NavList:

I took a break from composing a blog post on chronometers to try out your method again, this time setting the sextants on a flat and fairly level surface and setting both frames horizontal to within a few minutes. I was then able to get repeatable results with a x6 telescope to within about 0.4 minutes using sextants whose errors summed to 1 second at 30 degrees and 3 seconds at 60 degrees. However, the individual calibration errors themselves had an uncertainty of maybe 3 seconds.

Routinely, I cannot get repeatablity of index error to better than about 0.3 degrees, but I must try determining index error with a x26 theodolite 'scope some day.

Your method seems well adapted to checking micrometer error, though as most makers do not quote it, it may be difficult to know which sextant has the error. Using an autocollimator I've just checked a 1978 SNO-T and the micrometer error does not exceed 2 seconds. It is undetectable at 8 points out of 12. A 1975 Plath sextant is nearly as good, with a maximum error of 2.8 seconds, not exceeding 1 second at 5 out of 12 points. A 1953 Plath had a maximum error of 7 seconds, still very good. This was the one with the bent frame that I wrote about on my web site (www.sextantbook.com). If you look at the blog in the "Chasing tenths of an arcminute" category, you will see that many sextants are much worse than this.

Bill Morris
Pukenui
New Zealand
----------------------------------------------------------------
NavList message boards and member settings: www.fer3.com/NavList
Members may optionally receive posts by email.
To cancel email delivery, send a message to NoMail[at]fer3.com
----------------------------------------------------------------