But DME accuracy clearly is dependent on range, namely +-0.25NM+1.25% (of the range) (says Pans Ops / ICAO Doc 8168)
But that ICAO recommended DME accuracy might just determine the maximum declared range of the radar, it need not mean that the accuracy degrades with range in a linear way.
Presumably range accuracy depends on accurately measuring when the echo returns.
Surely this should be very accurate until the range-weakened return signal starts to get drowned in noise, which must tend to happen near a maximum range.
I’m no expert on this! If anybody knows how it works, I would be interested to know. Is there an electrical engineer on the forum? 
Peter, the maximum deviation for an airfield radar (less than 63nm range) is 300m. Consequently, the azimuth error at the extremities of range will be less than 0.15deg in order to meet the 300m requirement.
DavidS, the requirements for DME can be found at CAP 670 (for ease, go to page 429 onwards in your PDF viewer). In simple terms, the signal strength at receiver should not be less than –90 dBm within the DOC. My experience, backed-up by some very good calibration equipment, is that DME accuracy does not diminish with range.
@Dave_Phillips,
Thanks!
the maximum deviation for an airfield radar (less than 63nm range) is 300m
300m absolute accuracy at 63nm radius, laterally (i.e. in azimuth) – I am extremely impressed.
How do they check it? Does somebody fly directly inbound, using GPS?
Yes. Come to East Anglia in a couple of weeks when we’re doing just that. We will fly various radials/heights recording our position utilising DGPS. Our equipment accuracy (i.e. knowing where we are) is in the order of millimetres although the pilot is only required to achieve about +/-25ft or so in all directions as a measurement of uncertainty factor can be applied The data collected is then compared against that recorded at site.
The same system and principles can apply an extreme level of accuracy on equipment such as ILS. I’m relatively new to the job but am stunned by the granularity we can achieve.
Presumably range accuracy depends on accurately measuring when the echo returns. Surely this should be very accurate until the range-weakened return signal starts to get drowned in noise, which must tend to happen near a maximum range.
This is true for free space propagation in vacuum. We seldom operate DME’s in the vacuum, however. Even though there’s little water (vapour) in the atmosphere, it can still have an effect due to the high dielectric constant of water (otherwise, we wouldn’t see rain drops and weather radar would not work). It causes radio waves to bend, which means the distance travelled by the wave will be greater than the straight line distance. Also, the presence of inversion layers may cause ducting. Rain clouds may scatter radio waves, as well as local scatterers, which cause the pulses to widen. I haven’t found a reference with good quantitative data, but I’d still be cautious about stating that there’s no distance effect on DME error. Also, the FAA AIM even gives 3% of the distance as error.
ICAO Doc 8071, Vol1 Ch3 is the source document. Clearly the USA (or any other state) may differ from these SARPs but one would then expect to see such a difference filed in the appropriate AIP.
tomjnx, the atmosphere must have some effect, but I thought the DME L-band frequency was chosen to minimise that, or at least the weather-related stuff.
I found this Link this morning, which seems pretty well written (so far as I can tell :-), and describes the effects you mention.
Warning, some physics required!
I thought the DME L-band frequency was chosen to minimise that, or at least the weather-related stuff.
Well, people seem to be using GPS signal bending to measure atmospheric water content.
acp_5_1665_2005_pdf
I’ve personally witnessed ducting on 1.2GHz (i.e. close to the DME band at ~1GHz) which went on for hours. I made more than 100 contacts to the south of england from eastern switzerland from an altitude of about 8200ft, where you would expect a maximum range of around 120NM, yet the signals traveled more than 400NM. So they must have been heavily bent.
It appears that nobody has yet come up with any examples of enforcements of such “point” crossings?
Now let’s look at it differently. Say P2 was not there, so you just flew past the corner of P1, just touching it. Would that be enforced? I don’t think so.
