On 2 Jan 2009, at 05:59, John Denker wrote: > On 01/01/2009 10:05 PM, syd adams wrote: > >> I think i assumed long ago that the GS deflection had a limit of >> -10 to 10 >> like the heading-needle-deflection , and so scaled the needle to the >> outermost dot accordingly. > > That is not consistent with what is implemented in navradio.cxx > > That instrument, and all the CDI instruments I've seen, assume > -3.5 to +3.5 for the glideslope signal. I have no clue as to > the history or rationale for this, but it is what it is. > >> The manual states that in GS mode the vertical scale is in >> degrees , and >> captures at 0.5 degrees , and I'm assuming (again) , that each dot >> must >> represent 1 degree, since I havent come across exact numbers yet.In >> other >> words, my setup could be wrong, I'll have to research this more.
Note to John, this is the part of the discussion I care more about right now, because the Mk-VIII defines various rules and alerts in terms of 'dots'. It does however specify (in the installation guide) a mapping from DDM to dots, using fixed values. DDM seems to be expressed as a percentage of the loc or GS arc, eg +/- 0.4 for the loc, and +/- 0.8 for the GS. However, those definitions bring us back to: >> In the Real World, the entire usable part of the glideslope beam > is only 0.7 degrees thick. Having a scale in degrees would not > be very useful. Important thing - all the instruments John linked to do indeed have 5 dots for the CDI, but all the glass displays, and high-end HSIs I've seen, have two. This is interesting - 5 dots * 2 degrees (as John said) gives ten degrees. The reference I found (which shows a 2-dot CDI) says 5 degrees. And 2 dots * 5 degrees is ..... ten! :) This bring us back to, we should probably be working in percentage of full deviation (i.e DDM) for all these things, and exposing degrees deviation as a bonus for other users. There's still the issue of what full deviation is for both - I've heard 15 degrees for LOC stated, which would mean +/- 0.4 DDM (which is what the Mk-VIII uses as limits) would be +/- 6 degrees, and they state 0.0775 of DDM per dot, so full range would be 5.161 dots. Which close enough to 5 :) For the GS, John states that the 'usable' part is only 0.7 degrees thick. I'm unsure what usable means in that context, but given then +/- 0.8 DDM range used by the Mk-VIII GPWS, and their dots-to-DMM factor of 0.0875 for the GS, maximum deviation is 9.142 **dots**. Clearly that's a much larger deviation that could ever be seen in a real world-approach. Give that definition, it seems unlike that the 'full deviation' (as opposed to usable) for the GS should be taken as 0.7 degrees, since that would make each dot 0.076 of a degree - which is an order of magnitude less than the 0.3 or 0.6 definitions we've been talking about. > I wrote the code to make localizer sensitivity proportional to > runway length, as it is in Real Life, in accordance with ICAO > standards. > > I've shot a few approaches with it. > > If you fly from KMJX 6R to KJFK 31L there is quite a contrast: > the latter is about two-and-a-half times more sensitive. > > I tied the sensitivity into the property tree so you can look > at it and even monkey with it. > > The code implements the ICAO standard in some detail, even to > the point of producing slightly odd results for very short > runways e.g. San Diego KMYF (Montgomery Field). So, I read the ICAO document, but I'm not quite familiar enough with the radio model here to follow this discussion - could you start again with the definitions, and walk me through it in baby-steps? ------------------------------------------------------------------------------ _______________________________________________ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
