Author

Topic: How do you check for errors?

bruce

posted April 12, 2020 08:34 PM
I'm running into (what I hope isn't) a 'personal error'.... sometimes I do everything right but make a simple/stupid error, and don't know it until I check the answers.
For example, tonight I was reducing a starsight example, and got what looked like a perfectly reasonable result, but when I looked at the answer I found that my "a" and Zn were correct, but my Alon was off.
Turns out I had subtracted instead of adding the supplied WE. which screwed up my LHA, which cascaded through to my GHA m.s, and resulted in using the wrong "minutespart" for my aLon.
In the real world, how would I know that I've made an error? Is it only by coming up with a nonsensical LOP in the resulting plot? Are there tricks to find errors before it gets that far?
I used to be a mathmajor, years ago (probably why I can't add ) and one of the things my professors drilled into us was to "work the problem backwards", plugging the answer back into the equation to see if you get the right "inputs" on the other side, as a way of checking the math. Is there an analogous technique that for CelNav?
Bruce
From: Everett, WA


bruce

posted April 19, 2020 04:22 PM
I went through my notes and coursework exercises yesterday, and distilled a set of reminders that  hopefully  will help me reduce the number of careless errors as I go forward.
Some of this is as simple as reminders to myself to make sure I get the right value (with the right sign) from the right row in the right column on the right page.
Parking this mostly here for my reference, but... if you find it useful, feel free to use/add/revise.
Sights  Check and record IC / Index Error EVERY TIME
Time/Date  Watch Time: if watch is fast, subtract WE; if watch is slow, ADD WE  If minutes > 60, subtract 60m and add 1 hour to time  If minutes < 0, add 60m and subtract 1 hour from time  If hours > 24, subtract 24h and add 1 day to date  If hours < 0, add 24h and subtract 1 day from date  ZD = Zone Description of the WATCH, not of the timezone you're in.
Corrections  Index correction: if "on the scale", take it off  SUBTRACT  DIP correction is always NEGATIVE  dcorr depends on Declination: if declination DECREASING, dcorr is NEGATIVE  dcorr/sun is always 1 or less  dcorr/planet can be larger than 1 because of motion relative to Earth  vcorr is ALMOST ALWAYS positive / watch sign at bottom of page with Venus  GHA m.s is always POSITIVE  Altitude correction/stars and planets is ALWAYS NEGATIVE  Planets: additional altcorr / Mars and Venus is small and ALWAYS POSITIVE  Moon: both corrections  altitude, HP  are ALWAYS positive
Reduction  GHA/star = GHA.h/Aries GHA.m.s/Aries+ SHA/star  Make sure aLon is within 30' of DRLon / use actual DRLon for Polaris)  LHA: if aLon = W, LHA = GHA  aLon; if aLon = E, LHA = GHA + aLon  LHA: should be a whole number of degrees if aLon selected correctly / add/sub 360 to adjust  NA/Daily pages: doublecheck page, date, hour, body/column  NA/Altitude Corrections: doublecheck table, body/limb/season, Harange / plus Venus/Mars  NA/Polaris: use LHA for entry to ALL THREE sections / a0, a1, a2  NA/Polaris: use minutes of HA to interpolate/refine a0, if needed  NA/Increments: doublecheck page/minutes, row/seconds/, column/body  NA/Moon: doublecheck table/035 or 3590, daterange, Harange, limb  NA/Moon: use Ha/minutes to interpolate/refine altitudecorrection, horizontal parallax if needed  249v3: doublecheck Latitude, name /same/contrary, Decrange/014 or 1529  249v3: doublecheck column/Dec; if Decmins>30, look at Z in next column  Table5: doublecheck column/d, row/Decmins, sign of dvalue for dcorr  Table5: use Decmins to interpolate between rowvalues
Plotting  Use midLAT to select LON scale in nomogram  Use 60degrees of LON on nomogram to place meridian lines  Doublecheck UPS setup before starting plot  Use LONscale/nomogram to set AP on midLAT line  Use LATscale/miles to set altitudeintercept on Znline  Be sure to clearly identify key points / APs, altitudeintercepts??  Be sure to clearly label Zenithlines, LOP lines  If advancing an LOP, use LATscale/miles to advance along DRcourse
Dumb stuff  Doublecheck ALL math for stupid/careless math errors  Lat by LAN: realitycheck latitude against DRLan : add, or subtract smaller from larger  LHA rule defines HOW to calculate Zn, but be sure to use Z  NOT LHA  to calculate Zn
Yeah, I know a lot of this is simplistic. But I'm doing this to rebuilt the rigor in my process, knowing that I won't have an answer section in the back of the book to check while on the boat.
And, yeah, I know I can (and probably will) use "computed solutions"  e.g., the StarPilot app  to get a realitycheck on my handcalc'd solutions, but... for now, I just want to eliminate as many dumb mistakes as I can, before they become "bad habits"....
$.02 Bruce
From: Everett, WA


