Glideslope tracking

[kellym]

I note from the old archives a thread about Skyview autopilot not tracking the glideslope in an RV-10.
I have the Classic with latest software revision, the linear actuator version of the pitch servo in my RV-10. I am using GTN-650 as nav source.
It consistently lets the GS go 1/2 dot low before capturing and following the glideslope.
It does eventually get on glideslope. I have tried both increasing gain and increasing sensitivity. Both cause porpoising as soon as GS is received, so went back to original settings.

 

[Dynon]

We have some AP improvement coming (not sure off the top of my head whether this is one that will be improved). But, have you tried using the Autopilot Inflight Tuning Guide? It has helpful tips on how to tune autopilot performance.

 

[kellym]

Yes. I found that if I increased sensitivity to get IAS tracking to match the guide that it got twitchy and would porpoise and hunt on a glideslope, seeing some of it on the flight director bars. Returned the settings to close to default settings and it got better. Same for roll settings. It was best to ignore the needles going 1/2 dot off, as it usually would come back to centered. I have had a couple cases where the AP stopped tracking the localizer and stayed a full dot or more off-center, where I had to switch to heading mode to get it back, then put it back on localizer.
It would be helpful if you would publish recommended yaw settings. The defaults do not work and have just used what a few others found worked for them.

 

[DBRV10]

Kelly,
There is an improvement coming, I am testing it at present and the first flight was an immediate improvement. Of course with all these things it becomes more complex with tuning for cruise and for radio navaid and GPS approaches, but its just another couple of dedicate parameters.

I am partly responsible for the recent service bulletin. Well I didn't cause it but another RV10 owner and I discovered while trying to help him out, that there has been a material change in the SV42T actuator by a supplier, and YOU MUST read and verify if your servo is affaected immediately. No amount of tuning will engineer out a crack in the disc.

For now, with the servo not cracked, and current software, try the following;
Torque 100%
Sensitivity 16 or 17
Gain 1.5 to 1.9 We found 17/1.5 or 16/2.0 but in that range it was good.
Altitude gain 0.9
Everything else default.

Now the next thing is trim rates.....autotrim needs to be something like 105kts=100% and 130kts = 35-45%

I do not know if the next release will have the improvements. There is another bug we are trying to smash, have been for years and it seems it is nailed, so good results to follow.

 

[N42AH]

A

Kelly,
There is an improvement coming, I am testing it at present and the first flight was an immediate improvement. Of course with all these things it becomes more complex with tuning for cruise and for radio navaid and GPS approaches, but its just another couple of dedicate parameters.

I am partly responsible for the recent service bulletin. Well I didn't cause it but another RV10 owner and I discovered while trying to help him out, that there has been a material change in the SV42T actuator by a supplier, and YOU MUST read and verify if your servo is affaected immediately. No amount of tuning will engineer out a crack in the disc.

For now, with the servo not cracked, and current software, try the following;
Torque 100%
Sensitivity 16 or 17
Gain 1.5 to 1.9 We found 17/1.5 or 16/2.0 but in that range it was good.
Altitude gain 0.9
Everything else default.

Now the next thing is trim rates.....autotrim needs to be something like 105kts=100% and 130kts = 35-45%

I do not know if the next release will have the improvements. There is another bug we are trying to smash, have been for years and it seems it is nailed, so good results to follow.

Any update on when the update will be available?

 

[DBRV10]

Not yet, I have no control over that.

 

[kellym]

To add a different twist to this question, I find no documentation as to what parameters must be met for the VNAV to capture the GS.
I have difficulty with a local ILS approach to get the capture. I suspect the holding pattern entry to the approach places the aircraft too close to GS intercept (about 1.5 nm) when turning inbound on the localizer. Would appreciate any parameters such as how close to on course on localizer and whether capture is possible only below glideslope or how much level flight is needed to capture.

 

[airguy]

In my experience it can only be captured from below the glideslope.

 

[kellym]

In my experience it can only be captured from below the glideslope.

Matches my experience.

 

[GaryK]

Some may already know this but the replacement actuator is now shipping. I checked mine as soon as the SB came out and confirmed mine was affected. No cracks yet but Dynon confirmed it was bad based on high resolution pics. Installed the new actuator last week and sent back the old one with RMA number. Glad to have it changed out as it made me nervous thinking it could crack.

 

[DBRV10]

You should always intercept a GS from below. And it captures at half scale. I have never tried it slightly high, so I cant comment about that.

 

[jchang10]

I too have an issue which is not addressed in this thread. Is it just me or anybody else get this? Granted, I have only definitely noticed this at my home drome. Both GPS 12 and GPS 30 Approaches with a Garmin 400W.

The issue comes in at about the last 5 seconds of the approach to DH at 380 ft AGL. Around 600 ft or so, the AP consistently comes in low on the glidepath a good quarter to a half dot too low. After tracking the GS perfectly from FAF, why would the last few seconds consistently get that low? It is long enough where the AP does not attempt to correct by levelling out at all.

The last few times while VFR/VMC, I let it get low hoping it was just temporary and it would correct. However, it has never done it. I now realize it is consistently doing this.

Every other aspect of the AP behaviour, in cruise, in climb, in descents, are all good. It is frustrating to think that the last few seconds of the approach, the AP is misbehaving at such a critically low altitude.

Been flying this setup for a while but only consistently flying IFR the past year.

Jae
SV-52T (I think)
RV-10

 

[DBRV10]

As Dynon posted above there is a massive improvement coming. I have been playing with the new version and made even more ground. I will publish the full set when the new software comes out.

For now here are some improvements you might like to try, especially for turbulence.

Torque 100
Sensitivity 18
Pitch Gain 1.6
Altitude Gain 0.9
Pull rate 2.0
VSI Gain 1.5
G error Gain 1.2
G error limit 0.7

 

[DBRV10]

Make sure you have done the SV42T service bulletin.

 

[jchang10]

Thanks David for those suggestions. Ok, as of today, it seems like I was crying wolf. It just goes to show how difficult it is testing these devices. Hats off to Dynon for accomplishing what they have accomplished. It's not easy getting good tests.

I recorded another 2 practice approaches with my camera phone recording the screen. Of course now, it pretty much flew the approach perfectly, both 30 and 12. It did dip a quarter dot but it actually did recover nicely. I was doing this solo, which is really hard to do. I will try and post up the video later. I will continue testing, however. One anomaly for today, was both approaches were with a tailwind on final approach. How often does that happen?

One theory I had was that the reason for going low, was the reduction in the headwind component as you get closer to the ground. The AP was just not adjusting quickly enough for it, I wonder? I will have to wait for a better conditions to test that theory. Watching my video, it is quite amazing how fast things happen in the last few seconds.

I am hesitant to adjust the Pitch axis settings, just because everywhere else, the AP perfects great for me. I would hate to fix it for the last few seconds of approach, only to upset it elsewhere.

Bottomline for me, it worked perfectly today. However, I will wait for a day where there is a stronger headwind to test my theory above.

 

[DBRV10]

Take photos of your current settings, then go try mine. You can always go back.

The next version I expect will have the extra tuning available for both GPS and VOR/ILS approaches, and this seems to be the trick. But for now try mine. See how it goes. Also you want to be doing 95-120 knots, and your auto trim needs to be at 100% speed in the below 115 knots range.

 

[jchang10]

Cool. Looking forward to any improvements. Makes sense to me, to have separate settings for enroute versus approach. I believe Garmin APs have a specific "approach" mode setting for probably this reason, too.

https://photos.app.goo.gl/X9KLZpA7oHw4sQ567
Album should have both approaches to 30 and 12. On short approach, it shows up to 1/2 dot below GS. However, the AP did correct it, which is great. Previously, I did not see the correction before having to quit, but this time it actually did.

 

[DBRV10]

I see you do not have the trim wired through the AP panel. This should be done and then set the trim for max at or below 110-115 and about 35-40% at say 130 knots. I know thats a bit of work, but it helps a lot with the AP. In the interim you need to keep it in trim as per the trim flags.

 

[jchang10]

Actually I do have it wired. Works great. However it's really slow sometimes. I do see the trim indicator scrolling in the video.

 

[DBRV10]

OK....I did not see it just the trim flag......Mind you, if it trims for 3-4 seconds it stops, if it needs more trim you will need to tap the button and reset it.