Range rings are not accurate

[dtayls]

'Range rings' do not match the data returned by the 'measure tool'.

At 38S I am seeing errors of around 2% or more.

[nohal]

I'm not able to reproduce your problem. Can you please post a screenshot like this (In my case, I have 5 range rings 1 mile apart at the same latitude as you and everything seems just fine)?

[dtayls]

Try a single range ring at 100nm. That should make the 2nm+ error I mentioned quite clear.

So, why am I using such large range rings? I am using a sextant to accurately measure small angles (~6.5 degrees in this case) to test my skills. One of the objects is ~19nm away while the other is <1nm away. The center of the relevant circle is around 77nm away but using 3 range rings the errors in OpenCPN's calculations position me around 2nm away from my actual position.

[hoolie]

On my laptop at about 43N the difference between a 100nm ring and the range calculation is 0nm at 0deg, +2nm at 90deg and 270deg, and +3nm at 180deg.

The range calculation is (probably) rhumb line and the range circle is drawn on a Mercator projection. So at such a range there are bound to be discrepancies. To be accurate you need geodesic, or at least great circle, range calculations.

The facility is clearly supposed to be used at short ranges and I think you're trying to use it out of its normal context.

[rooiedirk]

I think this is due to the fact that the shown chart screen in OpenCPN is in fact a mercator projection. Drawing a (round) circle (based on pixels) will give offsets at larger ranges.
A solution could be, don't draw a circle, but calculate a polygone. Something like is done in the astro plugin, however this will add a lot of calculations and drawing delays.

[dtayls]

"Range Circles" are probably not the right tool for my purpose but they are all that is available in OpenCPN as far as I can see.

Clearly, range circles need to be quick and easy to compute particularly if there could be up to 10 for each point. But I need something that is accurate and repeatable. Is there another answer?