and
hypotheses.

If you were really trying to help, why didn't you try to read some of
the references that were cited?

One reference was mentioned. I do not have access to library journals.

That is a weak reply for several reasons. (a) More than one reference
was cited; (b) the originally mentioned Shelus paper is freely
available; (c) many municipal public libraries hold scientific
journals; (d) *Science* magazine articles are available for very
modest fee.


Why didn't you try to do some of yor
own (re)searching? Why didn't you consider any of the other losses
that were mentioned in the Dickey article?
There were offhand mentions of diffraction, etc. none of which is of
any value unless its an authenticated number.

I note that you did not respond to the question. Why didn't you
bother to do your own research?

Some of the answers to your questions are indeed in those articles.

I believe there are other systematic uncertainties
which begin to dominate. I found a description of many of these
uncertainties in Samain et al., 1998, Astronomy & Astrophysics, v.130,
p235.
It seems like you should have mentioned a few if they were pertinent.

It's a complicated subject. Would you like to be spoon-fed the
answers? Why not do some of your own research? Why not actually look
at the journal article (which is accessible to the public through NASA
ADS)? I suspect it's more convenient for you to sit back and take pot
shots at Peter or me, and picking and choosing whether to say too
little or too much information is available.

*You* started this thread by questioning whether the small number of
return photons would allow observers to determine the range so
precisely, even though the small number of returned photons does not
*necessarily* make the ranging inaccurate. [*]

The deeply ironic part is that after you made some ad hoc and totally
unsubstantiated estimates of the lunar ranging photon return rate, and
Peter and I attempted to inject some physics and some references to
the real observational issues, you are now criticizing *us* for not
providing enough information.

If you really care about the subject, why not get off of your petard
and find out some facts about the subject you are speculating upon?
Setting arbitrary boundaries about which journals you will or will not
look at is also rather silly. Scientific experiments and results are
published in scientific journals. If you don't look at scientific
journals, then you are pretty much relegated to uninformed observer
status.

And finally, while the discussion has focussed on the low quality
McDonald Observatory results, in fact there are other lunar laser
ranging observatories which detect many more photons such as the OCA
(see Samain ref above), and the Apache Point observtory (APOLLO,
Murphy et al 2007). The results from these observatories definitely
do not "beg the question" as your originally post asserted.

CM


References

Murphy et al. 2007, "APOLLO: Apache Point...", PASP, submitted,
arXiv:0710.0890v1

Shelus Ries Williams Dickey, "A Summary of LLR Activity and Science
Results,"
http://geodaf.mt.asi.it/html_old/news/iwlr/Shelus_et_al_LLR.pdf

[*] the more important question is whether return photons can be
distinguished from non-return photons.

.