New project to be flying this weekend! :: AttoPilot flying!

[mhaun5]

Woot!

I have everything I need to start flying this weekend autonomously. Yet, I am still missing a few items from the overall setup that I will install sometime next week or the following. I spy the following in my picture:

(1) MPX TwinStar 2
(4) HiTec HS-81 servos
(2) Scorpion 2215-18 motors
(2) Scorpion 35A ESC's
(1) TanicPack 5000mAh 3S Lipo
(1) TanicPack 2200mAH 3S Lipo (not shown)
(1) iCharger 106B-Plus
(1) Canon PowerShot SD1100IS with remote USB trigger
(1) 8GB SDHC card + CHDK framework (neat stuff!)
(1) 14mz transmitter (with 72mhz receiver)
(1) 2.4Ghz 1watt transmitter
(1) 2.4Ghz receiver
(2) 9XTend 1watt data modems
(*) Extension, Y Cables, Dean connectors
(1) Dell 830 Laptop Ground Control Station & Software
(4) APC-E 8X6E props
(2) CNC Custom made motor mounts
(1) AttoPilot 1.8 Navigation system & Sensors
(1) Castle Creations BEC (not shown)
(1) Parallax Prop Plug & Cable

What I don't have yet that I am still reviewing:

(1) 2 Channel Mini Video Switch V2
(1) OSD Camera (DPC-480A OSD Camera)
(1) Mounting bracket with 180o Pan / Tilt servos

Even though I don't have the above three items, I still will get everything going this weekend, and trimmed out on Saturday / Sunday. The build will start on Friday.

I will post videos and pictures as I go along.

Mike

[gumush]

it looks great , please daily update for progress.

Any detail for 2.4ghz transmitter and receiver ( not 14mz ) ?

How to plan to connect remote trigger to canon ?

9XTend 1watt data modems ? ( what for ? )

"(1) Mounting bracket with 180o Pan / Tilt servos" you can use http://www.servocity.com/html/pan_tilt_roll___accessories.html for camera.

[mhaun5]

Woot! The build process has started.

I picked Dean Goedde up from the airport at 10:00 AM, and we made a few last minute runs for some miscellanous parts.

We are in the process of dry fitting everything together, as we need to make sure we have a good balance and access to all the equipment. One of the issues with the MPX Elapor foam based TwinStar 2 plane is that it has to be glued together; making it very difficult to jam everything in.

Anyhow, here are a few pictures of the AttoPilot v1.8!

---

gumush:

The 2.4Ghz 1 watt is a Lawmate unit. I purchased it from Black Widow AV

The Trigger unit to the Canon Camera SD1100IS will be done through the AttoPilot trigger system. But the way it actually works is via a USB trigger, via power of 3.3 - 5.2 volts; using CHDK custom framework.

The 9xTend modems are for the AttoPilot's ground control telemetry unit. It will be able to have uplink and downlink.

Thanks,
Mike

[Vince53]

Looks good. You'll probably need to hack out some of the foam to get all that stuff mounted. It is much easier to fit it all up and cut it before joining the two fuse halves.

I also have a Multiplex Twinstar II which is anxiously awaiting an Attopilot. I won't be using any video or downlink for telemtry but I will have a camera. Last time I took it up it was just taped onto the fuse but it'll be interesting to see what you come up with.

I mounted my esc's right behind the motors, which saved some space, but I need to shorten up my connecting leads. I set everyting up with plugs for a two piece wing but when I clean it up I'll tape the wing together to reduce the wiring mess.

I received an Amateur Radio beacon from Bigredbee.com today. It is used with a directional antenna for finding lost models. I actually lost a plane with a beacon on it. I could tell the direction it went but lost the signal before long. I think it may have come down in one of our many area lakes.

I'm currently covering a SIG LT-25 which will probably wind up as the aerial photography platform. Slightly modified of course; 6" wheels with very tall landing gear so I can mount the camera under the plane.

Vince

[Frisk151]

Mike,

Great choice in motors and autopilot. I have a Scorpion HKII 2221-10 in my Rave R/C heli running off a Castle 45 HV speed controller and it has incredible power on a 4S pack.

What brand video switch are you using? I'm going to be running multiple cameras on my airplane and I was going to hack up a standard 4 channel switch, but would much rather have a micro or small version.

I read above you picked up "the man" (Dean) from the airport. Talk about special AttoPilot delivery... I must have missed that shipping option... ;)
I may do a build thread of my own on the Telemaster (has retracts and under-slung camera gimbal).

Look forward to more pictures on your build..

David

[icebear]

Mike,

Nice to see the 1.8 and your equipment.
Getting the designer over for the installation is really great customer service.

Good luck this weekend and looking forward to seeing the results! :)

/Björn

[mhaun5]

Thanks Vince, David, Bjorn!

I have tons of updates to give. I am in the process of unloading pictures during the build. But here is a teaser: the plane flies autonomously!

[mhaun5]

We worked through the night on Friday, finally quitting around 2:30 AM CST. Most of the time we spent on the build was design and layout, trying to figure out how to lay out all the equipment in the most efficient manner.

One thing I wanted to do was relocate the servos from the front of the MPX TwinStar 2 to the mid/rear. This would help with the overall needed CG weight, as well as give more room for us to work with up front.

There were a few moments that caused us to focus more time during the build on:

- Remote USB Camera trigger switch
- Canon SD1100IS fitment
- Canon CHDK Custom Framework
- 9xTend SMA to MMCX connector (lack of a rubber ducky antenna)

We first began laying out where we would want to place the Canon camera. Although it is small in size, it is still bulky enough to make it a pain in the ass for a clean fit. Cut after cut, we finally got a decent placement of the camera and USB cord.

Once we knew we could adequately house the camera, I moved on to looking at the CHDK framework. I wrote a custom script and placed it on the SD card. Made a few tweaks, and then went to check how much voltage would trigger the shutter. We found it was in the ~4 volts range. This was more than the default AttoPilot output trigger logic (3.3v), so we used a simple Transistor to power 5v - which was perfect in making the camera take pictures on demand.

One of the last issues we ran into was that I had the wrong rubber ducky 900 MHz connectors (I had MMCX, needed SMA). Dean happened to have an extra 9xTend 1watt transmitter with him, so I stole that and quickly glued it into the plane (making it mine :p ). I then used my MMCX antenna, and routed it up through the tail. By cutting some of the extra compartments out, we were able to put the gear in various places.

Overall, we stopped working about 2:30 AM CST. We had just glued the two halves together, with a small shopping list of items needed for the next day. Our delirious minds figured we would sleep for 6-7 hours, and pick it up at 8-9:00 AM CST; finishing everything by Noon!

... Oh were we wrong!

[airmcn_3]

We worked through the night on Friday, finally quitting around 2:30 AM CST. Most of the time we spent on the build was design and layout, trying to figure out how to lay out all the equipment in the most efficient manner.

One thing I wanted to do was relocate the servos from the front of the MPX TwinStar 2 to the mid/rear. This would help with the overall needed CG weight, as well as give more room for us to work with up front.

There were a few moments that caused us to focus more time during the build on:

- Remote USB Camera trigger switch
- Canon SD1100IS fitment
- Canon CHDK Custom Framework
- 9xTend SMA to MMCX connector (lack of a rubber ducky antenna)

We first began laying out where we would want to place the Canon camera. Although it is small in size, it is still bulky enough to make it a pain in the ass for a clean fit. Cut after cut, we finally got a decent placement of the camera and USB cord.

Once we knew we could adequately house the camera, I moved on to looking at the CHDK framework. I wrote a custom script and placed it on the SD card. Made a few tweaks, and then went to check how much voltage would trigger the shutter. We found it was in the ~4 volts range. This was more than the default AttoPilot output trigger logic (3.3v), so we used a simple Transistor to power 5v - which was perfect in making the camera take pictures on demand.

One of the last issues we ran into was that I had the wrong rubber ducky 900 MHz connectors (I had MMCX, needed SMA). Dean happened to have an extra 9xTend 1watt transmitter with him, so I stole that and quickly glued it into the plane (making it mine :p ). I then used my MMCX antenna, and routed it up through the tail. By cutting some of the extra compartments out, we were able to put the gear in various places.

Overall, we stopped working about 2:30 AM CST. We had just glued the two halves together, with a small shopping list of items needed for the next day. Our delirious minds figured we would sleep for 6-7 hours, and pick it up at 8-9:00 AM CST; finishing everything by Noon!

... Oh were we wrong!


Did you get the RX working?

Chris

[ios]

We worked through the night on Friday, finally quitting around 2:30 AM CST. Most of the time we spent on the build was design and layout, trying to figure out how to lay out all the equipment in the most efficient manner.

One thing I wanted to do was relocate the servos from the front of the MPX TwinStar 2 to the mid/rear. This would help with the overall needed CG weight, as well as give more room for us to work with up front.

There were a few moments that caused us to focus more time during the build on:

- Remote USB Camera trigger switch
- Canon SD1100IS fitment
- Canon CHDK Custom Framework
- 9xTend SMA to MMCX connector (lack of a rubber ducky antenna)

We first began laying out where we would want to place the Canon camera. Although it is small in size, it is still bulky enough to make it a pain in the ass for a clean fit. Cut after cut, we finally got a decent placement of the camera and USB cord.

Once we knew we could adequately house the camera, I moved on to looking at the CHDK framework. I wrote a custom script and placed it on the SD card. Made a few tweaks, and then went to check how much voltage would trigger the shutter. We found it was in the ~4 volts range. This was more than the default AttoPilot output trigger logic (3.3v), so we used a simple Transistor to power 5v - which was perfect in making the camera take pictures on demand.

One of the last issues we ran into was that I had the wrong rubber ducky 900 MHz connectors (I had MMCX, needed SMA). Dean happened to have an extra 9xTend 1watt transmitter with him, so I stole that and quickly glued it into the plane (making it mine :p ). I then used my MMCX antenna, and routed it up through the tail. By cutting some of the extra compartments out, we were able to put the gear in various places.

Overall, we stopped working about 2:30 AM CST. We had just glued the two halves together, with a small shopping list of items needed for the next day. Our delirious minds figured we would sleep for 6-7 hours, and pick it up at 8-9:00 AM CST; finishing everything by Noon!

... Oh were we wrong!


Looks great :)

Whens the first flight !?

Nick

[mhaun5]

For whatever reason, we figured we would simply throw the wing on and go. We needed some extra wire to connect and Y together the ESC's, which we would pick up on Saturday morning. After executing our plan, we were done shopping by 9:30 AM. Back to work!

The day started out well, things were going together nicely. Then the exhaustion set in, and we slowed down. Jimmy Prouty stopped by mid morning and offered a hand. We took it! He got us back on track and helped with the final push of getting the wiring and Scorpion ESC's programmed correctly.

We did a few things different than what the manual states. Specifically here, we glued the wing spar in-place, making it one entire wing surface. The reason we did this was mostly because we wanted to make it fly. I have enough space to transport without having to worry about completely breaking it down.

Once we had the wing and fuse completed, we started hooking up the AttoPilot. The install and configuration went extremely quick in comparison to everything else we had been doing in the past two days. In about 45 minutes, we had everything squared away - picking up Sat signals in the basement of my house.

We went through the entire ground test mode, and found we had to reverse a few items (rudder, aileron and elevator). We also set a few gains in the configuration file.

Fighting the time, it was now 5:00 PM CST and we hadn't even flown yet! We quickly scrambled to the field, and went to take it for its maiden flight. On the maiden power up, the NFL Tuck rule went into place; the result is this:

http://www.corvettedyno.com/attopilot/S1051280.JPG

[mhaun5]

If anyone is familiar with United States football (National Football League); there is this thing called the "Tuck Rule". If you don't know what it is, Google it. This is what occurred on our first attempt to launch.

We rushed to the field while we had some daylight left. We did all of the preflight checks, verified the AttoPilot and Manual modes were working correctly. We powered up the motors, Dean cocked his arm back to launch the plane with throttle around ~50% and on the forward motion a loud PLOP and odd pitched noise was made.

Lucky enough, Dean was able to hold onto the plan and shut it down before it all hit the ground. In review, the prop just broke while it was at 50% throttle! This caused the prop to become way off balance and fly off all together. A piece slapped into the side of the frame, and one nearly hit Dean! :)

The prop must have been defective from the factory, for there was no reason for the blade to just break. I came prepared though, and had an extra one to use. After a short pit stop to replace the Prop, we were back in order.

Then the moment came. We ran all the pre-flight checks again, and went for a launch. Dean through the plane and off it went! A few clicks of trim, else everything was flying near perfect. For our first test, we through into AttoPilot mode and it worked near perfectly! It began a nice loiter circle around our launch site.

It handled and flew just great. This was the best launch in Dean's AttoPilot airfoil thus far on its first flight. Needless to say, we all were super excited. We let it fly for 15 or so minutes, and landed to check the data. We made a few changes, putting in some Photo-Grid waypoints, and launched everything again. Another perfect flight!

We have video of it all, and it will be posted tomorrow. We are going to make a few tweaks to some gains and other settings, but it is great to know we are up in running!

In hindsight, if the plane would have been already built and trimmed, the entire install would have taken about an hour! This unit is awesome!

More to come later.

Mike

[mhaun5]

Did you get the RX working?

Chris

Chris,

No I did not. Thanks for trying to help though! I ended up using the 2.4 Ghz module from my large R/C Gas Helicopter. I need a DSC cable to make it work. :(

Thanks,
Mike

[airmcn_3]

Chris,

No I did not. Thanks for trying to help though! I ended up using the 2.4 Ghz module from my large R/C Gas Helicopter. I need a DSC cable to make it work. :(

Thanks,
Mike


Mike,

I am very glad to hear you guys got it worked out; I was feeling a bit bad for you. It’s great to hear it only took a couple of flight to tune it up; it really shows how far the development has come. We are looking very forward to seeing your GE tracks and video.

DSC cable........ Remember I said I never used my 72MHz RX? Well....... I always wondered what that cable was for sitting in my TX case :D

Great job guys!

Chris

[mhaun5]

Here is the first autonomous flight!

[mhaun5]

Mike,

I am very glad to hear you guys got it worked out; I was feeling a bit bad for you. It’s great to hear it only took a couple of flight to tune it up; it really shows how far the development has come. We are looking very forward to seeing your GE tracks and video.

DSC cable........ Remember I said I never used my 72MHz RX? Well....... I always wondered what that cable was for sitting in my TX case :D

Great job guys!

Chris

Thanks Chris!

In regards to making it fly out of the box, we have yet to change anything as it relates to the AttoPilot. Rather, it was pure WayPoint stuff as Dean wanted to show how he could do Photo Patterns. But you are right, this is really shows how far it has come, and proves it is an out of the box item!

We plan on making a few tweaks to settle everything in, but it is great as is!

Mike

[Gary Mortimer]

Well done on getting Dean to install an Atto into a fantastic airframe!!

No more ugly wings for him!

Looking forward to hearing more.

G

[Gary Mortimer]

Why not mount that antenna right on the fin??

Maybe it would have less of an aerodynamic effect there??

Or make up a wire equiv that fits in the foam???

But thats another subject.

[Tom Harper]

Congratulations! Looks like a tight pattern.

What was the flying speed?

[Connexxion]

I see alot of twisted wires,did they come with atto or did you twist them yourself?

Why did you leaf the servocables untwisted?

These untwisted cables could pick up interference and screw up your day.

[icebear]

Congratulations guys - great to see the build and the quick results!

Lookig forward to see some video too :D!

/Bjorn

[d_wheel]

mhaun5,

Thanks to our local post office, my atto didn't make it yesterday. Tracking shows that it left the post office but it never made it to my house.. Looks like tomorrow should be the day. However, with your help maybe I can get a start on the installation. If you will give me the measurements (h/w/l) I can go ahead and make the necessary cuts in my airframe.

TIA

D.W.

[mhaun5]

Dale,

Sorry to hear about the Post Office! That is no fun. :(

In regards to the final demensions, it will take me a little while to get to the unit. I have it under the wing, deep in the fuse. I can take it apart later, but I have no promises.

Tom just got his, so shoot him a quick PM to see if he can measure it for you; else let me know. I know Tom is currently in the process of mounting everything as well.

Thanks,
Mike

[mhaun5]

Just wanted to give an update on our ventures here...

Today we went out flying again, and had more great runs! I made my own custom paths to follow a non-used access road by some corn fields. The results came out great!

We were also able to use Google Earth v5.0 to stream LIVE via NMEA sentences from my 9xTend modem to the ground station. We were able to do this without the need of Goops or any other software (Thanks Google!).

The flight weather today in Kansas changed quite a bit from yesterday (72o F Sunny, ~14 mph winds) to today (37o F overcast, ~15 mph). We made no changes for today's flight in our Setup file. We made to more flights today, and both ended perfectly.

After reviewing the data, we made some very small tweaks to try for next time. The hope is that these tweaks make the plane follow our intended goals that much closer than they already do.

I attempted to make some videos from my crappy camera, and I will post one here. In addition, I hope to have the other videos from yesterday’s flight here soon.

Here is one video of our efforts:

http://www.corvettedyno.com/attopilot/S1051287.AVI

Thanks,
Mike

[Tom Harper]

Quick measurement comes up with: 2 1/4 X 1 11/16 X 11/16

Tom

[airmcn_3]

How much range are you expecting the video to transmit?

[markeydoo]

Mike
Great work
So with all the nuts and bolts what was the final cost of the project? Did you have to add or subtract form your original shopping list?
Thanks
Mark

[mhaun5]

I see alot of twisted wires,did they come with atto or did you twist them yourself?

Why did you leaf the servocables untwisted?

These untwisted cables could pick up interference and screw up your day.

Connexxion,

Dean had all of his wires twisted with the AttoPilot unit. The other wires were twisted as well to keep RF noise down. In regards to why the two servos didn't have their wires twisted, it really comes down to being tired and overlooking them.

In the four flights we have had thus far, we have not had any issues with these two servos. At this point, we are just crossing our fingers. :)


Congratulations guys - great to see the build and the quick results!

Lookig forward to see some video too !

/Bjorn

Bjorn,

Thanks! The last video was horrible, so I appologize for that. The sun was glaring, and my camcorder's battery wasn't charged. I used the standard record option on the point-and-shoot doo-dad I had. As you can tell, it was crap. Even though, you can see when Dean flips it into autonomous mode; the plane takes off to the desired speed (60 KM/H), Altitude (125 meters), and waypoint I had set.

My goal is to get the camcorder stuff going, and then take video of the plane flying various patterns; as well as setup / install points of interest. My next main goal is to have the camera and trigger system working perfectly. I need to alter some of the setting in the camera for shutter speed, etc.

On a side note, we did get video (I haven't seen it yet) from Jimmy on its maiden flight. Hopefully I can get that published here soon.


Congratulations! Looks like a tight pattern.

What was the flying speed?

Tom,

Thanks for putting the numbers out there! I still need to look at an ideal cruising target speed, but we currently have it cruising at 60 KM/H. In the flight logs we have seen it go as high as 90 KM/H (manual flight), and stall around the 25 KM/h range.


How much range are you expecting the video to transmit?

Chris,

I don't have a specific distance in mind, but I have the necessary wattage to do something like 8-10 miles LOS in perfect conditions. In reality, it will be much less than that. One thing that I have been working with Dean on is to make me a tracking antenna. :)

I plan on purchasing the correct antennas this week, working out the details for a more mobile base station.

Thanks,
Mike

[airmcn_3]

Thanks for the info, can you tell me what system and where you got it?

Chris

[mhaun5]

Mike
Great work
So with all the nuts and bolts what was the final cost of the project? Did you have to add or subtract form your original shopping list?
Thanks
Mark

Mark,

The project is never done! :)

Excluding the AttoPilot unit, I spent something like:

~$150.00 for the plane, servos, and rx.
~$350.00 for the upgraded ESC's, Motors, BEC, Props, Switches, Wire & Heatshrink, glue, and other misc nuts & bolts.
~$200.00 for the lipo batteries
~$225.00 for the 2.4Ghz transmitter & receiver
~$390.00 for the 9xTend data modems
~$160.00 for the Canon SD1100 IS camera

In all reality, I could have gone much cheaper and accomplished the same thing. Much of the stuff I was really nickle and dimed on; as it has been a little while since I have built anything. Specifically, the electric power system, as I come from the Gas/Nitro world. But, I wanted a smaller foamy plane that was easy to tear down and flown by just about anyone.

Mike

[mhaun5]

Thanks for the info, can you tell me what system and where you got it?

Chris

Chris,

I had purchased the video downlink system from BlackWidowAV.com originally. I believe they are Lawmate brand, and it does work very well. I have ranged tested this by getting in the car and driving off. I was impressed with the results.

They have it for sale here:
http://www.blackwidowav.com/bwav2401000components.html

I believe you can get the same stuff in other places cheaper (i.e. RangeVideo); but anyhow.

Mike

[markeydoo]

Thanks Mike
Are they selling the AttoPilot yet?

[Jack Crossfire]

The video was the first time we've heard Dean's voice. Unfortunately it's virtually impossible to get good video of autopilots unless you make videos for a living. There are just too many things going on. We also have another series of strong storms coming to you from the west coast. Regarding the cost question, he mentioned $800 a long time ago.

[airmcn_3]

Chris,

I had purchased the video downlink system from BlackWidowAV.com originally. I believe they are Lawmate brand, and it does work very well. I have ranged tested this by getting in the car and driving off. I was impressed with the results.

They have it for sale here:
http://www.blackwidowav.com/bwav2401000components.html

I believe you can get the same stuff in other places cheaper (i.e. RangeVideo); but anyhow.

Mike


Cool thanks, I have a 2.4 GHz 19dbi patch antenna sitting here collecting dust, may have to try the lawmate unit out.

Here is the link I found. http://www.nghobbies.com/cart/index.php?main_page=product_info&products_id=182

Is this the right one?

Chris

[mhaun5]

The video was the first time we've heard Dean's voice. Unfortunately it's virtually impossible to get good video of autopilots unless you make videos for a living. There are just too many things going on. We also have another series of strong storms coming to you from the west coast. Regarding the cost question, he mentioned $800 a long time ago.

Yes.

Yes. It is difficult as heck (even more so when you have a stupid point & click camera, not an actual camcorder with an eye-hole to watch what you want to video!) when the sun is out, everything is going on around you, and excitement is running through the veins.

Dean is selling the AttoPilot for $800.00.

Thanks for the heads up on the storms. :(

Mike

[mhaun5]

Cool thanks, I have a 2.4 GHz 19dbi patch antenna sitting here collecting dust, may have to try the lawmate unit out.

Here is the link I found. http://www.nghobbies.com/cart/index.php?main_page=product_info&products_id=182

Is this the right one?

Chris

Chris,

That would be the one! Let me know how the Patch works out.

I am sure it would be better than our quick & dirty tape job to the laptop (see attachment). :) :)

Mike

[airmcn_3]

Chris,

That would be the one! Let me know how the Patch works out.

I am sure it would be better than our quick & dirty tape job to the laptop (see attachment). :) :)

Mike

LOL, I thought that was the way it was supposed to be..... :D. I will defiantly let you know how well it works out.

Chris

[Gary Mortimer]

Would have loved to watch the video but 30mb eish, not easy over a flakey cellphone setup.

Even Vimeo or YouTube is tricky, but I sometimes endure the long download times.

Any idea from Dean if hes still around, the total airborne time for all Atto units now?? It should be possible to detect from all the log files. In fact it would be nice if it wrote its own logbook.

Working out power used and time in the air and then filing it. Would make complying with any log making rules our CAA comes up with easier.

Cheers

G

Heres the sort of image you can expect from your Twinstar when you get it tuned. From yesterday in Scotland.

[tekrunner]

Hey I like those radio modems. They're so small. Would you mind telling me where you got them? Thanks.

Tekrunner

[airmcn_3]

Hey I like those radio modems. They're so small. Would you mind telling me where you got them? Thanks.

Tekrunner


http://www.digi.com/products/

[mhaun5]

Hey I like those radio modems. They're so small. Would you mind telling me where you got them? Thanks.

Tekrunner

TekRunner,

As Chris indicated, you can get them from DigiKey. Most of the Maxstream products are very small, and depending on your needs they can get even smaller. Check here: http://dkc3.digikey.com/PDF/T091/P0532.pdf

The modems I am using are the 1 Watt 9xTend (XT09-SI / XT09-MI). They are 71.755 mm x 40.64 mm x 8.99 mm (LxWxH). Look at the little-bitty XBee / Xbee pros; they are 1/2 the size of the 9xTend modems.

Thanks,
Mike

[mhaun5]

I was reviewing the data I had received from one of the autonomous flights, and broke it down. I only evaluated a segment of flight, as I was really looking for something stable to see how my overall MPA went..

For some background, here is the power system I am using:

Scorpion 2215-18
Scorpion 35a ESC
APC 8x6E props
3s 5000 mAh battery

I have tons of data that I can analyze, but for a referenced example, I have selected two points along my flight path. In evaluating the data:

http://static.rcgroups.com/forums/attachments/2/4/8/4/1/7/a2335818-55-Graphs1.jpg

http://static.rcgroups.com/forums/attachments/2/4/8/4/1/7/a2335817-63-Graphs2.jpg

http://static.rcgroups.com/forums/attachments/2/4/8/4/1/7/a2335816-215-Graph3.jpg

http://static.rcgroups.com/forums/attachments/2/4/8/4/1/7/a2335819-21-ImageAnalysis.jpg


- I maintained a decent and level flight, with a variance of +/- 4 meters.
- I maintained an average in-flight airspeed of 61.36 km/h, with a variance of +/-4.1 km/h.
- I averaged 107.54 watts of usage during this distance of flying.
- I averaged 10.53 Amps of usage during this distance of flying.
- At this airspeed while flying into crosswind (~16 km/h), I was averaging ~7.5 km/Ah, or 4.66 miles per Amp Hour of battery usage.

As you can see many factors come into play here; what is your desired airspeed, how much crosswind is the plane trying to overcome, etc... In this case, I was still very happy with the performance figures. Averaging about 61.36 km/h (38 MPH) while earning 7.5 km/h (4.66 miles / Ah) of MPA. On top of that, I didn't really hit that high of Amps draw.

Mike

[small_rcer]

Mike and Dean;

When testing with a twin you could have some interesting scenarios. Could you test the TwinStar in single motor flight. Could you rig up the camera trigger to turn off one of the motors and see if Atto can maintain flight path and up to speed?

Other test I would like to see is if you could disable an aileron servo in flight and can Atto fly the plane?

Part of the appeal of the TwinStar is the redundancy of 2 motors. It is not unheard of to have an ESC failure or overheat causing a shutdown of a motor. With the TwinStar and Atto, if the plane can still fly, when crippled, then it is an even bigger reason to have Atto in use for critical applications.

The TwinStar can be loaded up to a fairly heavy weight for long endurance flight. In such circumstances a robust autopilot that can maintain flight with equipment failures is a real asset.

Jim H

[Gary Mortimer]

Flipping good suggestions Jim hear hear.

I have flown twinstar single engine, just because I wanted to see how it would go and it was fine.

If power and aileron went from the same side I doubt it could work. Although just leveling the working aileron and making the aircraft rudder elevator only would help. In fact in any aircraft with ailerons subject to failure that could be a save.

Of course it should also send out a mayday and the small fire engines should leave their shed.

The appeal of the twinstar for me was the ability to put a camera clear of prop views and a fair old lifting capacity.

It has proved to be a work horse.

The camera now is gallows mounted, looks ugly as anything but works very well!

Cheers

G

[boyisabird]

Mike-
Great to see data as I am a bit of a data nerd... but one criticism... What are the units on the X-axis. Is this 5 minutes of flight or 5 seconds? Kinda makes a big difference.

Also, I don't think I've seen the orientation of the camera on the Twinstar. Is it forward looking or down?

Thanks again for all the updates
Dave

[Tom Harper]

Mike,

Thanks for the data. It looks very tight. What was the wind like during this flight?

Tom

[mhaun5]

Mike-
Great to see data as I am a bit of a data nerd... but one criticism... What are the units on the X-axis. Is this 5 minutes of flight or 5 seconds? Kinda makes a big difference.

Also, I don't think I've seen the orientation of the camera on the Twinstar. Is it forward looking or down?

Thanks again for all the updates
Dave
Dave,

Thanks for the feedback. The units of measure across the X-Axis are row points retrieved from Excel. I believe they were recorded at 5 Hz intervals, or roughly every .2 seconds (I will check when I get home and look at the data again). I will have to go back and look at the data (not with me here at work), but the flight segment that I evaluated was about 7 seconds or so of data total. I know this isn't very much, but I was looking for a level and consistent period of flight to ultimately pull in MPA (miles per Amp/Hour).

When I get home tonight, I can grab the entire Autonomous run, and put the same charts against it. Then we can see how the gains responded (allowing me to fine tune them further) for Actual vs. Intended with the entire flight. In addition, there is another piece of data I didn't show here, which was the total mAh usage for the flight. It records this value at every Hz interval.

In regards to the Camera, I took many pictures and thought I had them in this thread. This was one of the more complex things we put inside the plane, so I have no clue how I left this out. I will have to dig back through the pictures again tonight and post them. But to answer your question, it is mounted inside the nose of the airplane (directly under the GPS module) to snap photos down.

The in-flight live video will have a camera that will be in the cockpit area of the TwinStar, mounted to a pan/tilt system. This will allow for FPV flying, eye in the sky controls when Autopilot is engaged.

Thanks,
Mike

[mhaun5]

Mike,

Thanks for the data. It looks very tight. What was the wind like during this flight?

Tom

Thanks!

Weather reports showed 12 - 17 MPH winds that day. During the flight, I would say it was between 12-14 MPH.

Mike

[mhaun5]

Mike and Dean;

When testing with a twin you could have some interesting scenarios. Could you test the TwinStar in single motor flight. Could you rig up the camera trigger to turn off one of the motors and see if Atto can maintain flight path and up to speed?

Other test I would like to see is if you could disable an aileron servo in flight and can Atto fly the plane?

Part of the appeal of the TwinStar is the redundancy of 2 motors. It is not unheard of to have an ESC failure or overheat causing a shutdown of a motor. With the TwinStar and Atto, if the plane can still fly, when crippled, then it is an even bigger reason to have Atto in use for critical applications.

The TwinStar can be loaded up to a fairly heavy weight for long endurance flight. In such circumstances a robust autopilot that can maintain flight with equipment failures is a real asset.

Jim H

Jim,

Thanks for the feedback and thoughts.

In regards to the items of question:

"Could you test the TwinStar in single motor flight?"

I plan on ordering another TwinStar to practice general flights with, and I wouldn't mind trying this on that airplane (without the AttoPilot and other equipment). This will tell us how the TwinStar flies with one motor in general; not so much about the AttoPilot itself.

Technically speaking, I believe the AttoPilot to be irrelevant in that situation. From a foundation point of view, the AttoPilot (while in Autonomous flight) adjusts flight characteristics based on various inputs: XY Sensor, Z Sensor, GPS Data, and Airspeed.

If the airplane was flying level and one of the motors gave out causing it to lose airspeed, the AttoPilot would then increase throttle to compensate for the loss of targeted airspeed. If it started to pitch or roll, the AttoPilot would then counter these values using all necessary control mechanisms to hit its target flight parameters. This would be like encountering a gust of crosswind.

Ultimately, if the TwinStar can fly with one engine, AttoPilot would be able to control it.

"Other test I would like to see is if you could disable an aileron servo in flight and can Atto fly the plane?"

In speaking with Dean, the AttoPilot flies much better when Ailerons are equipped on an airfoil; but Dean has had AttoPilot flying on airframes without Ailerons. So in the event both Ailerons were inactive, I believe the AttoPilot would fly as if it were equipped without; continuously making corrections to hit target parameters.

What I do not know is how AttoPilot would truly behave if one of the Ailerons went out, with another one still active. My guess is that AttoPilot would make an adjustment (intending to use both like normal), to only have that adjustment not fly in conjunction with the targeted parameters. Then it would continuously try to make adjustments until it reached its target. How this actually plays out, I will let someone else try this. :)

Mike

[FlyingStyroMan]

"Ultimately, if the TwinStar can fly with one engine, AttoPilot would be able to control it."

Just thinking out loud here....

Can your bird (loaded at max weight) maintain altitude on say 40% power on both engines? (40 % is just a total "pulled it out of my tail" number)

If it can then please ignore what I'm about to say, but if it can't, well, I might be thinking about Vmc rollover issues.

Do you know your approximate stall speed?

[Ginger Adam]

FWIW - the Twinstar2 on one engine just spins merrily into the ground :)

You should make sure the ESCs are set for Soft LVC (if the setting is available).

Adam

[small_rcer]

This is good to know. Maybe one will have to mount the engines with out-thrust (toe out in car terms ) to combat this. Rudder authority may have to be increased. I know of the stock props coming off in flight. I also know of prop collets coming off single motor planes, so testing and modifying the TwinStar for engine out flight might be prudent.

Dare I ask how you know the TwinStar spins merrily into the ground??

Jim H

[Ginger Adam]

It happened a few times before I set the LVC to soft - one motor will always quit before the other. Unless it's close enough for you to hear the pitch change it's not always clear what's happening, the plane just rolls over (quickly !) and goes into a spin. I'm only running with very low power plants so a hot brushless setup would likely be worse.

It will recover straight away if you throttle off, but if flying FPV or UAV you may not realise what's happening, and any sort of automated throttle (autopilot, or altitude hold) would probably lead to disaster if it prevented you throttling off.

I never did try 'nursing' the plane along at low throttle on one motor - maybe this is possible with some offset rudder/ailerons ?

Adam

ps Gary posted a while back that his TS flew fine on a single motor, so maybe my troubles are caused by my setup? :confused:

[tekrunner]

TekRunner,

As Chris indicated, you can get them from DigiKey. Most of the Maxstream products are very small, and depending on your needs they can get even smaller. Check here: http://dkc3.digikey.com/PDF/T091/P0532.pdf

The modems I am using are the 1 Watt 9xTend (XT09-SI / XT09-MI). They are 71.755 mm x 40.64 mm x 8.99 mm (LxWxH). Look at the little-bitty XBee / Xbee pros; they are 1/2 the size of the 9xTend modems.

Thanks,
Mike



Hey Mike,
Thanks for cluing me in on the radio modems. If you don't mind telling me, what's the difference between the XTO9-SI / XT09-MI ?

Thanks,
tekrunner

[dmgoedde]

Hey Mike,
Thanks for cluing me in on the radio modems. If you don't mind telling me, what's the difference between the XTO9-SI / XT09-MI ?

Thanks,
tekrunner-SI is SMA connector, -MI is the MMCX RF connector (very small snap on type)

[Gary Mortimer]

I have to confess on thinking about it, I flew single engine on the stock setup.

Then just lots of rudder required.

It was not an aircraft that you would have a lot of weight on either, a return to home and land situation.

I threw a prop a couple of times and once just gave it a go for kicks and giggles.

But it flew.

The brushless setup might have way too much asymetric thrust, tis true.

But there must be a low throttle setting that would work, if you got there quick enough.

I have never had a reduction in power in one side first, with my brushless setup.

Guess I have not tried hard enough for duration!!

Its a flipping winning airframe though!!

Perhaps we should nail the finer points over a beer at the RCAP meet Adam.

G

[brakar]

Mike,

I understand that you made a setup between a Canon PowerShot SD1100IS with CHDK on a (remote) SD-card and Atto. I think this means that you have (at least) two different softwares running on two different platforms - Atto software on Atto hardware and CHDK on the cameras internal processor.

Would you be kind and explain how this setup works, and how the two platforms communicates?

brakar

[mhaun5]

Hey all..

Sorry for the lack of updates to this thread. I promise I will do a lot of updates this weekend. It is my intention to get out and fly early in the morning before bad weather kicks in here; else I will just post some pictures and hope for more flights next week.

I plan to update the thread with pictures showing the mounting location of the Canon SD1100IS along with photos from the air! :)

brakar,

The reason I selected a Canon camera is due to the CHDK framework (Canon Hack Development Kit). This allows me to write custom code - or use the framework out of the box to further control items within the camera that would otherwise be unavailable. Such as:


RAW - CHDK can record raw files, giving you access to every bit of data the sensor saw, without compression or processing. Raw files can be manipulated on the camera, or processed on your PC. CHDK also has experimental support for the open DNG raw standard.

Override Camera parameters - Exposures from 64s to 1/60.000s with flash sync. Full manual or priority control over exposure, aperture, ISO and focus.

Bracketing - Bracketing is supported for exposure, aperture, ISO, and even focus.

Video Overrides - Control the quality or bitrate of video, or change it on the fly.

Scripting - Control CHDK and camera features using ubasic and Lua scripts. Enables time lapse, motion detection, advanced bracketing, and much more. Many user-written scripts are available on the forum and wiki.

Motion detection - Trigger exposure in response to motion, fast enough to catch lightning.

Edge overlay - Detect the edges in a scene, and display them later. Ideal for timelapses, stop-motion, stereography and much more.

Live Histogram - CHDK includes a customizable, live histogram display, like those typically found on more expensive cameras.

Zebra-Mode - Displays under and overexposure areas live on the screen.

GRIDS - Create custom grids and display whichever one suits your shooting conditions.

DOF Calculator - Display detailed DOF information on the screen.

USB remote - Simple DIY remote allows you to control your camera remotely.

Basically, you place the CHDK in a folder on the SD Card. The framework is loaded into RAM on the camera, so it isn't permanently writing over the firmware. But, you can set it up so that it actually loads every time the camera is powered on. This is how I have mine setup; each time I power the camera it loads the CHKD framework and executes the remote trigger script.

How it is all hooked up from a hardware perspective, I have a lead coming from the ATTOPilot tigger system, going to a USB cable. Based on waypoint (or distance/time triggers), the ATTOPilot sends a +5v signal. This +5v causes the camera to take a picture (based on the framework and custom script I wrote).

Therefore, I simply power the camera on before launch. I press the "picture" button - which starts the script to run in a loop mode (watching for a +5v power signal via USB to engage the shutter). Then at the selected waypoints, it will take a picture. :)

I will post some pictures this weekend.

Thanks,
Mike

[dmgoedde]

Small clarification: Atto's trigger, if set to a log controller is 3.3V high, 0V low. You can of course specify which state is active and which is rest. In my case with Olympus FE-230 low is active and high is rest. In Mike's case, high is the active and low is rest.

Anyway - the clarification is that Mike and I found his Canon camera would not trigger with a 3.3V logic high. So we made up a simple dongle between Atto and the camera's micro-USB which has a Radio Shack NPN transistor in the familiar TO-92 package, base is driven by the 3.3V logic from Atto, and the transistor swithes 5V from the servo rail. Therefore, Atto is now controlling a 5V logic signal to the camera's USB port as per the CHDK framework.

[brakar]

Thanks for the update and the clarification. Nice work-around with the transistor to activate the camera trigger. Would have been even more elegant to use CHDK to change the trigger settings in the camera - to allow Atto to trigger it directly with the 3.3V current. (Guess this might be difficult though. The trigger setting might be hardwired in the camera or the settings could be unaccessable via CHDK).

brakar

[mhaun5]

Yesterday was really windy and overcast, as we had a strong storm come through Kansas. It then poored rain all evening, soiling the ground today. Around 4:45 PM CST, the sun came out and it warmed up to 50 degrees. It was still windy though, but it had started to settle in wind gusts.

Looking outside, I decided I wanted to try and fly. I was eager to try the Photo Mission I had created; putting caution aside. :)

Flight #1

After getting everything setup in the field, I went to make my first flight of the day. Wind was switching directions from NorthWest to North with an occasional breeze from North East. I decided to launch into the wind, North. This is when the "bad" occured. :(

After powering up to 3/4 throttle, I hand launched. It started soaring like normal before the the plane rolled to the left, with the right wing going verticle. This pointed the the nose toward the ground. I wasn't that high off the ground when this occured; but I was able to get it leveled before nosing into the ground. It struck at a 45o angle or so, before rolling futher and throwing the leading edge of the right wing into the ground.

Then the tumble started. It didn't last long, but a few end-over-ends, it rested on its belly.

Damage Report:

- Bent & Mud filled Petit tube
- Two broken wing clips
- Mild bend in the right wing
- Out of round wing bolt holes (wing doesn't secure firmly, as it there is some out-of-round with where the bolt secures into).

Instead of taking everything home, I decided I was going to fly this weekend! With time running out, I grabbed my packaging tape and went to work. It didn't require much, as the wing clip was 1/2 there. I just taped it together. The other good thing was that when we assembled the plane, we made the wing one-piece.

In regards to the Petit tube, I just cut about 1/2 inch off the front. I then used a round body cutting tool to clean out the cut lines.

Flight #2

After powering everything on again, I decided to launch NorthWest. No wind gusts, and I was good to go. This flight went pretty well. I swapped over to AutoPilot after gaining some altitude, and away it went.

It flew well (still need to fine-tune some of the roll angle's and gain); more importantly it took pictures!

I will post the flight log data later when I get it all extracted. Here were a few shots it took.. I need to play with the camera settings, as this was just on the default Auto. There were a few blury pictures, but not too bad. :)

[ALtitudeap]

did you have this set to take pictures via atto triggers?

[mhaun5]

did you have this set to take pictures via atto triggers?

Yes.

Within my WP.txt file, I told it to take pictures at specific spots. Those photos are them. :)

[Tom Harper]

It rained here last night - I'm trying to decide whether to try a flight this morning.

Photos look good - What was your altitude?

[Gary Mortimer]

You have certainly got to get the speed up with a twinstar before you start manouvering.

Not so bad if like you its brushless, but I still keep low, build speed and then climb.

Hooray for tape!!!

[dmgoedde]

Photos look good - What was your altitude?I'd guess 100 to 150 meters based on my experience.

[mhaun5]

It rained here last night - I'm trying to decide whether to try a flight this morning.

Photos look good - What was your altitude?

Tom,

I had ATTOPilot set to fly at 125m AGL. Last flight, it was +/- 5 meters throughout the whole flight.

The photos came out OK. I need to adjust the shutter aperture to 1/1,000 or so. This will help out with some of the blury pictures.

Thanks,
Mike

[Lazman]

Could this system be used in a Helicopter?

Thanks,
Laszlo

[dmgoedde]

Could this system be used in a Helicopter?

Thanks,
LaszloNope, 1) it has no magnetometer, 2) related to #1, it operates under assumption of positive ground speed at all times and that aircraft body heading is more or less in line with ground track heading.

[Irv]

mhaun5 Tom,

Tom,

Does the AttoPilot system include the pitot tube? If not - how are they constructed?

Have you tried a 910MHZ video transmitter with the 900mhz spread spectrum modem modules?

Any special installation problems with the altitude pressure sensor?

Thanks.

[dmgoedde]

Let me jump in here in addition to the other guys
Does the AttoPilot system include the pitot tube? If not - how are they constructed?Per instruction manual you supply the tubing, and 1/8" OD or ID has been found to work well, you can use brass, plastic, or carbon fiber tubing. You mount the pitot square into the airstream and well outside the prop wash, and connect it to the Atto's Pitot sensor with clean fuel tubing.Have you tried a 910MHZ video transmitter with the 900mhz spread spectrum modem modules?Bad idea... the "900mhz spread spectrum " I think you mean ISM band, which extends up to 928 MHz if I am not mistaken.Any special installation problems with the altitude pressure sensor?Not applicable, as there is no user installation of this. This 2nd pressure sensor is hidden inside the AttoPilot control unit, and is distinct from the airspeed differential pressure sensor.

[mhaun5]

I put my 30th flight up today. As usual, it flew great. I have been really focusing on the Camera system more than the autopilot itself. In fact, I don't think I have made any changes to Atto since flight #8 or so. Rather, I send it up take about 50 pictures, land and adjust CHDK scripts.

It is a sweet thing!

Mike