I can send you the Oziexplorer track logs of numerous ILS approaches and you can check if they are straight or not. They look totally straight to me, laterally at least.
The straightness (or lack of it) only really matters to an autopilot.
What I was getting at are pilot reports of the LPV being easier / more stable to fly.
I find hand flying a full procedural LPV much easier for the simple reason that the procedures are easier. Even when you don't fly the full [procedure (ie. under vectors), the LPV is easier because setting it up is quicker, and you have better situational awareness. On all approaches I select the overlay on the GPS and then dial and identify in the necessary frequencies. On an LPV this is one operation vs two.
Actually hand flying down the LPV final approach track is just the same as an ILS. LNAV/VNAV and LNAV+V on the other hand are an absolute pleasure to hand fly - the greater full scale deflection makes them much less twitchy.
Peter,
You are correct that the FSD is essentially the same and angular on an ILS and a LPV. The portion that is fixed is inside the DH. When the WAAS GPS is used to fly the LNAV/VNAV, LNAV+V, or LNAV, it uses the same +/- 2 degree angular FSD for most of the final segment, but it becomes a fixed FSD of +/- 45 meters. This means that with these approaches when flown with the WAAS GPS, the FSD becomes fixed at about 1.8 NM from the touchdown point. BTW, most pilots are unaware that the FSD is angular +/- 2 degrees (unless this value exceeds .3 NM in which case .3 NM is used) for the LNAV/VNAV, LNAV+V, and LNAV when flown with a WAAS GPS. If they are flown with a non WAAS GPS, the FSD is fixed at +/- .3 NM on the final segment.
On a fully integrated system with a digital autopilot, for example the G1000 with GFC700, the autopilot flies the LPV vastly better than the ILS, particularly in a crosswind. This is because the normal CDI deviation control mechanism is used for the ILS and it needs to work out the crosswind. Since the winds normally will change both direction and velocity as one descends on the final approach course, the ILS tracker has the nasty habit of losing the localizer at around 500 feet, drifts off and S turns to attempt to redetermine the wind correction. I have found myself disengaging the autopilot under these circumstances and hand flying. The GFC700 will S turn with the best of autopilots in such conditions. However, with the LPV, the GFC700 is able to utilize roll steering for the lateral guidance, so it continuously corrects for any cross wind to hold the track constant. I have flown back to back ILS/LPV at my home airport in a Cirrus SR22TN with the G1000 Perspective system to demonstrate the difference to my customer. It was a day when the winds at 2500 feet were a direct crosswind at about 30 Kts, dropping down to about a 10 knot crosswind of about 40 degrees on the runway. There was a fair amount of turbulence on final. I took over on the ILS at around 500 AGL as it had drifted too far off course and was aggressively S turning back to find it. With the LPV, the crab was continuously and smoothly adjusted by the autopilot and I kept it on the autopilot down to 30 feet AGL. It was a thing of beauty.
That's really interesting.
most pilots are unaware that the FSD is angular +/- 2 degrees (unless this value exceeds .3 NM in which case .3 NM is used) for the LNAV/VNAV, LNAV+V, and LNAV when flown with a WAAS GPS. If they are flown with a non WAAS GPS, the FSD is fixed at +/- .3 NM on the final segment.
(my emphasis) is news to me...
Incidentally, does a WAAS box have any manual control over the FSD? With my ancient KLN94, and I believe with a GNS (non-W) box, one can set the FSD to 5.0 1.0 or 0.3nm, and this is useful in other contexts e.g. using the OBS mode to fly an NDB or VOR approach for which there is no GPS overlay, or the overlay is so stripped down so as to be useless or confusing. On the '94 one can even change the FSD when flying a GPS approach; not that one would want to in normal circumstances...
The GFC700 will S turn with the best of autopilots in such conditions.
This suprises me... I am perhaps spoilt by the KFC225 which - when its servos are not smoking - tracks an ILS perfectly centred all the way to the runway. It didn't used to be that good when driven from the KN72 lateral output (for L) and from the KX155A GS output (for V) but now that it is driven from the EHSI "FCS" output it is a lot better.
However, with the LPV, the GFC700 is able to utilize roll steering for the lateral guidance
Why only on an LPV approach? The S-turns you describe are presumably only in the lateral plane so roll steering should be available when following any track; an approach or not.
You're ignoring the interference that the ILS suffers from, Peter. The ILS beam is not a straight beam like the LPV "beam".
I think what you are getting at is that the aircraft receiver moves the CDI (or glide slope) by interpreting the relative strength of the 150Hz and 90Hz lobes....therefore if there are any local variations in these lobes the "beam" may not be perfectly straight....even if it is shown on the IAP as such?
Certainly there are no curved ILS approaches.
I just haven't seen any "funny business" on the ILS - as distinct from NDB approaches which are a total joke in terms of the depicted track.
If you wired up an autopilot to fly an NDB approach (which isn't possible legitimately because only a nav source capable of a "CDI presentation" can drive an autopilot, but it would not be hard to do) the resulting track would be all over the place between approx. 5D and 2D.
A VOR approach should in theory be as accurate as a LOC, but there is a lot less (1/4?) gain in the control loop so the tracking is sloppy and any crosswind correction will be that much less effective. Anyway one flies those with a GPS, hand-set to 0.3nm FSD...
I don't see any fundamental reason why interpreting the relative strength of the 150Hz and 90Hz lobes should not be possible to do very accurately. Something like the SN3500 EHSI does it purely in software and does it with superb accuracy.
I also do accept that it is easy to have a defective ILS receiver and have poor tracking, or even an angular tracking offset so e.g. a LOC with an actual track of 200 degrees is actually flown at say 204 degrees. The 4 degree error looks absolutely massive on the way in but it takes you to the runway just the same. Whereas a GPS cannot have this sort of error if it is working at all.
Peter the difference I think you are asking about is between Autopilot ILS tracking which is CDI driven and GPS roll steering which provides the lateral guidance for an LPV. My understanding is that the GPS can track your actual vs desired track and provide much better commands to the autopilot. An ILS is tracked in a less sophisticated way ie I am 1 dot left of track so probably should go to the right a bit.
An ILS is tracked in a less sophisticated way ie I am 1 dot left of track so probably should go to the right a bit.
How is that different from GPS i.e. I am 1 mile left of the track?
Peter the difference I think you are asking about is between Autopilot ILS tracking which is CDI driven and GPS roll steering which provides the lateral guidance for an LPV. My understanding is that the GPS can track your actual vs desired track and provide much better commands to the autopilot. An ILS is tracked in a less sophisticated way ie I am 1 dot left of track so probably should go to the right a bit.
Yes, probably.
I think an ILS is tracked by the AP flying a heading which is initially taken from the course pointer setting (as with any enroute "NAV" mode using a plain HSI) and then continually adjusting the heading by tracking the L-R bar deviation.
That is exactly what you do when flying an ILS by hand.
This will be at-best imperfect because the loop gain cannot be infinite. But evidently it can be very much good enough.
What I've never tested is flying an ILS with a mis-set CP. It's obvious what this will do on a VOR approach (it will intercept and fly a different radial) but it probably doesn't do much on a LOC, except resulting in a bad initial heading following the intercept.
How is that different from GPS i.e. I am 1 mile left of the track?
It's clearly not different in normal GPS flying i.e. NAV mode. Even GPSS flying is not different because the ARINC429 data from the GPS contains the current angular error and the autopilot just uses that error to adjust the heading (i.e. this is the same sort of proportional control loop as you have with a LOC).
But if you have a computed wind correction (which implies an airdata computation i.e. knowing the wind aloft) then you are giving the proportional loop a lot less to do, and a higher lateral accuracy can be achieved with the same proportional gain, by the use of an integral term (this is pre-WW2 stuff) which would bring the error right down.
I don't think current LOC tracking uses any integral term, but I really have no idea. I am sure others have been here before...
What I've never tested is flying an ILS with a mis-set CP. It's obvious what this will do on a VOR approach (it will intercept and fly a different radial) but it probably doesn't do much on a LOC, except resulting in a bad initial heading following the intercept.
I have once when I missed that the CP autoslew failed. It screws it completely. It tries to make adjustments to the wrong heading and never captured in my case. Obviously I caught it and hand flew.
Peter,
I have flown a fair number of ILS with a KFC225 in an A36 Bonanza. On balance it is a wonderful autopilot when it is working, but my experience with it is that it has the same S turn characteristic when the cross wind component changes while close in and low on the ILS. The GFC700 has about the same performance when flying the ILS.
