PDA

View Full Version : Dynon AP info from Dynon


Brantel
09-23-2009, 05:21 AM
Quote from Dynon Forum:

"This is where we are on the autopilot.

Dynon is totally committed to the autopilot, since it is a big part of both the current D10/D100 Series and SkyView. We haven't been able to move as quickly on the current software release (5.2) as we wish, but we will deliver it. And we will continue to release software for the D10/D100 Series after that, with much of the changes focused on the autopilot. For many existing and future customers the D10/D100 Series is a better match to their requirements than SkyView, and we will continue to support all of our products. I hope our track record has given you reason to believe in the company despite the autopilot not being perfect for everyone right out the door.

One lesson learned is that we need to be very cautious in discussing future plans. (We feel a little squeezed, since the number one question we are asked is "what new features and products are you coming out with, and when?") But with that in mind, these are the autopilots tasks we are working on:

* General autopilot performance. This is aircraft specific, and the Dynon autopilot works quite well on most aircraft. Some not so well. Based on our research this would be a pretty accurate statement for all AP's on the market, but we are working to improve the Dynon performance on all aircraft. For anyone who has an autopilot performance problem, we try to work with them, obtain data, and see if we can improve it. We are working actively with owners and OEMs to optimize performance. Frankly, our performance and feature set is often being compared to A/Ps that cost $6000, $9000 or more. We are flattered, and fully intend to work better than other autopilots even in those higher price ranges.

* There is an intermittent dropping of GPS. We do not yet have a firm handle on why, and so do not yet have a firm date for a fix.

* Does the Dynon autopilot require higher-torque servos? We are sometimes being compared to autopilot models that only do Altitude Hold, which obviously requires less torque than Altitude Capture and vertical speed changes. And in some cases we are compared to competitive installations that require a torque enhancer. It should be noted that the Dynon autopilot actually indicates servo slips graphically on the screen while TT doesn't show you at all, and in fact cannot show you given the design of their servos. So it's not surprising that people may never know or care that a TT servo is slipping in the same situation as the Dynon autopilot.

* Will the Dynon autopilot be improved with more customer settings? That is still being worked on. So until we know for sure, we cannot answer. We have tried to keep settings to a minimum by designing them out of the system. Where we find it's better to give the user an adjustment, we will. We did learn early on that even a little slop in the servo http://dynonavionics.com/public_html/yabbfiles/Templates/Forum/default/wink.gif, linkages, and cables can cause major performance differences, so more adjustments may be necessary.

* Dynon has a full-time autopilot expert, and he along with all of our development staff are working hard on both improving the autopilot and integrating it into SkyView. He is working with existing customers and installations to improve their autopilot performance, as well as designing new features. We have both customers and OEM engineers who we are working with to optimize performance. And when an existing customer has a problem, we request that they log their system data so that we can analyze their performance and help improve performance.

With all that said, we feel it is most helpful to respond to specific aircraft performance issues rather than answer generic questions. We ask, and hope, that anyone with an autopilot performance question get in touch with us directly. We actively analyze data from any customer who reports a problem and incorporate what we learn into our improvement program.

Dynon Support"


Overall I would say that this post is encouraging....

breister
09-23-2009, 08:55 AM
* Does the Dynon autopilot require higher-torque servos? We are sometimes being compared to autopilot models that only do Altitude Hold, which obviously requires less torque than Altitude Capture and vertical speed changes. And in some cases we are compared to competitive installations that require a torque enhancer. It should be noted that the Dynon autopilot actually indicates servo slips graphically on the screen while TT doesn't show you at all, and in fact cannot show you given the design of their servos. So it's not surprising that people may never know or care that a TT servo is slipping in the same situation as the Dynon autopilot.

TT does show you if you are out of trim. Since APs don't fly aerobatics, a trimmed airplane should never cause AP slippage. And although I can't speak for other airframes, I do know that in my airplane if the trim is off enough for slippage you REALLY feel the airplane bobble.

Thus, while a technically true statement on Dynon's part it is not a deficiency in TT. Rate this one as "good marketing..."

:D

dynonsupport
09-23-2009, 12:28 PM
If an airplane is in trim and you ask it to climb, the airplane slows down and is suddenly not in trim. It requires more and more force to make it climb more. It doesn't take aerobatics to need a strong servo to fly a plane.

Yes, if you keep trimming the whole time, the servo doesn't have to to much work. But then what's the point of the AP, since you are flying the plane anyway? The whole trim change problem is why autopilots usually require trim sensing servos when you buy a AP that can actually change altitude rather than just staying at one altitude. They expect you to stay on top of keeping the plane in trim, just like we do.

TT show you out of trim with some of their systems. We show it with all installs, partially because all of our APs will do altitude changes. What is different is that we show when the servo actually slips, which is totally different from just being out of trim. Since you never know if a TT slips, it's hard to know unless it really affects AP performance. We blink at you whenever we slip, even if we still are able to control the plane. For all you know with your TT, it's slipping 90% of the time it moves, but it's only when you are really climbing that you feel the bobble. Since slipping causes no damage to the system, this is totally fine, and clearly the TT systems work great so there's no reason to go to a larger servo.

We do have repeatable proof that our servos can exert more force than a TT servo of the same size. And we have a good technical reason that is true because we control the stepper motor in a more complex way than they do. Admittedly, the increase is small, and not something that means you can run a smaller Dynon servo. Given that, we're not totally convinced that we need a bigger servo in an install, given the EXACT same mission and installation method. We really do think that people often ask more of our servos than they do of other servos because we allow so much out of even our basic AP.

breister
09-25-2009, 12:20 PM
I started writing a longer response but it seemed like the longer it got the more it sounded like "making fun of your feature." That's not my intent, and at the core it is a matter of personal opinion as to how much that particular feature means to any one person.

I'm a big proponent of pilots learning everything about their systems. For example, while learning about their autopilot pilots should reduce the maximum torque setting and then run through climbs and descents (after first trimming to cruise airspeed) to discover how low the torque setting needs to be to purposefully cause their AP to slip (both in smooth and again in turbulent air). If you are actually looking for it it is trivial to sense - just keep your fingers lightly resting on the stick / wheel and you will feel it tick-tick-tick as it slips. If your installation requires close to maximum torque before the AP can climb and descend without slipping, IMHO you should order heavier servos.

However, once you have gone through this initial setup & familiarization, IMHO servo slippage really should not be an issue at all. Since I'm on the side of the fence that doesn't even think VISUAL TRIM annunciation is really a "critical safety of flight" item (for precisely the reason that I think slippage is unmistakable in the airplane, thus acting as primary annunciation by itself), I'm not likely to be swayed one way or another by a "slippage indicator." But, clearly others will feel differently, which is why a free market is a wonderful thing!

:D

One final thought on all of this pondering - it occurs to me that the "slippage problem" can be turned into an advantage. If you want to climb at a particular AIRSPEED rather than RATE, simply dial DOWN the maximum servo torque in settings and dial in a 5,000'/min climb. As the airspeed drops and the servo exceeds the force available, it will slip (harmlessly, as you have pointed out) and will continue to slip until the airspeed is once again high enough to put the trim back into a range where the AP can raise the nose once again. Small changes in trim would then alter the "breakout point," thus allowing you to select fairly precisely the speed at which you want to climb.

Think I'll try that trick this weekend....

Sam Buchanan
09-25-2009, 12:31 PM
One final thought on all of this pondering - it occurs to me that the "slippage problem" can be turned into an advantage. If you want to climb at a particular AIRSPEED rather than RATE, simply dial DOWN the maximum servo torque in settings and dial in a 5,000'/min climb. As the airspeed drops and the servo exceeds the force available, it will slip (harmlessly, as you have pointed out) and will continue to slip until the airspeed is once again high enough to put the trim back into a range where the AP can raise the nose once again. Small changes in trim would then alter the "breakout point," thus allowing you to select fairly precisely the speed at which you want to climb.



For the benefit of those still researching autopilots, some pitch autopilots allow you to select either rate or airspeed for climb/descent to selected altitude.

dynonsupport
09-25-2009, 06:24 PM
But many of those AP's don't let you dial in an altitude to stop climbing or descending at because they literally don't have any idea what absolute altitude you're at.

Lots of tradeoffs!

Sam Buchanan
09-25-2009, 07:40 PM
But many of those AP's don't let you dial in an altitude to stop climbing or descending at because they literally don't have any idea what absolute altitude you're at.

Lots of tradeoffs!

Uh.......check out the Trio Pro Pilot and some of the Tru-Traks. Select your target altitude (you sync the device to the altimeter using a barometric setting on the autopilot), select either rate or speed, punch the button, put your hands in your lap, and watch the plane climb on speed or rate to the target altitude, smoothly level off and capture altitude.

Every time. :)

Sorry for the thread creep, just wanted to clarify.

RV10inOz
09-25-2009, 08:57 PM
The Tru Trak climb rate and capture in an RV7 I have flown works much better than the climb to or descend to feature in out Dynon. We manually do the vertical part.

Unless we can do something different, but I expect the VNAV solution one day will have better Vertical software.

I think the Dynon determines a change in altitude at a given rate means get there in exactly X minutes.....and if any turbulence helps or hinders, it just keeps targetting the final time at a max rate.....

For example if you sink 500 feet quickly, it will level out at 0 VSI for a while then descend again some time much later rather than continuing down trying to maintain the VS.

I may be wrong but thats how it seems to work.

I love our Dynon AP..... I just think the Vertical part needs work and work that is most likely in the future releases. So patience patience it has to be........probably not good at it though :rolleyes:

DB:cool:

carguy614
09-26-2009, 06:55 AM
Following this very interesting thread. I use a Dynon 100/120 set in my 9A, and have a TT single axis AP. Love them both. Will definitely use Dynon in my current project, and intend on using the Dynon AP as well.
The message that I get from this thread is how integrated we as users are with the folks that design and manufacture out equipment. Dynon has proven it's dedication to the support and improvement of it's product, (as has TT I might add) and deserves applause.
I have no doubt that Dynon will work out all the glitches, and they rely on us for input to help accomplish that task.

Good job guys...
Chris

flickroll
09-26-2009, 06:55 AM
Overall I like my Dynon AP and it offers a lot of bang for the buck. I believe they still have some software tweaks to do:

Seems I am always messing with sensitivity (particularly with roll) to fine tune the performance on a given day. One day it will perform great, but next time it might be hunting or sluggish, requiring me to tweak the setting.

On the way to Osh this sumer I was climbing VFR to 12,500' to get over clouds over WV. Was doing fine til the last 2000' or so when it DRAMATICALLY pitched nose up. I quickly disconnected, stabilized, reengaged, and it did fine for the rest of the climb. Only time this has happened, but it makes me a little skittish to use the AP in IFR conditions.

Also, I'm one who is a little PO'd about the delay with the AP76, and somewhat based my purchase of the Dynon AP on its availability. I hope they roll it out in the not too distant future. My -8 has been in the air since April with a D-100 and a D-120, and I have no plans to switch to Skyview for at least a couple of years, if then.

rvmills
09-27-2009, 01:17 PM
I think the Dynon determines a change in altitude at a given rate means get there in exactly X minutes.....and if any turbulence helps or hinders, it just keeps targetting the final time at a max rate.....

For example if you sink 500 feet quickly, it will level out at 0 VSI for a while then descend again some time much later rather than continuing down trying to maintain the VS.

I may be wrong but thats how it seems to work.

DB:cool:

DB,

Interesting observation...has me wondering about how the Dynon AP does the vertical speed hold in climbs and descents (maybe they'll chime in here).

My impression was that the vertical speed, as set in the menu, is the target, and the AP pitches for that vertical speed, and adjusts pitch to maintain it (rather than calculate the altitude change, divide it by the target rate to establish a time-to-climb/descend, and then vary VSI to make that time). Not sayin' it ain't so, but it seems like a complex way of doing it.

I see those pitch (thus VSI) oscillations in climbs and descents as well, especially in bumpy air, and was assuming (uh-oh) that it was just the AP correcting (maybe sometimes overcorrecting) to get back to the target VSI. I will say that I have timed some descents, and even with the VSI oscillating a bit, the time to descend was usually pretty close to target (ex: 4 minutes to descend 2000' with 500 fpm set in the menu). Your premise is interesting, and I'd be keen to find out what the target really is (time or VSI...don't think it'd make a diference in how I fly it, just interested).

FWIW, I have seen a good improvement (decrease) in pitch oscillations in level flight since doing the service on the servo's shear screw, and to a degree, it has also improved the climb and descent feature. Smooth air peformance is solid in level flight, and acceptable in descents (haven't used it too much in climbs so far). When it's bumpy, I've been able to keep OK performance when level by decreasing the sensitivity a couple notches, but the descents are a little more "pitch chasing" than I'd like. Did a 4 leg X-C over the last few days, and saw the above (smooth air = fairly acceptable descents on AP, bumpy air = hand flown descents...for me anyway). So I would concur, that (VNAV/Vert Speed) aspect could use some improvement.

I too hope that continued AP software/capability improvements (and the AP-76 development for those that want it) go hand in hand with Skyview development and launch. Seems like there are some synergies there, and it sure seems Dynon is committed to its customer base and plans to do just that (and I'm a patient guy ;)).

Cheers,
Bob

RV10inOz
09-27-2009, 04:03 PM
Bob

I would have thought it just held a VS.....if that is what it is meant to do....its extremely poor at it.

We have ALT hold working quite well, so no complaints there, but the way I see it do an altitude change i think it creates a target and time and tries to achieve it.....feels that way...you soon get sick of it and disconnect.

I am keenly awaiting the AP76.....patience pays they say! :)

dynonsupport
09-28-2009, 05:25 PM
So the discussion here is in the ballpark. Without getting into the deep algorithmic nitty-gritty, the AP is currently a bit too sophisticated for its own good. We favored an overall approach that set us up well for the future development of VNAV features (which requires that you maintain a certain path through space, and that you erase position errors so they don't add up to big excursions off glideslope). But, as we're learning, this approach has some unwanted side effects for some in altitude change mode. In any case, this is one of the things that we're taking a hard look at when we talk about the generic "autopilot performance improvements" to make all Dynon AP equipped aircraft fly as well as they can. Know that we do have someone actively working on this and other AP improvements.

If any of you have any specific performance problems, the BEST thing you can do is get in touch with us directly. Taking a datalog at 1/sec interval (under the EFIS>SETUP>DATALOG menu) will help us see what you're seeing too. The best way to start a dialog about your aircraft is to email us at support at dynonavionics dot com with some details about your aircaft, AP setup (aircraft, servos, torque, sensitivity settings, any other details) and the datalog from your flying (download using the Dynon Support Program.)

RV10inOz
09-29-2009, 05:40 AM
Good reply DS! :)

That makes a lot of sense to me..... and is in line with what I have observed.

I do look forward to testing the VNAV some time next year! ;)

The AP74 or non VNAV system needs a different methodology for sure.

I am still thinking of coming to see you guys, offering to help where I can...from all the way down here!

Let me know who I should talk to!

DB

Andy_RR
09-29-2009, 07:21 AM
the pitch controller just take over control of the trim servo rather than adding an additional AP servo? Would save some weight, although obviously adding some design/FMEM complexity, perhaps. Probably the slew-rate of the trim servo would have to be a little higher too, but you'd never have to worry about servo slippage or out-of-trim conditions.

What am I missing here - I feel sure there's something! :)

PS: apologies for the slight thread drift...

SteinAir
09-29-2009, 08:51 AM
the pitch controller just take over control of the trim servo rather than adding an additional AP servo? Would save some weight, although obviously adding some design/FMEM complexity, perhaps. Probably the slew-rate of the trim servo would have to be a little higher too, but you'd never have to worry about servo slippage or out-of-trim conditions.

What am I missing here - I feel sure there's something! :)

PS: apologies for the slight thread drift...

There is a whole slew of reasons that it won't work well. It's been tried with little success in high performance airplanes. Remember the trim tab is not the positive control point for the elevator, the trim motors are slow and inherently behind the plane when it comes to positive control of said control surface. If the you sped up the trim tab to the point of what you'd need for positive control, it'd become an actual servo or anti-servo tab instead of a trim tab...

Anyway, this could be and has been discussed at length in the past. If you do a little searching there is a lot of good info here on VAF and on the Matronics list regarding use of the trim tab for an AP.

Cheers,
Stein