 Check out the bookstore at IRBS.com |
|
|
|
|||||
|
||||||
From: Bill (no email)
Date: Mon May 01 2006 - 14:31:20 EDT
Alex (in blue)
> All other star-to star distances from 18d to 104d become almost perfect if
> one uses IC= -0.8
As I recall you do not apply the Freiberger arc errors to your observations
as they are so small inmost cases.
>
> This indicates that the arc is non-uniform on the interval from 0 to 4d,
> and it is OK after that.
> I would be satisfied with this conclusion if not the Freiberger
> (and Cassens/Plath) tests that say that the arc is good everywhere.
I have not seen the testing equipment, but it has always bothered me that
one could slap any sextant on a machine without some known presets and come
up with the correct measurements relative to that instrument. Probably my
problem, but it still bothers me. Perhaps Ken, with some experience in
"proofing" his sextants can address my concerns.
> As you know, I experienced the problem BOTH before and after
> Freiberger/CP tests.
That is what really bothered me. If it were some newbie that treated the
sextant like a toy, I would be inclined to ignore it. But I know you to
take great care and have a very good eye and hand (relative to mine ;-)
>
> The only thing that remains to finally solve the question
> is to lend the sextant to some really experienced guy with
> good eyes who would
> care to make star-to-star tests to confirm my observations:-)
Dog. <G> We know you generally have greater precision (lower sigma on
average) when we observe together, and have a better eye (more accurate),
but doing the look-see with scope A on your sextant (seeing if your
observations agreed with mine) put us in the same ballpark, despite my lower
batting average.
> But of course, it is a scandal if both Freiberger and CP tests
> are so bad...
Backing up to, "As you know, I experienced the problem BOTH before and after
Freiberger/CP tests." I did note only the Freiberger gave finite values,
the CP tests just put it within some limits. (Maybe plus/minus 9"?) We have
to start measuring at some reference point, and any other measurement is
relative to that. As I recall Freiberger said that 0d (IE) was spot on,
and used that as a baseline for other errors. I do not recall the CP IE.
Soon thereafter your tests showed a significant difference in IE (for you)
of less than 1' IE. Before the boat if I recall.
Again, I have faith in your abilities. You do not *knowingly* play slop
pool IMHO. I have always questioned in my mind whether Freiberger adjusted
index error to 0d, then started; or used the existing baseline and proceeded
from there.
Going back a few posts, I had the chance to sit and chat with one of
Purdue's better/best machinists (and five-time Club Champ) Friday night at
the sailing-club callout . Asked him about "proofing" the teeth on a
sextant. He does have an X-Y table with a microscope capable of seeing
1/1000 of a millimeter. Problem being it cannot see the arc on the sextant
as is--laying flat.
An old friend who teaches manufacturing technology at PU will be returning
from sabbatical soon. He has some cool state-of-the-art equipment. I will
put the problem to him.
As I know from personal experience and professionals (as stated a few posts
ago) static testing is just a starting point. As my machinist friend said
Friday, "I can chuck up an end mill, bring it down until it *just* kisses,
back it off 0.002", turn on the milling machine, bring it down--and it
*bites* into the surface. Static vs. dynamic bearing."
We trying to measure 1/10th of 1/60th of 1/360th by eye, with at best
adequate optics (the eye and scope), and *handheld*. Often on a moving
platform, with an instrument advertised as plus/minus 20" to 9" of an arc.
By expanding the instrument range by 2X with mirrors, I reason any fault
with the worm gear or arc teeth will also be double.
Best way I know to put it is the old saw, "Don't blame poor craftsmanship on
a poor tool. If you doubt that, put a quality tool in the hands of a poor
craftsman." (The "old" definition of a "tool" is a handle on one end and a
use on the other.) Before you get your (former) USSR equivalent of
"knickers" get in a twist, I believe you to be an excellent craftsman.
What we are doing from the sextant perspective is not the pure science you
are trained in. A "tool" has idiosyncrasies. The better the tool, the
fewer the idiosyncrasies, But they all have them, and the craftsman has to
adjust to those idiosyncrasies to do better than the tool can do if
controlled by a machine without a constant feedback loop.
Talking with Ken Gebhart, it seems even the top of the line sextants have
their own little quirks. I suspect that even with the "best" units out there
someone with your talents will always be chasing a ghost in the machine ;-)
While I really like your concept of short star-to-star measurements on and
off the arc, I wonder how many of those one can do without starting to
"spoil" those spots on the arc and worm gear.
Bill
|