Telemaster FPV PicoPilot project

[smitty505000]

Hey all,

Just a few details of my latest project. The 6' Telemaster Electro with FPV and PicoPilot. My goal is two fold: 1 is to be able to fly FPV and have the plane return home in the event of radio or video loss. 2 is to have the plane follow a set of waypoints and just look around while it does.

Equipment

Telemaster Electro
Axi 2826/12
PH 40 ESC
4x TP 2100 PL (3 for motor and rx's, 1 for video eqipment)
2x Dynam BEC
2x Berg 7H RX's (1 for channels 1-6 & 1 for channels 7-8)
Inspire OSD
Geko 201 GPS
PicoPilot NA
Pandora Pan & Tilt with KX131 cam
Bwav 200mw video tx
Diversity video rx
Icuity DV920 goggles
Track R1 Head tracker

First flights made this weekend were good but she tended to pull left and pitch up. I added some down thrust and right angle to the motor but I think I need more down because she glides nice with nutral elevator but when power is applied she pitches up quite a bit.

Smitty

[RaptorAP]

Pitch up is due to airfoil when power is applied. You'll drive yourself nuts trying to trim it out, just realize its not trim or motor mount angle.
Nice project, be sure to post video!
Mike

[clolson]

The Senior Telemaster has a lifting tail which seems to do a pretty good job of keeping constant attitude as the speed changes.

Curt.

[workshop]

I have a very similar system: http://www.rcgroups.com/forums/showthread.php?t=631757

Looking around while the plane flies autonomously is REALLY fun. One can concentrate on observation and leave the flying comfortably to the PicoPilot.

Jeff

[tomsal]

The pitch up is, in part, because of the airfoil but you do need more downthrust.

Without downthrust the aircraft will climb when you add power even though it glides great power off. You want that good glide power off and can trim it to fly level at the cruise power setting you want with downthrust.

Once trimmed this way the plane will fly level at the cruise setting and climb with power settings above that. This is a desirable condition with FPV especially beyond visual range because you can get the plane to descend in a controlled way by reducing power and get it to climb by adding some. No need to move the elevator which can lead to trouble.

An aircraft with a flat bottomed high lift airfoil is basically a constant speed device when properly trimmed. Full sized aircraft are the same way. Trimmed for level flight at cruise power they will descend when power is reduced and climb when power is added all to maintain the trimmed speed.

In full sized aircraft this tendency to climb or descend is counteracted with elevator trim, in models level flight at the desired cruise power setting is achieved through downthrust.

Hope this helps.

Tom

[smitty505000]

Wow, thanks for all the replies!

Mike, I will post some video hopefully next week. Should have it trimmed out by then.

Curt, The Tele Electro has a lifting tail as well.

Jeff, yes! That is one of the threads that gave me the Idea! With the pod on top, I bet you dont have much problem with the nose wanting to pitch up.

Tom, I understand perfectly now. I am adding some more down thrust and will give her a go. I am shooting for a crusing speed of about half stick.

Wondering if I should be using Pico Pilot with alt-t instead of my alt-e version??

Smitty

[icebear]

Great setup you have there Smitty - really impressive!

I think you will be just fine with the Alt-E. I have used the Alt-T a lot and it works well for airframes with a positive pitch response.
However, with the TM you should be able to set a comfortable cruise speed and let the Alt-E manage altitude if you just make sure that say half throttle cruise gives you a straight flight path just as you suggest.
I believe Jeff is using the Alt-E on his TM with good results.

Bjorn

[smitty505000]

Thanks for the kind words Icebear.

I added some more downthrust and hopefully get to give it a go Saturday. If she goes good, I will then move on to the PicoPilot and make adjustments as needed. So my goal for the weekend is to get her flying good and get the PicoPilot working.

How critical is it that the PicoPilot is level in respect to pitch? The manual recomends it to be level with the plane while flying and says most planes fly slightly nose up 5 or 10 degs.

Smitty

[icebear]

Hi Smitty!

I have always placed mine with a good "forward tilt" so that you make sure that it is always at least neutral. I used foam & rubber bands around mine and put an extra piece under the rear of about 1/2" and that has worked really well.

Maybe you have already done this, but set your mechanical throws (ail+rud) quite conservative and set the gain low when you start testing and try to increase gain until you are happy with winglevelling & navigation.
I recall you use a Y-cable between ail and rudder for the Picopilot as recommended nowadays?

Good luck!

/Icebear

[smitty505000]

Thanks for the advice IceBear. I gave it a bit of forward tilt, all control throws are set for conservative throws (all the way in on servo and all the way out on control surface horn).

Was able to fly this morning and she holds good flight path at half throttle with little to no pitch up and PicoPilot off. Turn on PP with 4 waypoints set and it tried to follow them but the plane is crawling with a 8-10mph headwind. Seamed to miss a couple of waypoints after it searched for a while then came and hovered overhead.

How close does it have to be to a waypoint before it goes to the next? Will it skip one if it dosnt find it in a certain amount of time?

I wish I could get a nice calm day when not at work!!!!

Smitty

[kd7ost]

The co-pilot pitch when mounted is not that critical unless it's way off. During the calibration process is when things become critical. You need to set the plane on the ground in the pitch angle you want it to maintain while in flight when you hit the aileron stick to memorize the cal setting. If the sensor unit is pitched up or down a little, if there is nose or fuselage rear in the co-pilot sensor view, it doesn't matter. Those are factored into the calibration.

Dan

[kd7ost]

How close does it have to be to a waypoint before it goes to the next? Will it skip one if it dosnt find it in a certain amount of time?
Smitty

The way UNAV sequences the waypoints is as such.

The sequencer is "in range" of waypoint once it gets within .05 miles. (264') That's not enough to sequence to the next point though. There are two accepted methods of sequencing waypoints. Arrival and departure sequencing. In an arrival type system, the plane will make the waypoint once it gets within the preset range then sequence to the next. This isn't good in small UAV's because many times the plane will turn before it gets over the way point. We fly slow and turn pretty tight compared to full size planes. Anyway, they also then incorporate departure sequencing. What this means is first you must detect the arrival range. For UNAV that means you get within the .05 mile range. Then it looks for the range to begin to increase. This supposes you've flown as close as conditions will allow to the waypoint and are now moving beyond it. Then the waypoint sequencing happens.

Naturally cross winds and what have you might prevent your plane from hitting the exact spot but those are dynamics we can't control.

If the sequencer doesn't "Make" a waypoint then it can't be departing a waypoint. It will not sequence on to the next waypoint until it makes the one loaded in the sequencer at that time. You can't fly tight, zig zag little courses as it puts the plane in an undue position to turn on a dime. Since it needs a little airspace to fly around, make sure to plot the course accordingly.

Dan

[icebear]

Smitty,

I think Dan's on to what problem might be. How far are the WP's placed? I have found that while as little as 150 meters might work you should rather go for at least 200-300 meters in a square to start testing navigation. Otherwise you may not notice when the WP's are 'made'.
What you describe sound like they are made and since the circling overhead works, your navigation probably does too.

I suppose you don't use the FMA co-pilot at all and that you are using the NAV2A Picopilot with winglevelling?

A few questions;

- I see the Garmin GPS but no Picopilot GPS in the picture - are you using the Geko 201 for both Picopilot and OSD? If you download the recorded "track" during flight you can easily compare that to the waypoints programmed in the Picopilot and see how it performed (if you have Mapsource for instance) - very useful when setting up.

- How did the altitude hold function (also visible in the Geko recording I guess)?

Like Dan indicated each WP must be "made" before the NAV unit sequences off to the next, so if the WP's are placed out enough you'd see clearly when they are passed as the plane would turn and navigate to the next WP.

Keeping my finger crossed for better weather so we can see some more progress for you soon... :D

/Bjorn

[smitty505000]

Hey guys,

Yup I think you guys are correct.

No, I am not using FMA co-pilot. Its PicoPilot NAV2A and ALT3-E. Yes I am using just the Geko 201. I tried to DL the track but I accedentally errased them at the field. I normally save after each flight then clear it and I forgot I had to press enter twice to save :( I do that so I can select each flight to view in google earth.

The alt hold seamed to work very well as the OSD showed a consistent number.

I checked google earth and measured my waypoints and I see now that it did make all the waypoints to within the 264'. My waypoints were only about 800' apart and it was a bit windy. She was fighting the wind the whole time. I will make them farther apart next time.

Very strange watching the plane fly with no hands. When the plane started hanging out over head it just kinda just sat there heading into the wind. I started having a conversation with my buddy (who was watching the live video) and he said dude where is your plane? I said right there and he said why arent you watching it? I said its not going anywhere. He then started coming up with nic-names for my plane. Stuff like SLK String Less Kite and eslwk Expensive string less wonder kite. :D

I think if we get a good calm morning, it will work great. I am really ready to do some longer legs but want to make sure all the bugs are out.

This is a lot of fun but Mondays are crule. Gotta wait all the way till Saturday again. :D

Smitty

[smitty505000]

Weather was cloudy but calm this morning and I decided to go to the field. I had spread the waypoints out so I could be sure it was following the points. Flew three times and the plane followed every wayoint on every flight. :D
I also saved the tracks and imported them into Google Earth. The alt hold looks good also but might tweek it a bit more depending on the next test.

If weather holds out tomorrow, I plan on sending it out about half mile.

Smitty

[kd7ost]

Good for you. It's nice to get those test flights under your belt.

Dan

[smitty505000]

When I got up this am there was still a bit of wind and pretty cloudy but I figured I would try to dodge the rain. :cool:

Made one flight with three legs at about .7 miles each and the plane performed very well. I got a bit nervous as it was becoming smaller and smaller but it made the first turn and was on track. The second leg was into the wind and she just kept chugging along. On the third leg it was coming back down wind and missed a point and had to circle once but no problems.

I am very pleased with the results and now plan on checking out some landmarks arround the field. Also incorporate my goggles and head tracker.

I did get about 107 glitches as shown on the inspire so I need to track down where they are coming from. I am wondering if the bergs are going into failsafe when these glitches are happening. Or if it is between the rx and the servos. I suppose I could rig something up to the rx failsafe led to detect and advance the event counter on the inspire to let me know when a failsafe has taken place.

Smitty

[kd7ost]

Smitty,

Aren't you using the failsafe to engage the plane to fly anyway? Whenever I did fully autonomous testing I turned off my transmitter and let the onboard receiver send the needed pulses while in failsafe mode.

Dan

[icebear]

Great to hear about your success Smitty - good work!

Would you mind posting a picture from the tracking? Would be nice to compare to my Viking setup which is very similar and I have tweaked it for months now.

Cheers,

Bjorn

[smitty505000]

Hey Dan, no I have been leaving the TX on in case something happens I can take over. Plus I want to be able to pan and tilt the camera.

IceBear, yes here is a photo from Google Earth.

Smitty

[smitty505000]

I did a bit of testing and found the Berg 7 rx's do glitch. They are single conversion but do some sort of filtering with an algorythm. They do not glitch very bad and would not effect control surfaces because the movement is so little. That little movement though is enough to wreak havoc on the pan and tilt unit. Also while watching the reply, the RSSI drops to nearly nothing more than a few times. A couple times at around 1/4 mile away!

I have now switched them out for my Futaba 149DP 9ch PCM rx with failsafe. I have to use the three boosters for the PicoPilot but it works. I found where to solder the RSSI wire and it works perfect also. I did the same range test (inside my apartment) as with the bergs and no glitches at all! Will have to see how it does at the field.

It has been pretty windy here in florida but I am hopping it will calm down so I can fly at least one day this weekend.

Smitty

[smitty505000]

Had two good flights this morning. Decided to fly FPV with failsafe set to turn on the autopilot and bring it home. Using my Futaba 149DP PCM RX I had very few glitches displayed on the Inspire. Took her out about 1 mile while buddy watched through binoculars. The 149DP was rock solid and never went into fs. Camera Pan and tilt was also rock solid.

Thinking about making a Pic interface that will attach to the PicoPilot waypoint prox pin and the camera shutter input pin on the inspire. This will increment the event counter once a waypoint is reached and allow visual confirmation. It could also be used to drive a servo for payload release or whatever. Its pretty simple and should only take a couple hours to make.

Smitty

[Connexxion]

Smitty,

did you notice any altitude loss during turns?

What amount of rudder deflection do you have on your TM?

I'm struggling with my plane at the moment since it loses altitude during turns.

I've even been considering to mix some UP-elevator with the rudder to maintain altitude during turns.

Btw: I've the Picopilot NAT version.

Cnnxxn

[smitty505000]

Hello Connexxion,
Mine only drops if it is turning downwind and then slowly regains alt. My mechanical throw on my elevator is about 3/8" up and down and alt-e gain set at about 25%. As for Rudder its mechanical throw is set at about 1/2" in each direction and about 50% gain.

Hope this helps

Smitty

[icebear]

Thanks for the picture Smitty - looks like you have really good tracking!

Nice idea also to use the Wprox to get WP count on the OSD!

Cheers,

Bjorn

[arocholl]

I did a bit of testing and found the Berg 7 rx's do glitch. They are single conversion but do some sort of filtering with an algorythm. They do not glitch very bad and would not effect control surfaces because the movement is so little. That little movement though is enough to wreak havoc on the pan and tilt unit. Also while watching the reply, the RSSI drops to nearly nothing more than a few times. A couple times at around 1/4 mile away!

I have now switched them out for my Futaba 149DP 9ch PCM rx with failsafe. I have to use the three boosters for the PicoPilot but it works. I found where to solder the RSSI wire and it works perfect also. I did the same range test (inside my apartment) as with the bergs and no glitches at all! Will have to see how it does at the field.

It has been pretty windy here in florida but I am hopping it will calm down so I can fly at least one day this weekend.

Smitty

Hi, can you publish (or point to a link) where there is a description on the places to hook RSSI up on receivers of several brands?