GPSS & Waypoints

[KLRDMD]

Installed Equipment: Avidyne 550, Aspen 2500 (three panel), S-Tec 60-2 Autopilot

GPSS has never done a full standard rate turn since I've owned this airplane. The adjustment is as far as it can be adjusted and it gets kinda close but not ideal. As a result, every time it overshoots each turn a bit. I’m looking at the S-Tec 3100 autopilot when it becomes available for the Baron which would solve that problem but I don’t think it is worth throwing money at an autopilot that I plan to replace soon.

I just flew a flight plan: 57AZ-CROME-KAVQ-57AZ and GPSS flew it. The autopilot in GPSS will intersect the leg at a 45º angle initially and in the first photo on the iPad you can see it doing that, I activated the autopilot after turning downwind from taking off RWY19 at 57AZ. The autopilot tracked directly to CROME, as it should have after intercepting the course. Notice the 550 shows a magenta line to CROME and a “barber pole” after CROME signifying the next leg. 

BUT . . .

Also notice the angle of the flight path on the 550 from 57AZ to CROME to KAVQ and that it is not curved as would be expected with GPSS. 

After CROME, the airplane continued straight ahead for a while so I just let it do it’s thing to see what would happen. It made a right 270º turn to re-intercept at CROME and then went merrily along to KAVQ. Notice on the 550, at KAVQ the flight path is curved as would be expected with GPSS and in fact the flight path was curved.

 I guess CROME must a “fly over” waypoint where KAVQ is a "fly by” waypoint. Interesting. I don’t know if you can force a fly over to be a fly by waypoint or not.

[Marauder]

Installed Equipment: Avidyne 550, Aspen 2500 (three panel), S-Tec 60-2 Autopilot

GPSS has never done a full standard rate turn since I've owned this airplane. The adjustment is as far as it can be adjusted and it gets kinda close but not ideal. As a result, every time it overshoots each turn a bit. I’m looking at the S-Tec 3100 autopilot when it becomes available for the Baron which would solve that problem but I don’t think it is worth throwing money at an autopilot that I plan to replace soon.

I just flew a flight plan: 57AZ-CROME-KAVQ-57AZ and GPSS flew it. The autopilot in GPSS will intersect the leg at a 45º angle initially and in the first photo on the iPad you can see it doing that, I activated the autopilot after turning downwind from taking off RWY19 at 57AZ. The autopilot tracked directly to CROME, as it should have after intercepting the course. Notice the 550 shows a magenta line to CROME and a “barber pole” after CROME signifying the next leg. 

BUT . . .

Also notice the angle of the flight path on the 550 from 57AZ to CROME to KAVQ and that it is not curved as would be expected with GPSS. 

After CROME, the airplane continued straight ahead for a while so I just let it do it’s thing to see what would happen. It made a right 270º turn to re-intercept at CROME and then went merrily along to KAVQ. Notice on the 550, at KAVQ the flight path is curved as would be expected with GPSS and in fact the flight path was curved.

 I guess CROME must a “fly over” waypoint where KAVQ is a "fly by” waypoint. Interesting. I don’t know if you can force a fly over to be a fly by waypoint or not.

I have an Aspen setup with the 60-2 and a GTN driving it. One thing I have noticed is that the STEC as an analog AP is dependent on input for winds aloft to make the correct guess for when to begin the turn. 

Here are two approaches (apologize for the busy track log). The turn at Certs intersection was done completely with the AP. The track shown between Certs and Efeco was hand flown. Just like the AP I guess wrong too. 

 

 

 

[Awful_Charlie]

Not that it's going to be much consolation to you I'm afraid, but with IFD540, Aspen PFD1000 and KFC150 I don't get this problem, even with turns >120 deg.

Hate to ask you to suck eggs, but you are running with GPSS enabled, and not just via NAV?  The turning behaviour could be explained at a stretch perhaps if you were in NAV mode (on my installation this would mean autopilot in HDG mode and GPSS indicated on the Aspen)

Will the S-TEC do rate 1 turns (or very nearly) with the heading bug or in NAV mode? If not it might point towards servo limit switches of rigging tensions (I know nothing about the S-TEC installation), if it does than maybe you can up the GPSS gain in the Aspen? (but where is the S-TEC getting the GPSS from? Aspen or direct from the IFD?) Do you have a flight director, and if so, does that show a demand for more bank in the turns than the autopilot executes?

It does seem rather odd that your flight plan shows the immediate turns - normally all waypoints in the flight plan are "fly by" unless they are coded by Jepp as "fly over" as part of a procedure, did you have anything "odd" in the flight plan (eg intercept a radial). You could try this in the simulator to see what that does, and either way feed it back to Avidyne (on their forum, or to tech support - I've generally found them responsive on both)

[KLRDMD]

It was in HDG mode, not NAV. The autopilot will not do standard rate turns under any circumstances. The gain is adjusted to the maximum possible value in the Aspen, twice normal and it brought it from less than half standard rate to more than three quarters. The GPSS should be coming from the Aspen, I would think. The FD shows as you would expect for the turn being done. The flight plan in the 550 was 57AZ-CROME-KAVQ-57AZ.

[Marauder]

Just now, KLRDMD said:

It was in HDG mode, not NAV. The autopilot will not do standard rate turns under any circumstances. The gain is adjusted to the maximum possible value in the Aspen, twice normal and it brought it from less than half standard rate to more than three quarters. The GPSS should be coming from the Aspen, I would think. The FD shows as you would expect for the turn being done. The flight plan in the 550 was 57AZ-CROME-KAVQ-57AZ.

When you do turns without the GPSS (just using the heading bug) does it do standard rate turns? 

[KLRDMD]

Just now, Marauder said:

When you do turns without the GPSS (just using the heading bug) does it do standard rate turns? 

No

[Marauder]

2 hours ago, Awful_Charlie said:

Not that it's going to be much consolation to you I'm afraid, but with IFD540, Aspen PFD1000 and KFC150 I don't get this problem, even with turns >120 deg.

Hate to ask you to suck eggs, but you are running with GPSS enabled, and not just via NAV?  The turning behaviour could be explained at a stretch perhaps if you were in NAV mode (on my installation this would mean autopilot in HDG mode and GPSS indicated on the Aspen)

Will the S-TEC do rate 1 turns (or very nearly) with the heading bug or in NAV mode? If not it might point towards servo limit switches of rigging tensions (I know nothing about the S-TEC installation), if it does than maybe you can up the GPSS gain in the Aspen? (but where is the S-TEC getting the GPSS from? Aspen or direct from the IFD?) Do you have a flight director, and if so, does that show a demand for more bank in the turns than the autopilot executes?

It does seem rather odd that your flight plan shows the immediate turns - normally all waypoints in the flight plan are "fly by" unless they are coded by Jepp as "fly over" as part of a procedure, did you have anything "odd" in the flight plan (eg intercept a radial). You could try this in the simulator to see what that does, and either way feed it back to Avidyne (on their forum, or to tech support - I've generally found them responsive on both)

The Aspen has GPSS built into it, so I am guessing that is where he gets it. 

[Marauder]

Just now, KLRDMD said:

No

Mine will. I am going out tomorrow to shoot some approaches and will bring my GoPro with me and make a video of an approach fully coupled. I often wondered how the intercept angle feature worked (or not work) with GPSS. Before I had the Aspens, I could do an intercept using this method (pushing both HDG and NAV at the same time). It would always do a standard rate turn at the right time.

[KLRDMD]

Just now, Marauder said:

Mine will. I am going out tomorrow to shoot some approaches and will bring my GoPro with me and make a video of an approach fully coupled. I often wondered how the intercept angle feature worked (or not work) with GPSS. Before I had the Aspens, I could do an intercept using this method (pushing both HDG and NAV at the same time). It would always do a standard rate turn at the right time.

I know it is in the autopilot and that's fine. That wasn't the important message of my original post. The "fly over" versus "fly by" aspect is the part I didn't expect.

[Marauder]

3 minutes ago, KLRDMD said:

I know it is in the autopilot and that's fine. That wasn't the important message of my original post. The "fly over" versus "fly by" aspect is the part I didn't expect.

Ken -- maybe I am not following your thread correctly. I read it over again and saw you mentioned the 270 degree loop. I have never had my 60-2 using GPSS do that turn for either a waypoint that is part of an approach or just a waypoint in space. If I were using my GTN, Aspen and 60-2 to fly what you shown, what I would see is a curve on the GTN and just sharp angles on the Aspen. It will follow what the GTN is depicting. The autopilot will always begin the turn before the waypoint and it will be standard rate. If the turn is sharp like your photos, the turn starts really early. I'll shoot a video of this tomorrow.

[KLRDMD]

10 minutes ago, Marauder said:

Ken -- maybe I am not following your thread correctly. I read it over again and saw you mentioned the 270 degree loop. I have never had my 60-2 using GPSS do that turn for either a waypoint that is part of an approach or just a waypoint in space. If I were using my GTN, Aspen and 60-2 to fly what you shown, what I would see is a curve on the GTN and just sharp angles on the Aspen. It will follow what the GTN is depicting. The autopilot will always begin the turn before the waypoint and it will be standard rate. If the turn is sharp like your photos, the turn starts really early. I'll shoot a video of this tomorrow.

I'm curious to see what your setup would depict if you put the same flight plan into your equipment tomorrow. 57AZ-CROME-KAVQ-57AZ. Mine depicts a sharp, acute angle at CROME but a smooth lead in turn at KAVQ. The smooth lead-in turn at KAVQ is typical for GPSS and what I regularly see. I'm not concerned about the less than perfect standard rate turns and minor under/over corrections. The lack of such a lead-in turn depiction at CROME has me wondering if that is a "fly over" waypoint that cannot have GPSS make the turn smoothly. I wouldn't expect it to be at that location, but who knows ? My autopilot has never done this 270º turn before either so I'm curious what another setup would show with similar but different equipment.

[carusoam]

My older generation BK 150 AP likes Standard rate turns...

The KLN90B will anticipate the turn early and give the pilot a warning signal to begin the turn...

Pilot then rotates the heading knob or course selector on the HSI to match what the KLN90B request...

 

Ken,

It seems a bit odd to not turn at a standard rate if able...

Best regards,

-a-

[KLRDMD]

CROME is an intersection (fix) so it is neither a fly by nor fly over waypoint. It is not a waypoint at all. Left = NACO, Right = Jeppesen

Hmmm . . . .

[KLRDMD]

I may be getting this figured out. CROME is an intersection, a fix on an airway defined by VORs. Fixes, by definition are fly over, you do not make the turn to the next heading when on an airway until flying over that fix. Maybe next time I'll use PICLI instead of CROME as it is not a fix defined by VORs and should be fly by.

[Marauder]

I may be getting this figured out. CROME is an intersection, a fix on an airway defined by VORs. Fixes, by definition are fly over, you do not make the turn to the next heading when on an airway until flying over that fix. Maybe next time I'll use PICLI instead of CROME as it is not a fix defined by VORs and should be fly by.

Well, I think we need to go back to the drawing board. I loaded your flight plan up in the GTN Trainer and it flew it like I expected it would. This is the way I see it handle any course, it will turn before the actual point in space.










Sent from my iPad using Tapatalk Pro

[KLRDMD]

That's what I would expect. I don't have any simulators installed.

[Marauder]

That's what I would expect. I don't have any simulators installed.

I’m running another simulation with PICLI. I’ll see what that comes up with.


Sent from my iPhone using Tapatalk Pro

[KLRDMD]

Here it is on the Avidyne simulator (I just downloaded it). Exactly as it was flown. No lead-in turn at CROME but one at KAVQ. Hmmm . . . .

[Marauder]

Well, with PICLI in the simulator, it is doing the same thing. Does Avidyne have a simulator? Wonder if you can duplicate it on their simulator.




Sent from my iPad using Tapatalk Pro

[Marauder]

Here it is on the Avidyne simulator (I just downloaded it). Exactly as it was flown. No lead-in turn at CROME but one at KAVQ. Hmmm . . . .

 

Hmmm indeed. From what the Avidyne simulator is showing, the AP flew it as posted and then did the circle to get back to the route to KAVQ. Wonder if this is a programming error.

 

 

Sent from my iPad using Tapatalk Pro

[carusoam]

What decides that the angle is too sharp to execute, and then turns the other way for the 270? (Procedure turn?)

Something again would need to show lateral safe clearance altitude for the big procedure turn...

Kind of an uneasy feeling expecting a turn to the left and getting one to the right...?

The GPSS is trying to execute a tight turn, but the AP is unable to reach standard rate... the pilot is going to need to be on his toes regarding the overshoot and possible ground obstructions...

These are questions by an IR pilot with a few years to go before being current again...

Best regards,

-a-

[kortopates]

The control information (or GPS stearing on what to fly, when to turn) is provided by your Avidyne GPS. The Aspen GPSS just translates that to heading commands for an AP that doesn't know how to interpret GPS stearing commands.

Thus the issue is with the Avidyne GPS.

There is no such thing as a fly over Fix - only fly over RNAV waypoints - and CROME is a low Altitude enroute Fix thus has to be coded as a Fly By waypoint by definition.

 

So the Avidyne stearing is just plain wrong.

Why I don't know. Time to call up their support. Perhaps only a DB coding error.

 

But flying GPSS this way on an IFR flight plan could quickly lead to a pilot deviation as you turn the opposite direction from expected over CROME and IFR separation is lost due to nearby traffic. Let's hope it's an isolated database screw up. Really doubt it's anything to do with your less than standard rate turn issue.

 

Glad you figured this out VFR.

 

Let us know what they tell you.

 

 

Sent from my iPhone using Tapatalk

 

 

[Awful_Charlie]

5 hours ago, KLRDMD said:

It was in HDG mode, not NAV. The autopilot will not do standard rate turns under any circumstances. The gain is adjusted to the maximum possible value in the Aspen, twice normal and it brought it from less than half standard rate to more than three quarters. The GPSS should be coming from the Aspen, I would think. The FD shows as you would expect for the turn being done. The flight plan in the 550 was 57AZ-CROME-KAVQ-57AZ.

If the FD is only demanding a half standard rate turn then it suggests the servo is responding correctly. However, if the AP will not turn enough in any mode, I think this needs to be addressed, as it is not turning enough even in heading mode then my feeling is you are on a bit of a losing battle (unless you fly very slowly!).

Have you tried resetting the Aspen GPSS gain to standard? I have a feeling the values in that are "inverted" (eg GPSS=2 means half rate, and GPSS=0.5 is double), but in any case setting standard can remove that unknown

5 hours ago, Marauder said:

The Aspen has GPSS built into it, so I am guessing that is where he gets it. 

I don't know the S-TEC, but some more modern autopilots have the GPSS built into them, and if this is the case then the Aspen GPSS is out of the loop

4 hours ago, KLRDMD said:

I may be getting this figured out. CROME is an intersection, a fix on an airway defined by VORs. Fixes, by definition are fly over, you do not make the turn to the next heading when on an airway until flying over that fix. Maybe next time I'll use PICLI instead of CROME as it is not a fix defined by VORs and should be fly by.

In a flight plan route, ALL waypoints/intersections/VORs/NDBs etc are fly by. You can't demand code a fly-over in a route, that is the sole domain of Jepp (as the Navdata provider) in a procedure

I ran the route through the PC sim (SW 10.2.0) and got the same result. I tried creating a user waypoint at the same position as CROME and using that made no difference. By adding a waypoint before CROME to reduce the turn it behaved as I would expect:

Notice how these are both now displayed as fly-by.

I'd have to go flying to see how mine behaves with a very acute turn as in this case, but I would expect it to do as Marauder shows above

[Marauder]

4 hours ago, Awful_Charlie said:

If the FD is only demanding a half standard rate turn then it suggests the servo is responding correctly. However, if the AP will not turn enough in any mode, I think this needs to be addressed, as it is not turning enough even in heading mode then my feeling is you are on a bit of a losing battle (unless you fly very slowly!).

Have you tried resetting the Aspen GPSS gain to standard? I have a feeling the values in that are "inverted" (eg GPSS=2 means half rate, and GPSS=0.5 is double), but in any case setting standard can remove that unknown

I don't know the S-TEC, but some more modern autopilots have the GPSS built into them, and if this is the case then the Aspen GPSS is out of the loop

In a flight plan route, ALL waypoints/intersections/VORs/NDBs etc are fly by. You can't demand code a fly-over in a route, that is the sole domain of Jepp (as the Navdata provider) in a procedure

I ran the route through the PC sim (SW 10.2.0) and got the same result. I tried creating a user waypoint at the same position as CROME and using that made no difference. By adding a waypoint before CROME to reduce the turn it behaved as I would expect:

Notice how these are both now displayed as fly-by.

I'd have to go flying to see how mine behaves with a very acute turn as in this case, but I would expect it to do as Marauder shows above

I think we are all coming to the conclusion that something isn't behaving correctly. As Paul suggests above, it may be related to the Avidyne GPS database. The STEC 60-2 that Ken & I have in our planes don't have GPSS built into them like the STEC 55 series. We both get our GPSS from the Aspen. 

I have flown a number of sharp angle turns with my STEC. If there is a waypoint involved, the autopilot will always make a turn early to intercept the next course. If I intercept a leg, there is no lead in and the autopilot, if in normal GPSS mode, it will turn as much as standard rate if the turn is steep. If I have the autopilot in APR mode, the autopilot will fly through the final approach course a bit.

[Marcopolo]

Ken,

 

  I am almost 99% sure that this is standard operating procedure for the IFD units.  The pilots guide references this for any turns greater than 135*.  I have seen this behavior many times in my local area and on long cross countries (IFD 540, KFC200, GPSS).  When I see a turn on the map that doesn't depict the smooth radius I simply fly toward the intended waypoint and then activate the barber poled leg when I get within a mile or so of that waypoint where the high angled turn is depicted.  You can try this on the simulator.

 

Ron

 

From the pilot's guide "On any course change greater than 135 degrees, the FMS will treat the waypoint as a flyover"