Late turn after GPS waypoint

Before we installed the GTN650 we had the KLN94 and KAP140 – without the roll steering.
We need to align the heading indicator, set the heading bug and then set the cdi correctly to run the Autopilot in NAV mode. Still so with the GTN650.
One pilot told me he has experience with a newer machine where – with the same avionics – he was able to set a course on the GPS and get the autopilot to track that without setting the heading indicator and CDI.
Has that anything todo with the roll steering connection?

Yes, the KLN94-KFC225 combination does offer an analogue form of roll steering.

The KLN94-KFC225 analog roll steering connection requires a KLN94 P/N 069-01034-0102 (which can be done as a software upgrade from the 069-01034-0101 most commonly installed) and is only a few wires, shown at the top of here plus a config change in the KFC225.

I have only ever flown with one pilot who had this wired up, but he was not able to get it to do anything obvious. It was, I believe, installed by Air Touring who despite being CAA Level 2 approved could not even install a KI229 RMI – they had to have 2 or 3 goes at it. However there is no reason why it should not work; normal “dumb” steering (whereby the autopilot uses the HSI/EHSI course pointer as the heading, and uses the deviation bar to tweak the tracking) is good enough for everything I have ever seen including the T-shaped GPS approaches. And this is why I never bothered to wire it up… also in my case there is difficult access to get those particular wires in place. Also the KLN94 cannot fly any curved paths and has no holds or arcs in its database so again there was no point.

The wider issue with roll steering is that if used with a mechanical HSI, the HSI course pointer is totally ignored by the autopilot (it still functions as usual in VOR/LOC modes) which is not ideal for situational awareness which is compromised by having a “dead” HSI course pointer, which makes GPSS in conjunction with a mechanical HSI a slightly dodgy idea…

If I got a penny every time I met a pilot with a GNS or GTN box who is less than clear about what it does on an ILS, I would have enough for a second TB20 Somehow, Garmin have managed to really confuse people with the terminology.

I thought the later KLN94 with KAP140 had roll steering:

-0102 includes with DC roll steering output for use with KAP-140 and KFC-225 autopilots
(click Here for diagram)

http://www.sea-avionics.com/lc/cart.php?target=productDetails&model=KLN-94+System+2&substring=kln94

I find it hard to get my head around how this works… do you mean NCYankee that, on an ILS, the Garmins will fly using GPS LNAV till the FAF and only then use the LOC signal?

That would produce very poor guidance.

In my plane, I press APR when approaching the LOC so the system is tracking the LOC from much earlier. There is zero “snaking” although it can take a while to take out a strong wind drift.

Roll steering on the GNS is only available on the ARINC 429 channel. It is a bank angle to fly and does not have any direct relationship to the analog CDI L/R signal. With the KFC225, NAV mode does not use CDI deviation for tracking a course when GPS is the navigation source as long as roll steering data is available. IOW it simply flies the bank angle commanded by the roll steering. This permits flying curved paths such as a DME arc, PT, Hold or turn anticipation where as long as the bank angle does not exceed the auto[pilot limits, it will roll out on the new course on center, because the bank angle needed is dynamically changes as needed.

In the case of the GNS/W or GTN series, the roll steering label is transmitted on a 100 MS basis with the bank angle it is asking the autopilot to fly. When on an ILS, the GNS/W or GTN stops transmitting a bank angle on reaching the FAF. For most autopilots using a GPSS adapter, the adapter will output a wings level (no heading bug error voltage). Interestingly enough, the legacy units would continue to output the roll steering bank angle all the way to the MAP. This behavior exists only on ILS or localizer procedures and not on VOR procedures. There are two cases for how the KFC225 would track on an ILS, case 1, VLOC is selected on the GNS/W or GTN. Case 2, GPS is selected on the GNS/W or GTN. In either case, the CDI deviations will guide the autopilot, but in the latter case, there will not be a GS on the HSI and the GS won’t be tracked.

The KFC225 does a good job on the ILS in no or low cross wind conditions, but it will S turn if the winds change close in to the runway like I described. I have flown enough time with this autopilot to verify this, although in general it does a better job than most adjusting to an updated wind correction angle.

On the G1000/KAP140 combo I fly (DA40), when flying a full procedure, the CDI changes from GPS to VLOC when turning inbound on the procedure turn.

The KAP140 does not have roll steering inputs to NAV.

Ok, but then the G1000 has work some magic, because in the G1000+KAP140 combination it behaves like it does.

Re the KAP140 roll steering, has there ever been a roll steering converter which instead of faking a heading bug it fakes the course pointer + deviation bar signals, allowing the AP to remain in NAV mode?

Well, I only know that with the KAP140 in NAV mode it will follow the G1000 flight plan, including turn anticipation. It will also do holds and racetracks.

In principle it could be done. Basically you would fake the outputs of a KI525 HSI.

The HSI is of course done entirely with software in an EFIS/FMS system like the G1000, so that should be a simple matter.

I find it hard to get my head around how this works… do you mean NCYankee that, on an ILS, the Garmins will fly using GPS LNAV till the FAF and only then use the LOC signal?

That would produce very poor guidance.

In my plane, I press APR when approaching the LOC so the system is tracking the LOC from much earlier. There is zero “snaking” although it can take a while to take out a strong wind drift.

Re the KAP140 roll steering, has there ever been a roll steering converter which instead of faking a heading bug it fakes the course pointer + deviation bar signals, allowing the AP to remain in NAV mode? In principle it could be done. Basically you would fake the outputs of a KI525 HSI.

The KAP140 does not have roll steering inputs to NAV.

NAV or APR mode on the GFC700 are dependent on the selected source for navigation. With VOR or Localizer as the source, the tracking logic and operation is old school, selected heading sets the intercept angle and selected course selects the tracking course. CDI deflections left or right and the rate of change are used to track the ILS or VOR. When GPS is the selected course, roll steering is used for the course guidance. This can be demonstrated on a day when there is a crosswind and the runway is served by both an ILS and an LPV approach along the identical path. Typically the winds at GS intercept altitude are stronger and 90 degrees to the surface winds, and most of this change occurs in the last 500 feet. The autopilot tracks the final approach course to the FAF and then starts the descent when the GS is intercepted. By the time the aircraft reaches the FAF, the wind correction has been established and the autopilot reduces its tracking sensitivity. As you descend, the direction and velocity of the cross wind changes and the autopilot usually gets behind the change and has to S turn back to reestablish the WCA. Often near 500 feet AGL, I find myself taking over manually under these circumstances. The GFC700 exibits this S turn tendency with the best of autopilots. I then fly the LPV. It is a thing of beauty. The track holds steady the entire way without a hint of an S turn as the heading is continuously adjusted by the roll steering to maintain a constant track to the runway. The CDI remains stuck in the center and never moves off center, but the heading smoothly and continuously adjusts to hold a constant track. Under identical conditions, I have had to take over for the ILS and could let the LPV continue without touching the controls until just before the flare.

In the case of the Garmin GNS430W/530W or GTN navigators, the GPS units will output label 121 on an ILS up until the FAF and then they will suspend sending the roll steering label. This causes an aftermarket GPSS to command wings level from that point as it is no longer receiving the label it needs. For this reason, one can’t use roll steering on the final approach leg of an ILS. Some installations also disable the roll steering anytime the navigation source is changed from GPS to VLOC. Many autopilots could take advantage of the roll steering to provide better lateral navigation on an ILS, but would have to fly the vertical GS manually. The G500/600 enforces the source must be GPS in order for it to supply GPSS, the Aspen does not and neither will some of the after market roll steering adapters.

Is this “GPS to ILS” issue the one concerning the automatic detection of the “final” and “real” localiser? That issue is difficult to solve automatically, because you can have fake localiser radiation.