I have found there are – with apologies to Clint Eastwood again – two kinds of GPSs: those that return the right altitude, and those that don’t.
And they don’t seem to be related to how old, how recent, or how expensive.
My ancient Samsung T705 tablet is within 10ft – as I rediscover every time I use the ex Oziexplorer track to do flying movie subtitles. Most other devices are 100-200ft off.
Peter, you can compare the performance of SPS service and WAAS service by looking at the respective PAN reports at this site for WAAS: https://www.nstb.tc.faa.gov/index.htm
Accuracy of non IFR FAA certified devices are marketing numbers.
I had a look at that 200 page PDF but can’t easily find a SBAS / non SBAS comparison.
I am sure much marketing data is bogus, but I would still expect a consumer GPS to do a much better job with SBAS. I am sure the technology is the same as in IFR units. Plus, the consumer units support many more satellites, while avionics ones support only NAVSTAR.
And the latest devices like the NEO-M9N (which I am currently playing with) has anti-spoofing and anti-jamming features which I have not heard of previously, relative to any non-mil GPS. I am actually using this module
and you can see the price – $39 
Against the developments in this area, we have most certified units having been designed 10+ years ago. Going back further, to the KLN94 (late 1990s, although the KLN89 is probably the same) which used what was back then a standard consumer GPS module.
As a data point, I have the GPS running with the antenna indoors, on a shelf under a window, picking up 10/11 satellites (can’t tell you which ones right now – another story) and the position is varying around (this is the minutes part of the latitude; I am not posting the lat/long of my house :) ) 53.59564 and 53.59615 which, based on 1 minute = 1nm, is a variation of 1 metre. This is a module which (100+) is £10. It will be interesting if it is using SBAS. HDOP VDOP & TDOP are 1.31, 1.97 and 1.46 (not sure of the units).
This shows HDOP VDOP TDOP just over 1m – the values after the double comma
Indoor antenna – an old Garmin one from 1999
The altitude, at 120m, doesn’t seem quite right but I can’t tell for sure. Google Maps says 70m here. I suppose I could drive the whole rig to the beach
My spare KEA130A, when set to EGKK QNH of 1009, shows 250ft which is 80m.
However, I have already worked out that SBAS doesn’t give you any geoid correction.
The error, 160ft, is exactly what most consumer GPSs deliver unless you tick the box “geoid correction”.
I wonder if @AF is around?
However, I have already worked out that SBAS doesn’t give you any geoid correction.
What about the reliability of data?
It is tracking 12 satellites (the 12 after the 1.21) and shows a 3D fix (D3).
The variation I am seeing in the coordinates is 0.0005 of a minute which works out to 1m which roughly corresponds to the HDOP.
I haven’t yet sussed out how to detect SBAS usage. I am extracting three sentences and one of them is binary data.
The basic stratux GPS is not that good with SBAS…
Peter wrote:
However, I have already worked out that SBAS doesn’t give you any geoid correction.
Geoid has no use for the internal GPS calculations, everything is expressed in terms of the WGS84 spheroid or the geometric altitude. For example, terrain and obstacles are expressed in the DB in terms of height above the spheroid or their geometric altitude, LPV GP is expressed in terms of geometric altitude. ADS-B Out GPS altitude is expressed in terms of geometric altitude. Only time that geoid is applied is when an altitude is displayed to the pilot and is often termed GPS altitude or GSL altitude.
OK; every day is a school day
So to display the correct height AGL one needs to do the WGS84 correction. IFR navigators do it with a lookup table AFAIK, but there is also a polynomial surface method which works if you restrict the latitude a bit.
We did this here and there was another thread I can’t find right now.
Luckily in my current project I have no need to display height AGL. And there must be public domain correction methods out there.
@greg_mp it’s funny to see they are using UBX also – they are using a similar U-BLOX module to what I am using, although it will be an older version.
Mine is their very latest one, and unobtainable right now except in a development board form like I have. UBX is a horrible binary protocol; I have been writing the code to parse the stuff for the last few days…
I’ve been getting 1-2m “precision” for a few days now, indoor antenna.
I found this WGS84 to AGL converter which is spot on.
They also refer to the data they use. The 15 minute data isn’t small, so I reckon IFR boxes like the old KLN94, or the more recent Garmin handhelds, don’t use the same resolution.
