Skywalker X8 options.h file (control gains flight tested and proven)

[Philip Giacalone]

HI all,

Tom and I just flew our Skywalker X8 for the first time. It included an 'interesting' launch and landing, but otherwise the flight was flawless under MatrixPilot/UDB4 control. 

The entire flight was flown in stabilized mode. The options.h file is attached. 

The link below show the flight from launch to landing. 

https://www.youtube.com/watch?v=NEkHUa9ZmzY

Best regards,

Phil

[crashmatt]

Congratulations on the maiden.

Could that launch glitch be caused by the IMU not having settled in direction?  Are you using a magnetometer?  If not then I would not launch in stabilized.  Fly in manual for 30s before you commit to stabilized.

/Matt

[Tom Pittenger]

We have a magnetometer, the problem was the gusty winds. The wing was lifting up before it even left the launcher.

--
--
---
You received this message because you are subscribed to the Google Groups "uavdevboard" group.
To unsubscribe from this group and stop receiving emails from it, send an email to uavdevboard...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[crashmatt]

I didn't notice that.  Do you think that more gain on the roll axis would help? 

The X8 disappears in most of the video so it is difficult to see what it is doing.  Any chance of a telemetry log?

[Philip Giacalone]

Hi Matt,

We will fly with an openlog next time. We had a 3DR radio connected and were planning to adjust the gains in flight on the second flight. But the winds kept increasing so we called it a day after that first flight. 

Overall, the flight stability was very good and the gains seemed reasonable. The best visible examples are during the climb out and landing. In both cases, I was barely touching the controls. During the climb I was making a slight adjustment to the roll trim and then trimmed the elevator at the top of the climb. We launched with 2 or 3 clicks of up elevator and the ailerons evenly trimmed. 

During the turns, the X8 felt very sluggish -- I had to push the stick pretty hard to get the plane to turn. The X8 performs like a heavy truck with great aerodynamics. Bad in the turns and very hard to slow down. The launch weight was 7 pounds. During the landing, I had the throttle off from the start of the descent above the trees, but the plane just never slowed down. On the video, we can still see the prop turning, so there's some question about what was actually going on with the motor. We wish we'd flown with the log, too. And also an onboard camera. Something for next time... with calmer winds. 

After watching the launch video several times, I was reminded of high performance kite flying. The plane tacked horizontally across the wind with the bungee pulling back against them. As Tom mentioned, the wind caught the left wing and started the roll while still on the launch ramp. We exited the ramp at pretty low speed, ~19 MPH ground speed. The plane then tacked across the wind while accelerating, the motor was then engaged, and the plane finally picked up enough airspeed to let the ailerons level the plane off. 

One other lesson learned was that our ESC is slow to engage the motor. It took more than 1/2 second to engage after pushing the throttle forward. Those were some precious clock ticks during the launch phase. 

Best regards,

Phil

[crashmatt]

Phil,

I see that the towhook is placed fairly well back.  That might contribute to the kite effect father than slingshot effect.  When gliders do highstart launches they need rudder to control them but you don't have those.  It would be amusing to see you properly bungee launch it.

When you say that the turn was sluggish, was the aircraft rolling enough?

The roll->elevator gain has traditionally been too low in the standard options.h.  This might contribute to a slow turn rate with higher than expected airspeed.

At 7lb your wing loading is about the same as my 3m glider.  Your stall speed should be quite low.  With some headwind it should be a gentle arrival.

You need to put the motor brake on for those folding props or they just keep windmilling around.  Having said that, they do add drag when they windmill which probably made your landing easier.  Some pilots with folding props actively use the motor as an airbrake by running the motor slowly.

/Matt

[Philip Giacalone]

Some very helpful info in your post, Matt. Thanks.

I was good to see your comments about props windmilling. While I was sure I had the throttle off during the landing, we could see and hear the motor turning on the video. Maybe that explains it. The log would have been nice to confirm it. Next time.

Hadn't thought of running the motor slowly as a brake. The instinct was to turn it completely off. It was a bit of a helpless feeling when the plane didn't slow down into a strong headwind. 

Our hook is 12 cm in front of the CG. We chose that spot to avoid a long moment arm in pitch, since the bungee is pulling from ground level. We also put it there so we could mechanically attach to the carbon fiber wing spar. I didn't think about the effect on 'kiting' behaviors. When you say 'properly bungee launch it' do you mean by placing the hook much further forward? 

Yes, the plane seemed to roll ok. I intentionally kept the roll angles small since I didn't want to push it on the first flight. And since there's no rudder control, there's no swinging the tail around. That combination made the turns sluggish. I didn't get the sense that I was fighting the autopilot. The autopilot performance was really good -- very stable with no surprises. 

We plan to adjust the gains in flight the next chance we get so we'll play with the roll-elevator gain. 

Best regards,

Phil

[Philip Giacalone]

Hi Matt,

Good to hear that the launch hook location is ok. Seems that the weather was our biggest problem. It was way too windy that day, but we were anxious to get our first flight into the air. Lucky for us it worked out ok. 

At this point, we're moving ahead with the build of a much better catapult and hope to have it ready in about a week. This should be much safer in all conditions. 

http://www.rcgroups.com/forums/showthread.php?t=1806237

The X8 is the test aircraft we'll be using to test our systems for the nasa challenge. We expect to do many flight tests, so hopefully this launcher will be much more reliable and safe. 

'S' turns are a good idea to bleed off airspeed. We'll do some flight tests of those at altitude next time and get a feel for the response. Been thinking about some sort of air brake but would rather avoid the complexity unless it's really needed. 

Anyway, I appreciate all of your advice, Matt. Good stuff! 

Best regards,

Phil 

 

On Tue, Jul 16, 2013 at 7:57 AM, crashmatt <uavfligh...@gmail.com> wrote:

Phil,

At 12cm ahead of CG you should not get much kiting effect.  Glider highstart towhooks are placed much closer to CG.

Try doing a stall test at altitude to check just how much you can slow it down.  Best to be in manual so you get to know if it likes to drop a wing.  The autopilot will hide that.

Not much you can do with a wing to loose energy other than some S turns.  If it becomes a problem then you might fit some airbrakes. Maybe part of the canopy popping up or some nice draggy landing gear popping down.

Regards Matt

[Phil Giacalone]

Hey Matt,

Just had the thought of an electrical thrust reverser for the X8. It would be simple to do using a double pole switch and a servo to throw it. No modifications required to the airframe. Thoughts? 

Phil

[Tom Pittenger]

PhilG,

Care to elaborate on how it reverses thrust with a single servo? I was hoping we could turn the motor backwards, or split the ailerons to add some flaps.

-TomP

--

[Philip Giacalone]

Hi Tom,

Yes, it's very simple. A double pole switch to reverse 2 motor wires and turn the motor backwards. A servo to throw the switch. I'd rather not modify the airframe. 

Best regards,

PhilG

[Tom Pittenger]

So an H-ridge or a physical relay to swap the wires around?

- An H-Bridge would change the impedance so i would stay away from that one.

- A physical relay could work.

[Philip Giacalone]

Just a double pole switch and a servo. The KISS approach :-)

[Murali- VU3MUV]

Tom,
I assume Matt means to say is,
Two of the three wire to the motor if crossed will  change the direction.

connect these two from motor to common poles of two way  dpdt ,  other pole of the switch is cross connected to speed contoller. This way switch can change the polarity and therefore the motor will reverse. The lever of switch is connected to servo to toggle its position thereby reversing the motor. I have done a similar mod to  my dads car which had faulty electric window controller. It wouldn;t go either up or down depending on the polarity of wire. adding such crossed connection did the trick. The original controller from Nissan was priced pretty high and so this mod for less than 5$.

here is the pic of what  Matt means.

cheers

Morli

[Murali- VU3MUV]

Oops,

I meant PhilG not Matt. my bad.

Morli

[Philip Giacalone]

Thanks Morli. I thought this was so obvious that it didn't require such an elaborate explanation :-)  

PhilG

[Murali- VU3MUV]

Hi Phil,
 It dosn't usually but I like to help. :).I assume that drawing is what you meant then. And there was talk about Impedance change ?!  Noob question what is H bridge and why would changing the polarity this way change the impedance?

Thanks

Morli

[Murali- VU3MUV]

BTW , you would need high current DPDT which I guess would be hefty IMHO.

[Tom Pittenger]

Hi Murali,

That is exactly what I was thinking, thanks for the great image!

The DPDT relay is a mechanical method to swap two wires, in this case it is the best solution. However, an H-Bridge is an electrical way to do it: no moving parts. It is a common method to reverse DC motors, such as your car window example. We would need two bridges, one for each wire that needs to get reversed. I mentioned the impedance matching because we would have two channels passing through the bridges but another channel not, it would electrically be imbalanced and our efficiency would drop. Plus the motor controller may not like it, it is expecting a balanced circuit and may overheat due to the efficiency loss.

-TomP

[Murali- VU3MUV]

Thanks Tom for the H bridge link. Reading . :)

[Tom Pittenger]

Here is a nice link with a good picture that explains it well with an end result of 3 IOs. With some simple tricks you can typically combine the FWD and REV pins into a single FWD pin which also gets inverted to feed the equivalent of the REV pin. By pulsing ENA you can control the speed too. If you get super fancy on your circuit, you can use it to route current backwards when the motor during braking to charge your battery. This is the basis of regenerative braking and is used in Hybrid and electric cars (convert your kinetic power into a power source to charge the battery). When you start getting complicated mechanical solutions quickly become a bad idea.

-TomP

[Murali- VU3MUV]

Thanks Tom,

Ya that explains it better.  Now does it not mean we still have to connect one  wire direct to brushless motor?

Off  topic question:

I am trying to make mini windmill for my daughter's school project. This is same as jewel generator(but cannot buy it),  need/looking for the circuit for 9-12V out put. cheers and thanks

[Phil Giacalone]

Hi Morli,

We've got a 50 amp ESC so you're right about the DPDT switch having to handle high DC current. Turns out these switches are rather expensive! Given this, what I'd like to do instead is figure out a way to program the ESC to reverse the direction of spin under Tx control. If this were doable, this problem would be solved without any additional hardware. If not, then I guess we'll go back to trying Matt's original suggestion -- which is to simply run the prop at very low speed to increase drag during the landing approach. 

We're using a Thrust Ultra ESC but the programming instructions aren't very helpful. Wondering if anyone has experience with ESC programming to be able to control/reverse the spin direction. 

Best regards,

Phil

[Phil Giacalone]

Thanks, Tom. Seems we're all posting to this topic at about the same time :-) 

For our next flight let's try Matt's idea of turning the prop slowly to see if this effectively slows down the plane. We can get into more elaborate solutions if needed. 

Best regards,

Phil

[Andrius Narbutas]

On Sunday, July 21, 2013 8:50:04 PM UTC+3, Phil Giacalone wrote:

Yes, it's very simple. A double pole switch to reverse 2 motor wires and turn the motor backwards. A servo to throw the switch. I'd rather not modify the airframe. 

How complex...
Why not flash ESC with SimonK FW, enabling SLOW_THROTTLE, BRAKE and MOTOR_REVERSE? Then you won't need any aditional stuff (also, COMP_PWM recommended if you plan to do reverse in air - for "smooth braking") - just set-up endpoints correctly.
People use this stuff on copters, for 3D flying, for example:
http://www.youtube.com/watch?v=7Ldntzjhm1g
Also, could be used on planes, i used the same reverse on DIY polaris, to pull out of weeds if landing isn't perfect:
http://goo.gl/i31hR0
(not using in air - probably would disintegrate)

P.S. almost on topic - what's the point to share gains settings on various planes? You change servo - need to re-tune, changed linkage hole - need to re-tune - so many different factors affecting gains, airframe plays very little role here...

[crashmatt]

If you are using a hk controller then you might get Simonk on it but it depends on it being an arduino based controller.

Alternative is to buy a car esc that already has the reverse capability.

Another problem is how to control this from the transmitter. I have a feeling you will need to modify the Servo mixing on matrixpilot.

The very simplest approach is to check if it will fly more slowly.

[Philip Giacalone]

Thanks, guys. More good food for thought. 

Ok, we're going to avoid adding any complexity for now and see how we can pilot the plane to slow it down during our next test flights. 'S' turns, as well as a slowly turning propeller, will be tried... 

Andrius: You mentioned, "flash ESC with SimonK FW." It's good to know that this exists. And of course your right about the gain settings being highly dependent on many factors. The idea of posting them was just to share something that works for the X8, since nobody had previously posted data here. We've been posting/sharing gain settings for other aircraft, too, in the hope these might be helpful to others -- at least as a starting point. 

Best regards,

Phil

[Murali- VU3MUV]

Car Esc have pretty big sink and looks heavy ,so don't know if they would suit in plane without little mod,  however Car Esc is good idea. Why did I not think of that :o(

[crashmatt]

Morli, I cheated. It's not my idea. It came from a diyd post a few years back.

[Andrius Narbutas]

On Monday, July 22, 2013 12:09:25 AM UTC+3, Phil Giacalone wrote:

Wondering if anyone has experience with ESC programming to be able to control/reverse the spin direction. 

What exactly do you need?
No idea what's "trust ultra" ESC is, but most ESC's could be re-programmed to alternative FW - SimonK (with Atmega chip), BLHeli (with Silabs). Both supports RC-car style reverse and could be used even in air (dangerous, used improperly probably will torn plane in half... Easy).
Start here, if you need alternative FW for ESC:
http://www.rcgroups.com/forums/showthread.php?t=1513678

[Murali- VU3MUV]

It still is a good Idea Matt. 

On Tue, Jul 23, 2013 at 9:47 AM, crashmatt <uavfligh...@gmail.com> wrote:

Morli, I cheated.  It's not my idea.  It came from a diyd post a few years back.

[Riccardo Kuebler]

Hi Phil,

if you can find a 50A SimonK flashed ESC, like this 30A one, you could easily reflash it with motor reverse option, with a simple USB programmer through the PWM cable.

An idea I have from longtime now for the Funcub to have reverse thrust while diving vertically with full flaps for landings.

Best regards,

Ric

2013/7/21 Phil Giacalone <scienc...@gmail.com>

[Andrius Siug]

Hi, can anybody share options.h file for matrixpilot4? Necessarily for X8, i`m new in this and want to see example. Thanks

[Andrius Narbutas]

On Saturday, January 18, 2014 10:25:06 PM UTC+2, Andrius Siug wrote:

Hi, can anybody share options.h file for matrixpilot4? Necessarily for X8, i`m new in this and want to see example. Thanks

http://code.google.com/p/gentlenav/source/browse/#svn%2Ftrunk%2FMatrixPilot%2Fexample-options-files
?

[j]

--
--
---
You received this message because you are subscribed to the Google Groups "uavdevboard" group.
To unsubscribe from this group and stop receiving emails from it, send an email to uavdevboard...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

dont think there is an X8 example there.

[Andrius Narbutas]

On Saturday, January 18, 2014 11:31:11 PM UTC+2, vwyoda wrote:

dont think there is an X8 example there.

Misunderstood question.
http://code.google.com/p/gentlenav/source/browse/branches/MatrixPilot_NASA_challenge/MatrixPilot/options_X8.h
Needs to be slightly modified for current trunk (also, it's not for UDB4 IIRC).
Have in mind, that gain settings depends on servo and their throws, so do not expect copy-pasted settings to work.
Also, the same "spin on loiter" but applies to X8 as it's flying wing - go with low gain settings, or do not include loiter in your flight plan.

[Tom Pittenger]

To my knowledge, Phil G flies the X8 Skywalker with the nasa code with the options.h file that was kindly linked above by Andrius (Thanks!). PhilG hasn't flown on trunk in a very long time. Therefore, he has been flying the X8 with a UDB4.

As an update to the initial video on this thread, I'd like to post this video which is now a 'typical' launch using PhilG's custom catapult.

https://www.youtube.com/watch?v=YKX3npC_f3o

-TomP

--