Accuracy of converting semicircles to degrees

  • Thread starter Thread starter Bruce Stemplewski
  • Start date Start date
B

Bruce Stemplewski

Garmin's formula for converting degrees to semicircles to degrees is:

d = s * (180.0D / 2^31)

When s = 119304640

d = 9.9999994039536

But 10 was entered on the GPS. Does anyone know of a more accurate
formula for converting semicircles to degrees?
 
Bruce said:
Garmin's formula for converting degrees to semicircles to degrees
is:

d = s * (180.0D / 2^31)

When s = 119304640

d = 9.9999994039536

But 10 was entered on the GPS. Does anyone know of a more
accurate formula for converting semicircles to degrees?

The difference lies not in the formula, but in the representation of
an angle by 31 bits (plus sign) in the computer. Here the difference
is less than one millionth of one degree.
 
Garmin's formula for converting degrees to semicircles to degrees is:

d = s * (180.0D / 2^31)

When s = 119304640

d = 9.9999994039536

But 10 was entered on the GPS. Does anyone know of a more accurate
formula for converting semicircles to degrees?

d = s * (180.0D / 2^31)+0.000002
When s = 119304640 and D=1 then d = 10.000001403953552525650000000000

Does that look better?

This reminds me of an operator that was continually writing up a
system that was supposed to maintain 1500 pounds and the digital meter
kept reading 1499. Tweak the meter to read 1501 and all the complaints
stopped. How important is the difference other than it doesn't look
right when there are a string of 9's rather than a string of 0's?

You should never expect more accuracy from an answer to a formula than
the accuracy of the input. Therefore your d should be no more than 9
significant digits or 9.99999940. If we consider the 180.0 to really
be 180 and an infinite number of zeros then the real max accuracy of
the formula would be limited by 2^31 or again 9 digits. 9 digit
accuracy is VERY good. It is the difference between 30 years and 30
years plus 1 second. Do you really need more than that or is it just a
string of 9's isn't as personally satisfying as a string of 0's??
 
Dale said:
accuracy is VERY good. It is the difference between 30 years and 30
years plus 1 second.

Or almost TWOmm in the distance from LAX to JFK. How can we possibly
condone such GROSS inacuracy in our hand-held GPSRs? ;-)

Ted
 
Or almost TWOmm in the distance from LAX to JFK. How can we possibly
condone such GROSS inacuracy in our hand-held GPSRs? ;-)

Ted


I corrected it by taking a millimeter off each end of my handheld with a
dremel tool.

--Gene
 

Members online

Back
Top