AttoPilot V1.8 in the Green 'Solution'

[Tom Harper]

AttoPilot deliveries are beginning and I got mine. Thanks Dean!

Dean's year long Beta Site efforts have paid off - this unit is very professional and complete. The terminology is clear and there are lot's of step by step instructions. It's all coming together!

My system was delivered ahead of the labels, but they will arrive tomorrow. I got some details from Mike's thread (thanks Mike). The wind is howling outside so there's no hurry to get into the air. Maybe first flights by the weekend, if the weather calms down.

The airplane is a Greensleeve motor glider wing and empenage on Mike Evan's 'Solution' profile, AP fuselage. Layout and construction were aimed at AttoPilot from the beginning. There are platforms for the GPS, Thermopile sensors, the status LED and the SD memory socket. Conduits were cut in the fuselage to accomodate the numerous wire runs. The system is compact and complete, but there is still a lot of plumbing involved. Especially when you add the camera and it's support. The model has been flown several times and the camera gear has been used for stills and video. The first two photos were taken during first flights.

The third photo is the XY sensor head. It's velcro mounted on the motor box which gives it 0-0 orientation. I gave some thought to a pylon to raise it above the prop, but for now stuck to simplicity. The indicator LED is on the lower part of the platform.

The fourth photo is the platform on the rear of the wing mount. The GPS unit has a solid mount (velcro) and a clear view of the sky. The SD socket is velcro mounted next to it - easy access!

The fifth photo is the electronics compartment. Obviously needs work. The receiver and Attopilot both rest in foam pockets. An important feature is that the receiver connectors are held in place by the top of the compartment. The AttoPilot connectors will be held in place by the compartment door - yet to be added.

So, I'm taking a week (the weather gives me no choice) to see if I can get it right the first time.

[Gary Mortimer]

Is that wing slightly smaller than the normal solution Tom??

I guess it has ailerons as well, something that the solution did'nt have??

It obviously flys, looking forward to your Atto test flight results.

Everyones talking about the stickers, are they critical and made of some special secret thing??

[Tom Harper]

Gary,

The model is smaller than a Solution. It's made of foam board instead of foam and Depron. Has ailerons on the 78", high aspect ratio wing. Flys nice.

The stickers are important because they tell you where to plug the connectors. I have some diagarms but I'll wait until I have the directions written on the case (stickers/labels).

Tom

[Gary Mortimer]

Ah I see, thanks Tom

[Tom Harper]

Drilled a 1/8" hole in the leading edge and main spar to accept the Pitot tube. Epoxied it in place and let everything sit overnight.

Drilled a 5/16" hole in a sheet of 3/32" balsa. Coated the edges with white glue and pressed it into the upper curvature of the airfoil section. Tried to do it without getting glue all over the place and was partially successful. When the glue set I went over the top of the wing with a hot iron to bond the covering to the balsa sheet.

Connected the Pitot tube and ran it and the Z lead down the wing through the ribs. Then covered the bottom with 3/32" sheet.

[airmcn_3]

Drilled a 1/8" hole in the leading edge and main spar to accept the Pitot tube. Epoxied it in place and let everything sit overnight.

Drilled a 5/16" hole in a sheet of 3/32" balsa. Coated the edges with white glue and pressed it into the upper curvature of the airfoil section. Tried to do it without getting glue all over the place and was partially successful. When the glue set I went over the top of the wing with a hot iron to bond the covering to the balsa sheet.

Connected the Pitot tube and ran it and the Z lead down the wing through the ribs. Then covered the bottom with 3/32" sheet.


Looking good Tom,

We use hot glue to mount the sensors. Makes cleaning up easy. You are going to have one tidy setup there!

Chris

[Tom Harper]

Thanks Chris.

I've got a hot glue gun around here someplace. I probably should use it.

Tom

[Tom Harper]

Finished the Z sensor and Pitot tube installation. Don't have enough of that thin green covering the wing came with, so I reverted to Monokote. Will have to redo the whole wing one of these days.

Now I have to see how much trouble it is to get all of that stuff into the fuselage conduit. I prefer minimal field assembly. Gonna have to work on that part.

[d_wheel]

Connected the Pitot tube

It's probably in the instructions somewhere, but if so I missed it. Which one is the pitot sensor?

Looking good! I'm almost ready for a test flight. Just a couple more things to clean up.

Later;

D.W.

[mhaun5]

D.W.

You can plug it into either port.

Mike

[d_wheel]

D.W.

You can plug it into either port.

Mike

I assumed that one was airspeed and the other altitude????? Oh Heck! I'm terrible at making decisions. Which shall it be..... ;) ..

Later;

D.W.

[mhaun5]

I assumed that one was airspeed and the other altitude????? Oh Heck! I'm terrible at making decisions. Which shall it be..... ;) ..

Later;

D.W.

LOL. Just so I am clear for others that read this...

On the AttoPilot itself, you have two ports for the Pitot airspeed tube (two nipples):

http://static.rcgroups.com/forums/attachments/2/4/8/4/1/7/a2328824-188-S1051250.jpg

Plug a fuel line into one of the nipples on the side of the AttoPilot case. Then take that fuel tube and run it to a brass tube or some other type of tube to collect air externally on the airplane. Make sure you leave one non-obstructed and vented. Also make sure your Pitot tube is collecting undisturbed air, as you don't want to create a situation where diverted airflow or prop wash causes issues.

In regards to the sensors, check my thread as I have pictures of all of them. You have your XY sensor (looks like an FMA sensor) that reviews the horizon, you have your Z sensor (looks like a pill with two eyes on it), and then you have your GPS unit.

Hopefully the weather is nice enough for you to fly this week. The wind here is horrible right now (40 mph!).

Mike

[d_wheel]

Hopefully the weather is nice enough for you to fly this week. The wind here is horrible right now (40 mph!).

Mike

The wind is almost that bad here. Supposed to blow all week. :( I will be getting up early each morning in hopes of finding better conditions for a test flight.

D.W.

[Tom Harper]

Looking forward to your first flight. Have you got all the gear installed?

The wind has been bad here also and the storm is headed your way. It is supposed to blow today with rain and snow.

I widened the conduit to make assembly easier. Might even transport it with the left wing in place.

Gotta go up to the big city today so won't make much progress.

[d_wheel]

Looking forward to your first flight. Have you got all the gear installed?

The wind has been bad here also and the storm is headed your way. It is supposed to blow today with rain and snow.

I widened the conduit to make assembly easier. Might even transport it with the left wing in place.

Gotta go up to the big city today so won't make much progress.

Yes. It's installed and ready to go.

The wind was already 20mph at sunup this morning with predictions for more wind and possible thunder storms. The next 5 days also have high wind in the forecast. Maybe a break this weekend.

Later;

D.W.

[Tom Harper]

Wind was 40-60 MPH yesterday. Had snow on the way back from Albuquerque then a major sand storm at the Rio Salado.

It's headed your way!

Tom

[d_wheel]

Got in a couple of tuning flights today. Nothing to brag about yet but I can see the potential. Couldn't resist trying autonomous mode at one point and because I had my speed set to slow, almost lost the aircraft when it went into a spin. Recovered maybe 20 feet above the lake! :eek: It flew fine in pilot assist mode, but the turns were verrrrry slow. I haven't figured out how to increase the turn rate yet. It might be because my transmitter rates are set pretty low. I may have to redo the linkages in order to increase the rates to a higher level. At any rate, I got a lot of data to use for the setup. Tomorrow morning is supposed to be good as far as wind is concerned so I need to get up early and try again.

Later;

D.W.

[alphacenturi]

whats the attopilot like? i am a bit of a dummy when it comes to this level of electronics. does it require any code writing/programming? or was it more of a plug and play thing, to use very layman terms?

[Tom Harper]

alpha,

AttoPilot is plug and play. No programming. You do have to enter some parameters to describe your aircraft and the waypoints.

Tom

[d_wheel]

I would say "plug and play" is a yes and no. No, you don't have write or compile any code to get in the air. However, there is a large SET file that requires one to input quite a few parameters in order to get reliable and safe flights. I may be a little slower (dumber) than the average guy, but I estimate it will take 10 or 15 flights to get Atto dialed in to my airframe. I suggest that everyone make at least 4 or 5 "tuning" flights before committing to any ambitious flights. From there, review the data from each flight and make fine adjustments to maximize efficiency and precision.

I made 2 more flights this morning before the wind increased. It was about 6mph before the first flight and 10 after the second.

All I can say at this time is WOW! AttoPilot is going to be the autopilot I have dreamed of using for many, many years!

I didn't use a waypoint file for these flights. Atto defaults to a holding pattern around the launch point when no WP.txt file is present. The aircraft held the circle path and altitude so well that I still can't believe it. Watching the live downlink was like seeing a video playing in an endless loop. It passed over the same clump of trees, fence line, small island in the edge of the lake, etc., each time around. Each time I glanced at the altitude it was within 5 feet of the last, always right at 190 feet.

Reviewing the data and recorded downlinked videos confirms a precision that is uncanny. Dean certainly did his homework and obviously wrote algorithms that work very well indeed. I can hardly wait to get a good waypoint file loaded, using the recorded data as a guideline, and go out again. Unfortunately, the forecast is for 20+ mph winds over the next few days. Oh well, guess I need to catch up on a few chores anyway....

By the way Tom, I hope you don't mind me using your thread for reports of my progress. If you would rather, I will move it to another thread.

Later;

D.W.

[icebear]

DW,

Great to hear about your progress!

I know for a fact that I'd trust you anytime with setting up a new autopilot and without knowing about Atto, I would agree that anyone should allow some tuninig flights with ANY autopilot before venturing into more ambitious projects. I think I had at least 20 or 30 flights on a test flight before attempting any of my missions...

Anyway, great to hear about your field experiences!

And, oh - great thread Tom! :)

/Björn

[dmgoedde]

Got in a couple of tuning flights today. Nothing to brag about yet but I can see the potential. Couldn't resist trying autonomous mode at one point and because I had my speed set to slow, almost lost the aircraft when it went into a spin. Recovered maybe 20 feet above the lake! :eek: It flew fine in pilot assist mode, but the turns were verrrrry slow. I haven't figured out how to increase the turn rate yet. It might be because my transmitter rates are set pretty low. I may have to redo the linkages in order to increase the rates to a higher level. At any rate, I got a lot of data to use for the setup. Tomorrow morning is supposed to be good as far as wind is concerned so I need to get up early and try again.

Later;

D.W.I apologize for my absence. Once the roll gain is set right (backed off perhaps to 75% of where it creates oscillations) then you can confidently increase the max allowed roll angle. This also assumes the pitch control is set correct. Then of course there is the Kd gain for each axis which will tame oscillations, but the Kd gains usually need to be used like spice: sparingly.

If the Pilot assist mode has sluggish turns, I would normally think that either the max roll angle is too low, or the Kp gail for roll is too low and not forcing the plane to roll target. Sounds like you got a great autonomous hold pattern and are ready for a WP.txt flight.

[Tom Harper]

D Wheel - I'm happy to have the company and I can sure use your information.

[Tom Harper]

Got the labels and the generic SET file.

This is a smallish, profile model so there's no fuselage room. That's not all bad. The wire runs are short and it all fits, but there are a still alot of cables going various places.

The wing attachment went better than expected. Turns out that the easiest procedure is to run the pitot line down to AttoPilot. The intermediate connector is too cumbersome when you are balancing the wing and fuselage.

Anyway, here's the photos. Should power up tomorrow.

Tom

[dmgoedde]

Tom, looks great. I would suggest possibly a better place for the XY sensor is in an "X" configuration on top of the fuselage boom mid way between tail feather and wing saddle. This would allow the horizon sensors to always see horizon in level flight, and minimize any deadband in attitude measurement. I am certain the location you picked is fine, but if I were there with you during the install, that is a suggestion I would make.

Good luck!
Dean

[dmgoedde]

couldn't get you on the phone... so I will add 1 more comment. The pitot tube looks too close to the prop wash. I would move it out at least 8" past edge of prop.

[d_wheel]

Sounds like you got a great autonomous hold pattern and are ready for a WP.txt flight.

Dean, is there any way to offset the center of the holding circle? The field I fly at has a rule about not flying west of the runway. During hold, Atto flies its holding circle around the spot where it was turned on. Most of the people I fly with don't mind when I fly behind them, but there are a few who are not so forgiving. If I could offset the holding circle, or designate a holding pattern I could keep it in the designated area.

Thanks for all of your help with the set file. I am going to try a waypoint file next time out. I'm sure part of my original problem was in setting the speed too slow. This made the ailerons less effective and caused the aircraft to kind of wallow along until it finally stalled and went into a spin. Increasing the speed seemed to make a huge difference. I'm sure that your suggested changes in the set file will be another positive change. I'll let you know.

Later;

D.W.

[dmgoedde]

Dean, is there any way to offset the center of the holding circle? The field I fly at has a rule about not flying west of the runway. During hold, Atto flies its holding circle around the spot where it was turned on. Most of the people I fly with don't mind when I fly behind them, but there are a few who are not so forgiving. If I could offset the holding circle, or designate a holding pattern I could keep it in the designated area.

Thanks for all of your help with the set file. I am going to try a waypoint file next time out. I'm sure part of my original problem was in setting the speed too slow. This made the ailerons less effective and caused the aircraft to kind of wallow along until it finally stalled and went into a spin. Increasing the speed seemed to make a huge difference. I'm sure that your suggested changes in the set file will be another positive change. I'll let you know.

Later;

D.W.D.W. Yes - there are two things that cause Atto to "capture" the home position for RTL loiter:
1) After GPS locks you briefly put the mode to Auto, then back to manual (back to manual for RC takeoff)
OR
2) after GPS lock and plane ground speed goes above 15 km/h (approx 10 mph)

So what you could do is once GPS locks (GPS unit LED starts to flash at 1Hz, and Atto's status LED starts flashing out the # of sats) you taxi the model out to the runway keeping the ground speed well below 10mph, then take off in manual mode.

I will probably be changing this ground speed trigger limit to make it user-configurable via te SET file. Yo might want to make it higher like 25 km/h so you can taxi more quickly without inadvertantly setting home before you want.

THIS is why the SET file is over 100 parameters now... Atto Pilot is NOT that complex. It really comes down to adjusting 3 key gains as the first step: pitch, roll, and throttle response to get a decent loiter. Once that is done you can tune secondary parameters. The ground control software is a bit buggy, but we are stamping out some small issues and will publish it to customers soon. A key feature is ability to playback the PID charts form a LOG.txt file. People don't need a modem connection to see the PID charts. By simple visual cues, you can quickly see what gain adjustments are needed for each of the 5 key parameters (pitch, roll, airspeed, altitude, and heading control). Each is displayed as a running strip chart of target and actual. Low Kp gains cause actual to lag and not hold on target, too high Kp gain causes rapid oscillation around target. Once you find the cliff to induce oscillation, back the Kp gain down to around 75% of the oscillation point, then add a small amount of Kd gain. BTW, Ki gains in Atto are 100% automatic and generally set based on the Kp gain, and the Ki reset times are fixed. I can defend this by reiterating the high accuracy Atto routinely achieves.

[dmgoedde]

I'm sure part of my original problem was in setting the speed too slow. This made the ailerons less effective and caused the aircraft to kind of wallow along until it finally stalled and went into a spin. Well, Atto automatically increases control gains as airspeed goes down. This is the airspeed based gain scheduling that sets Atto apart (among other things like SD data logging). I think (based on experience across some 20 airframes with Atto) is the low airspeed simply caused a tip stall, which the ailerons were powerless to stop. How much washout do your wings have? I have flown a number of planes, especially flying wings, with 2-3 degrees of washout, and invariably they stall in a very well controlled manner meaning the nose drops instead of a wingtip dropping.

In my latest devlopment firmware, I added a 4th control mode, which is assisted RC (stabilization), but the throttle stick is a direct pass through to the throttle meaning you can turn the motor off and pull back on the stick to hold nose high and wings level via 50Hz stabilization, and bring the plane down in a controlled stall. I did a number of assisted landings yesterday onto asphalt using this mode with a UAV company's flying wing. This is a hop skip and a jump from 100% autonomous landings. The key is that airspeed gain scheduling will automatically increase gains to maintain control at low speeds. At this point I still think it is important to have enough washout in the airframe, otherwise this won't work. In the case your describe with your plane, the wingtips might keep trying to stall.

[d_wheel]

Yes. I am sure it was a tip stall which resulted in a spin. I am testing Atto in an Easy Star that is overloaded with equipment. At over 3.5 pounds, it is a pig! It may be a tad tail heavy also. I will adjust the CG forward before the next flight.

When the aircraft stalled, it appeared that Atto was doing exactly what it was told to do. It was trying to hold the airspeed, altitude, and stay on track. However since it was flying on the edge of a stall it couldn't keep it under control. What I might need to do is decrease the maximum elevator throw in the set file. That way at slow speed it could never get the nose high enough to result in a stall.

Later;

D.W.

[Tom Harper]

Dean,

I'll look into moving the pitot tube - major operation.

Do I need a firmware update for the camera control?

On wings with no washout, how about creating washout by raising the TE of both ailerons.

[small_rcer]

D_Wheel;

With your easyStar having become somewhat 'porcine' , you may want to try using the easyGlider Pro wings. Virtually identical root section and mounting area. It has been installed by some people in the foamies group. About 18" more span and the corresponding area and the other part is the wing has ailerons.

I have the wings and joiner spar here and when time permits in the next week or so that is what I am readying for my EasyStar. Will replace the stock brushed motor with brushless at the same time.

Jim H

[Tom Harper]

I didn't want to drill more holes in the spar and LE so I just peeled off some mylar and blopped epoxy on the balsa.

[d_wheel]

D_Wheel;

With your easyStar having become somewhat 'porcine' , you may want to try using the easyGlider Pro wings. Virtually identical root section and mounting area. It has been installed by some people in the foamies group. About 18" more span and the corresponding area and the other part is the wing has ailerons.

I have the wings and joiner spar here and when time permits in the next week or so that is what I am readying for my EasyStar. Will replace the stock brushed motor with brushless at the same time.

Jim H

Thanks for the suggestion, and great timing. I flew the old pig too long today and ran the battery down. At about 20 feet up, battery power was so low that the receiver quit and it just rolled over and tried to plant itself. I'm now looking for something with more wing area and more room inside for equipment. It would be nice if Multiplex would come out with a double size easy star. That would be the perfect size for a camera plane!

Later;

D.W.

[zlite]

D_Wheel;

With your easyStar having become somewhat 'porcine' , you may want to try using the easyGlider Pro wings. Virtually identical root section and mounting area. It has been installed by some people in the foamies group. About 18" more span and the corresponding area and the other part is the wing has ailerons.


I tried that and found that it was a terrible flier. Too "short coupled" with those big wings, and aileron drag created major countersteer. It was almost uncontrollable at slow speeds.

[Tom Harper]

D Wheel,

I'd use the larger wing and tail but make a longer profile fuselage out of foam board.

Tom

[Tom Harper]

I moved the pitot tube then strengthened the hatch cover. Also changed from Tygon tubing on the barometer to clean fuel line. Tygon is too stiff.

It goes together reasonably well, but I am thinking of changes to my next design.

So, next up is initializing the SET file and the first ground test.

[albertocvr]

When attopilot is going to be available , and what about the small rtl Dean was working on ?

[small_rcer]

The counter steer or adverse Yaw that zlite refers to is common on full size gliders as well. There is a very simple and long-used solution, which I was intending to use.

The key is to have differential in the ailerons. The up going aileron must travel up, as much as 3 times the travel of the down going aileron. The high pressure on the underside of the wing causes a lot more drag on the down-going aileron. This causes the plane to rotate or yaw towards the down going aileron. So you reduce the down travel and increase the up travel.

In full size you use rudder to do a coordinated turn. Quite normal. A pain in the butt for models so you use differential.

If you are using a single channel and 'Y' adapter, then you have to hook up the aileron servos for mechanical differential. Start with a 2 to 1 ratio of up to down.

Rather than make a new fuse to have a larger tail volume coefficient, just enlarge the fin. The horizontal tail and elevator have lots of authority so they are big enough.

Jim H

[Gary Evans]

In full size you use rudder to do a coordinated turn. Quite normal. A pain in the butt for models so you use differential.

Jim H

What would the reason be to not just link the two in the radio? Simpler than separate rudder in a full sized aircraft.

[small_rcer]

From reading the discussion on AttoPilot, I believe at this time it only supports aileron by way of a 'Y' adapter. As this plane is guided by AttoPilot, using mechanical differentials on the aileron is probably the easiest and most appropriate way. Also many people still use a non-programmable radio, or gave up on the programming due to poor or unclear documentation, or one of many other reasons.

Best of all if the differentials are setup via mechanical changes, any radio at any time can be used. Also tuning the mechanical differentials are less likely to be affected by inadvertent trim settings changes or program changes on a radio.

It is also old-school in that you learn to trim the plane and better understand how and why it flies as it does.

"Old School Jim"

[Gary Evans]

Also many people still use a non-programmable radio, or gave up on the programming due to poor or unclear documentation, or one of many other reasons.

"Old School Jim"

Ya, I hear those new fangled radios are just a fad and that you can pick up them up for a song since many people are reverting back to levers and bellcranks to get things done. If only there was a mechanical autopilot.

[dmgoedde]

In full size you use rudder to do a coordinated turn. Quite normal. A pain in the butt for models so you use differential.
Jim HJim - Atto is setup to use rudder in a special way for coordinated turns.

Ailerons - used to hit and hold target roll angle. Once target is hit, it takes very little aileron input to hold that target. Aileron position is driven by roll error (Kp and Ki) and roll rate (Kd)

Rudder - NOT simply "Y" linked in code to aileron position, rather driven by target roll angle through a Kp term. So long as there is command roll angle, there will be rudder deflection to assist the turn. You can turn off rudder in non-RC mode by setting the Kp gain to zero.

I setup rudder like this to help full-house planes navigate better.

[dmgoedde]

From reading the discussion on AttoPilot, I believe at this time it only supports aileron by way of a 'Y' adapter. As this plane is guided by AttoPilot, using mechanical differentials on the aileron is probably the easiest and most appropriate way. Also many people still use a non-programmable radio, or gave up on the programming due to poor or unclear documentation, or one of many other reasons.

Best of all if the differentials are setup via mechanical changes, any radio at any time can be used. Also tuning the mechanical differentials are less likely to be affected by inadvertent trim settings changes or program changes on a radio.

It is also old-school in that you learn to trim the plane and better understand how and why it flies as it does.

"Old School Jim"Jim - by "Y" all I meant was Atto is not setup to drive two seperate aileron servos. If your model has two seperate aileron servos, then they are "Y" connected and driven by one output of Atto (the aileron output).

Because Atto properly uses rudder for coordinated turns, I don't think you need mechanical differential in the ailerons. The rudder is handled seperately from ailerons, and the control of rudder is completely different than how ailerons are driven. I tried a simple "Y" in code of rudder to ailerons at first, and got almost zero help from the rudder. Then I realized that as plane rolled over and met roll target, the rudder deflection was going back near zero like the ailerons. At that time I changed rudder deflection to be driven by roll target, rather than roll error. This way the plane maintains rudder deflection in the turn.

At all times, the elevator is used to hold target pitch, and in this way causes the model to automatically pull through turns. See my latest Vimeo for clear examples of pitch control helping pull-through on turns:
http://www.vimeo.com/3249191

[Ginger Adam]

I still run my 78" wings on the EZ* whenever I need to reduce speed or increase payload.

Adam

D_Wheel;

With your easyStar having become somewhat 'porcine' , you may want to try using the easyGlider Pro wings. Virtually identical root section and mounting area. It has been installed by some people in the foamies group. About 18" more span and the corresponding area and the other part is the wing has ailerons.

I have the wings and joiner spar here and when time permits in the next week or so that is what I am readying for my EasyStar. Will replace the stock brushed motor with brushless at the same time.

Jim H

[John O'Sullivan]

I believe what Jim was referring to was Y-connecting both ailerons to the Aileron output of the ATTO.
I usually use the aileron RX output for right aileron and Aux 6 or 7 for left aileron, which allows for independent aileron adjustments. On my Attopilot 1.6, I cannot do this and so use a Y lead in the single aileron output. I do not have any adverse yaw on my Eyespy and don't need mechanical differential.

Don't know if the Receiver outputs 1 and 6 are usable for independent ailerons (or even necessary) in version 1.8.

John

[dmgoedde]

Don't know if the Receiver outputs 1 and 6 are usable for independent ailerons (or even necessary) in version 1.8.

JohnThey are not. This is what I referred to in the post above. If I had just one more IO pin, I would gladly have seperate aileron outputs, and even software-driven differential controls, and independent servo reversing for the two ailerons, plus flaperon option. Actually, based on recent advancements, I could free up 1 IO pin and do that, however the hardware design for v1.8 is currently set in stone.

[Tom Harper]

After a bit of delay, I am back on the UAV project.

I have the IV unit mounted and have gone through the basic check out steps. Everything works fine. I even get GPS lock on my workbench with 6-7 sats. The external status LED is a blessing. You get a running monologue of what the system is up to.

The installation looks messy but it is actually pretty functional. This model was designed to have everything external and accessible. It is that!

AttoPilot has nice long leads. I definitely would not make any changes there. But, on a small model with a 3"x5" equipment bay you have a lot of extra wire. That will not be the case with the model that is under construction.

The photo is of the present installation. The IV unit is velcro mounted on the profile fuse. The two unconnected wires are IV and throttle. I checked it out with an external battery. I need to add a layer of 3/16" foam board to give a little more depth to the equipment bay.

It is getting windy but maybe the Saturday very early.

[Tom Harper]

Added 3/16" thickness to the equipment bay. Added stiffeners to the door. All of the wires fit and everything is snug.

The wind is howling with 55 mph gusts. Maybe very early tomorrow - but I have my doubts.

[Tom Harper]

By 6:45 AM the wind was above 30 MPH - whistling and howling gusts. It must be March!

No flying today.

[Tom Harper]

While the wind was howling outside I decided to complete the set up on the bench. AttoPilot 1.8 does everything on the bench that it does outside (except for the XYZ sensor). That's handy because you can even confirm the number of GPS satellites.

I connected everything and closed up the equipment bay. Then connected the flight battery - and - zip, nada. I unbuttoned the bay and checked for voltage at the rcvr. It looked good but still nothing. So, I went back to the last known good state, with an external battery. I added connectors one at a time while checking operation each time. It worked at each stage - even the last one. So, I buttoned it up again and plugged in the flight battery. It all works - I must've had something in wrong.

Took it outside and did the final check. Looks good. The thermopile sensors are very responsive. I expect some oscillation there. We'll see.

The wind has suddenly dropped, but it's still unreliable, quiet for a while then heavy gusts. But, tomorrow the forecast is for calm at last! So, hopefully tomorrow!!

[rcdoma]

Good luck on the maiden Tom!

[Tom Harper]

Thanks - due to scheduling problems it looks like Monday.

[Tom Harper]

Rained last night but was calm and a little overcast this AM.

Checked out everything - got 6, 7 then 8 satellites. Gave it throttle and took off. Made a couple of passes - one at full speed and one at minimum speed then climbed to altitude and switched to AUTO.

It turned and headed for the waypoint. Dean says the LOG looks like it was sluggish on controls, but it looked great to me. It found the first WP and made a 180 as expected.

Then, the rod that joins the panels on a Greensleeves wing pulled through the bottom of the ribs on the left panel and the wing folded. That's not good. The wing panel floated away and the rest of the model spun in.

All the AttoPilot parts survived and the model can be salvaged. But, it sure ruined a good flying day. I will replace the steel rod with 3/8" sq spruce spar to better distribute the load.

Tom

[airmcn_3]

Rained last night but was calm and a little overcast this AM.

Checked out everything - got 6, 7 then 8 satellites. Gave it throttle and took off. Made a couple of passes - one at full speed and one at minimum speed then climbed to altitude and switched to AUTO.

It turned and headed for the waypoint. Dean says the LOG looks like it was sluggish on controls, but it looked great to me. It found the first WP and made a 180 as expected.

Then, the rod that joins the panels on a Greensleeves wing pulled through the bottom of the ribs on the left panel and the wing folded. That's not good. The wing panel floated away and the rest of the model spun in.

All the AttoPilot parts survived and the model can be salvaged. But, it sure ruined a good flying day. I will replace the steel rod with 3/8" sq spruce spar to better distribute the load.

Tom


OUCH.... Sorry to hear about this Tom.

Good luck on your next one.

[Gary Mortimer]

Well thats just not cricket, poor show from johnny construction foreigner.

I can almost feel that credit card going out and buying a RTF something or other.

Bad luck old man.

[Tom Harper]

Thanks guys - I've lost my zeal for the 'Solution' camera mount. Not yet sure how I'll handle it on the replacement. Sparky Paul said "don't put anything on the front of the model that you are not willing to lose".

In defense of the foreigners, the wing was carrying more weight than the glider. However, if I use the wire spar technique, I'll definitely encase it in a tube. A hole through the ribs is not sufficient. The second photo is the spruce joiner. I cut square holes through the first 3 ribs so the joiner is 12" long - that should do it.

The first photo is the new model under construction. Bigger and better with the camera mount raised to the thrust line. Rather than rush it, I'll go ahead and rebuild the Green one.

Tom

[FAKHREALAM]

Rained last night but was calm and a little overcast this AM.

Checked out everything - got 6, 7 then 8 satellites. Gave it throttle and took off. Made a couple of passes - one at full speed and one at minimum speed then climbed to altitude and switched to AUTO.

It turned and headed for the waypoint. Dean says the LOG looks like it was sluggish on controls, but it looked great to me. It found the first WP and made a 180 as expected.

Then, the rod that joins the panels on a Greensleeves wing pulled through the bottom of the ribs on the left panel and the wing folded. That's not good. The wing panel floated away and the rest of the model spun in.

Tom

I think you stress out your plane during auto flight, the result of that stressing is the center wing pulled out. The common casue is due to plane were exposed to high G during the auto flight. I think by checking the G's we can save our planes (these are just RC toys) from this kind of mishaps.

Fakhre Alam

[John O'Sullivan]

Tom:
I doubt if 3/8" square spruce is sufficient to take the wing loads. If you can get some 3/8" square birch or maple the load carrying capability will be doubled.

On re-reading your post it appears that the failure was in the wing rather than the rod itself.
I guess that if "the rod that joins the panels on a Greensleeves wing pulled through the bottom of the ribs", that the rod itself was not the problem, but it was the wing tube integrity with the wing structure. it is important that the wing tube is integrated to the wing with real ply ribs (not just Lite-ply).

John

[Tom Harper]

John,

There was no wing tube. The steel rod just ran through the ribs. The rod was fine, the ribs failed.

I agree that maple would be stronger, but it would far exceed the strength of the rest of the structure. I think spruce will be more compatible.

Tom

[Tom Harper]

I built a new fuselage with 4 layers of 1/2" foam board. That makes the equipment bay deeper so everything fits comfortably. The equipment bay hatch is secured with two servo mounting screws. Changed to a pusher configuration and raised the camera mount to near the thrust line. The camera mount encloses the camera in plywood and foam. For testing I have a balsa box with lead weights inside.


The wing has the 3/8" sq spruce spar described above and 1/4"x1/8" diagonals between ribs. Removed the light weight green film and covered the structure with white Monokote.

You can get at everything from the outside. The equipment bay stays secured. Auw came out a little less than before.


Specs:

Span - 6.6 ft
Area - 4 sq ft
AR - 10.5:1
AUW 4.5 lbs

17.75 ounces per sq ft

Measured power - 258 watts

[airmcn_3]

I built a new fuselage with 4 layers of 1/2" foam board. That makes the equipment bay deeper so everything fits comfortably. The equipment bay hatch is secured with two servo mounting screws. Changed to a pusher configuration and raised the camera mount to near the thrust line. The camera mount encloses the camera in plywood and foam. For testing I have a balsa box with lead weights inside.


The wing has the 3/8" sq spruce spar described above and 1/4"x1/8" diagonals between ribs. Removed the light weight green film and covered the structure with white Monokote.

You can get at everything from the outside. The equipment bay stays secured. Auw came out a little less than before.


Specs:

Span - 6.6 ft
Area - 4 sq ft
AR - 10.5:1
AUW 4.5 lbs

17.75 ounces per sq ft

Measured power - 258 watts


Looks good Tom,

You may want to make sure the rear sensor on the XY head can see past the GPS. We have just about made it a ritual to mount the XY head about 1" up with a piece of foam or balsa.

Anyhow you got that new bird together quick! Good luck with the flight.

Chris

[Tom Harper]

Chris,

Thanks - I've considered that but wondered if it would just be cosmetic.Easy enough to do.

Tom

[dmgoedde]

Nah - don't worry about it. MrBlucor's 8' Telemaster has the XY sensor mounted on top of fuselage in the fore/aft/left/right rotation with GPS right in front of it and that thick clark-Y airfoil blocking about 10 degrees of view from level. His pitch control is just fine.

[Tom Harper]

Here we are ready to go again. There is a wind warning of 45MPH gales for today, so we headed for the field early. No wind at the surface but there was a very ominous cloud over M mountain.

The new layout is much better than the last one. Everything outside, easy wing attachment. It all checked out except the right aileron - I must have pulled the connector loose during assembly. Every AP model I've built in the last 3 years had a single aileron, so it should be good enough for this one.

With a 7 foot span and 17 oz loading this is not a 3D performer, but the takeoff and climb was smooth and ample for AP. Made a couple of passes on RC and Assisted RC (Modes 0 and 1) then gained altitude and switched to full AUTO (no WP file present).

It was inside the radius of the loiter circle, so it turned West for a bit then North as it picked up the path of the circle. After a couple of circles we switched to assisted mode (1) then to RC and brought it down.

Checking the LOG file, I see that the ESC cut off just after 2AH of battery power was consumed. My eBay, el cheapo specials are rated at 2.2 AH so they did OK. I'll buy some more.

Very satisfactory flight!

[Gary Mortimer]

Thats not a cloud.........

Nice one Tom.

How have you found adjusting the gains?? Or is that something that you are now doing with the data in?

[Tom Harper]

Gary,

I made some changes based on Dean's critique of the last flight. It will take time to gain cofidence fondling the gain values.

The guidance looks good, but I assume Dean will have some further suggestions. The throttle hunts during full Auto, so I'll reduce the throttle gain.

On this flight I was still working with the obvious - like the wing stayed on this time!

Tom

[Gary Evans]

Thats a lenticular cloud indicating high wind speed and a standing wave at high altitude.

[Tom Harper]

Sounds ominous!

It was dead calm on the ground, but the log file indicates ~10 mph winds aloft.

[dmgoedde]

Gary,

I made some changes based on Dean's critique of the last flight. It will take time to gain cofidence fondling the gain values.

The guidance looks good, but I assume Dean will have some further suggestions. The throttle hunts during full Auto, so I'll reduce the throttle gain.

On this flight I was still working with the obvious - like the wing stayed on this time!

TomHi Tom - A quick playback of the flight in the GCS PID window will confirm that throttle gain is too high if the actual airspeed averages on target but otherwise oscillates around it. I just got back from a UAV startup in Georgia, and their throttle gain was also about 2X too high in the initial flight. The airspeed was on target on average but oscillated +/- 10 km/h with a period of about 5 seconds. Same thing for MrBlucor's 8' Telemaster. In both cases we dropped the gain to about 1/2 of initial trial value and viola the airspeed and throttle were smooth.

Mountains AND clouds??? I thought "hobby autopilots" with their thermopiles were crippled in these conditions. And to boot, your loiter circle is VERY round... what's up with some people talking trash about thermopiles?

[Tom Harper]

Overcast and a big, nasty mountain didn't seem to have any bad effects.

The circle looks great to me! The throttle gain is 30. I'll back it off to 15.

Will try a simple WP flight down the access road later this week.

Tom

[Tom Harper]

Hmmm... servo was unplugged. Could have happened while mounting the wing.

Put a bit of glue (rtv) on the plugs. Used some 1/8 x 3/8 balsa to make a hatch frame and covered it with clear packing tape. If it happens again it can be reconnected in the field.

[d_wheel]

Very satisfactory flight!

Congratulations. Feels good to see it working doesn't it..

Later;

D.W.

[Tom Harper]

Yeah, I like this part!

[Tom Harper]

Removed the ballast box and mounted the camera. It just bolts on with 1/4:20 nylon screw into the tripod mount. Two layers of foam keep it aligned.

[Tom Harper]

Added the linear regulator with heat sink. The camera runs off of the LiPo without going through the ESC. Everything is buttoned up and operating.

The camera is an Olympus C-50. I'll use that for the learning process.


Tom

[Mick Molloy]

Hmmm... servo was unplugged. Could have happened while mounting the wing.

Put a bit of glue (rtv) on the plugs. Used some 1/8 x 3/8 balsa to make a hatch frame and covered it with clear packing tape. If it happens again it can be reconnected in the field.

Heat shrink over the join works a treat and is easy to remove when required

[Tom Harper]

Mick,

Thanks, I didn't think of that. I have some large diameter stuff that should work.

Tom

[Tom Harper]

Howling, 50 mph winds again!

Maybe by Saturday.

[aperture]

a drop of thick CA on the mating edge (after connected) on one side does the job for servo extensions, and can be scratched off with a fingernail later to remove. Got that suggestion from another member here somewhere...

[dmgoedde]

A dab of hot glue works very well also. Tenacious, but can be pulled off with a needle nose pliers later.

[dmgoedde]

Howling, 50 mph winds again!

Maybe by Saturday.If you had a flying wing (like mine), you could be out flying in the 50mph wind with your Atto.

I added 2 important things to the code yesterday:
1) Min ground speed. Even if airspeed is on target but ground speed starts to drop below 15 km/h, Atto will crank in extra power to not let ground speed get too low. This worked fabulous in Gilbert AZ yesterday where the winds were reported to be 37 mph. At end of rectangular flight plan my wing was downwind 500m, and normally I'd have to either set the min airspeed high to keep ground speed positive, OR take manual control if I saw the plane was going too slow on ground track. Yesterday the plane steadily crawled back against the wind. Of course your plane has to have the power to penetrate, but now you can leave the Min Airpseed in SET quite low and not worry about penetration in the wind.
2) Added exponential gain to pitch and roll. Through analysis of LOG, I realized that the required gain for level flight is somewhat less than gain required for those 30 or 40 degree banking turns. Same goes for pitch. In the past it has been a compromise of choosing a high enough gain that will hold 30 degree roll but also not cause too much oscllation at level flight. Now the gains at both endpoints are set independently, and yesterday's flights (yes, in 35-40mph winds) were smoother than a no-wind Ketchup Lake flight with a motor glider.

[mhaun5]

If you had a flying wing (like mine), you could be out flying in the 50mph wind with your Atto.

I added 2 important things to the code yesterday:
1) Min ground speed. Even if airspeed is on target but ground speed starts to drop below 15 km/h, Atto will crank in extra power to not let ground speed get too low. This worked fabulous in Gilbert AZ yesterday where the winds were reported to be 37 mph. At end of rectangular flight plan my wing was downwind 500m, and normally I'd have to either set the min airspeed high to keep ground speed positive, OR take manual control if I saw the plane was going too slow on ground track. Yesterday the plane steadily crawled back against the wind. Of course your plane has to have the power to penetrate, but now you can leave the Min Airpseed in SET quite low and not worry about penetration in the wind.
2) Added exponential gain to pitch and roll. Through analysis of LOG, I realized that the required gain for level flight is somewhat less than gain required for those 30 or 40 degree banking turns. Same goes for pitch. In the past it has been a compromise of choosing a high enough gain that will hold 30 degree roll but also not cause too much oscllation at level flight. Now the gains at both endpoints are set independently, and yesterday's flights (yes, in 35-40mph winds) were smoother than a no-wind Lake Ketchless flight with a motor glider.

Dean,

I am actually thinking about building another Airframe to use. You seem to favor the Wings... What are the characteristics that you like about these wings over say, a high-wing trainer of some sort?

Thanks,
Mike

[dmgoedde]

Dean,

I am actually thinking about building another Airframe to use. You seem to favor the Wings... What are the characteristics that you like about these wings over say, a high-wing trainer of some sort?

Thanks,
MikeNumerous:
1) High resistance to wind and wind gusts. I always favor aileron planes with no dihedral wings as low and close to the C.G. as possible. wings are a major source of drag, and if they are far above the C.G. (like in a high wing trainer) then gust of wind can generate torque moment and knock the craft around. If instead entire craft's center of pressure is line up near the C.G., then all a gust of wind can do is cause translation of position, not so much a torque moment. Buffeting - YES, roll and pitch disturbances - not so much
2) More efficient OK, written like that it is a generalization... but without a fuselage and seperate tail feathers it should, and seems to be, much more efficient. I routinely see my flying wings able to cruise and hold altitude on < 20 Watts/Lb... THAT'S RIGHT. For example, I have a 3.3 Lb flying wing (with Clark-Y looking airfoil nonetheless!) that loiters at 65 km/h (41 mph) with median power of 50 Watts, which is 15 Watts/Lb. On a 3s Lipo that's only around 4.5 Amps. My flight pack in that plane is 4400 mAh and I have safely used 4300 without puffing the pack... so theoretical max flight time is approaching 1 hour.
3) More compact, less delicate As a sinlge object, rather than wings + fuselage + tail feathers (that can get damaged and broken off) I find flying wing's compactness very convenient. My standard wing these days is 48" span and about 12" chord. This easily fits into the back of my Jeep.
4) Minimal/No assembly at the field - Nothing to assemble at the field. Put batteries in, get GPS lock in 15 seconds, and I'm in the air. Always < 1 minute from getting out of the car to having plane in the air and in Autonomous cranking away to the 1st waypoint. This would be very convenient for SAR (Search and Rescue).

I don't have patience these days for wasting time, and wings are in my opinion 10x better than anything else out there. If you want compact and don't mind a slight efficiency hit, then go with a low aspect ration stubby wing. You still have plenty of efficeiency. If you need really high duration, go with high aspect ration setup, possibly with plug in wing panels. You can't lose.

[Tom Harper]

What motor and prop and planform - any taper or sweep?

Was that it in #2307?

Looks like about 20% taper (.8 C at tips) and maybe 15 deg sweep and elevons.

[ios]

Numerous:
1) High resistance to wind and wind gusts. I always favor aileron planes with no dihedral wings as low and close to the C.G. as possible. wings are a major source of drag, and if they are far above the C.G. (like in a high wing trainer) then gust of wind can generate torque moment and knock the craft around. If instead entire craft's center of pressure is line up near the C.G., then all a gust of wind can do is cause translation of position, not so much a torque moment. Buffeting - YES, roll and pitch disturbances - not so much
2) More efficient OK, written like that it is a generalization... but without a fuselage and seperate tail feathers it should, and seems to be, much more efficient. I routinely see my flying wings able to cruise and hold altitude on < 20 Watts/Lb... THAT'S RIGHT. For example, I have a 3.3 Lb flying wing (with Clark-Y looking airfoil nonetheless!) that loiters at 65 km/h (41 mph) with median power of 50 Watts, which is 15 Watts/Lb. On a 3s Lipo that's only around 4.5 Amps. My flight pack in that plane is 4400 mAh and I have safely used 4300 without puffing the pack... so theoretical max flight time is approaching 1 hour.
3) More compact, less delicate As a sinlge object, rather than wings + fuselage + tail feathers (that can get damaged and broken off) I find flying wing's compactness very convenient. My standard wing these days is 48" span and about 12" chord. This easily fits into the back of my Jeep.
4) Minimal/No assembly at the field - Nothing to assemble at the field. Put batteries in, get GPS lock in 15 seconds, and I'm in the air. Always < 1 minute from getting out of the car to having plane in the air and in Autonomous cranking away to the 1st waypoint. This would be very convenient for SAR (Search and Rescue).

I don't have patience these days for wasting time, and wings are in my opinion 10x better than anything else out there. If you want compact and don't mind a slight efficiency hit, then go with a low aspect ration stubby wing. You still have plenty of efficeiency. If you need really high duration, go with high aspect ration setup, possibly with plug in wing panels. You can't lose.

I totally agree on all points above :)

I've been known to arrive at the flying field, be in the air within a few minutes, have my fun for around 20 minutes, land and be gone before the gas guys have even had a chance to fuel up!

I think the highest winds I've flown in were around 35-40 knots on a few occations, and I wouldn't hesitate to do it again - in fact its fun to buzz the pits when no one else is willing to fly!.

Wings have fantastic wind penetration, and I suppose what you loose in (very) slightly less efficiency, you gain in the overall lighter overall airframe weight.

Dean - have you thought of selling an AttoRTF 'package' - consisting of an attopilot installed onto a small flying wing? It could be the hottest christmas present of the year!! :)

[hobbyboy]

If you had a flying wing (like mine), you could be out flying in the 50mph wind with your Atto.

I added 2 important things to the code yesterday:
1) Min ground speed. Even if airspeed is on target but ground speed starts to drop below 15 km/h, Atto will crank in extra power to not let ground speed get too low. This worked fabulous in Gilbert AZ yesterday where the winds were reported to be 37 mph. At end of rectangular flight plan my wing was downwind 500m, and normally I'd have to either set the min airspeed high to keep ground speed positive, OR take manual control if I saw the plane was going too slow on ground track. Yesterday the plane steadily crawled back against the wind. Of course your plane has to have the power to penetrate, but now you can leave the Min Airpseed in SET quite low and not worry about penetration in the wind.
2) Added exponential gain to pitch and roll. Through analysis of LOG, I realized that the required gain for level flight is somewhat less than gain required for those 30 or 40 degree banking turns. Same goes for pitch. In the past it has been a compromise of choosing a high enough gain that will hold 30 degree roll but also not cause too much oscllation at level flight. Now the gains at both endpoints are set independently, and yesterday's flights (yes, in 35-40mph winds) were smoother than a no-wind Ketchup Lake flight with a motor glider.



Dean,
How do we get that code?

Bob

[Gary Mortimer]

Um

I'm getting 40 to 60 minutes with a smaller battery and more traditional airframe.

It might not fly well in 20 knots but would I get the shot in those conditions?

More importantly can I mount and protect cameras on a flying wing better than the norm??

Is the market for your product in a robust device that flys in whatever or somebody looking at the forecast and spotting a reasonable day and going and getting the shots.

Just chucking it out there

G

Oh I would like to be able to sell a ready to go system that could stooge around for 30 minutes taking shots within LOS

Thats right I already have an airframe that can.

[CenTexFlyer]

I think I might be able to contribute a bit on the "wing" subject.... :)

We have been working with the flying wing concept for photo/AP/UA/data collection since about the turn of the century, and have worked very hard to disprove all the naysayers (Too fast, to unstable, tip stalls, hard to orient, blah, blah, blah). Not only have we done this through our airframe design and configuration, we've taken it to the field and proven it. In other words, we "talk the talk and walk the walk". No, you will not do it as well with a Zagi/Clique/WingWarrior airfoil..... they are made for speed and manuverability which is a completely different animal. Dean has taken a page right out of our book when it comes to explaining "Why a wing?". I don't know what his position was before, but I know there as trip out to a California desert that might have done some convincing. I will add a few other things that I believe make a wing a great platform for lofting sensors :

1: Versatility - As long as you maintain some semblence of a C of G, you can load just about anything you want into the pod. This has become even less critical with the Atto's ability to keep up with flight parameters!

2: Wide envelope of operation - you can lazily fly a properly set up wing on a calm blue bird day, or crank it up into a windy, cold miserable environment to get the data you need. You only need an area the size of a small yard to get out of and a convenient tree or bush (environmental decelerators :cool: ) to put it into for landing.

3. Rugged - properly designed and built, these things will take a beating and keep flying. In 5 years of HARD field flying, we have NEVER been unable to complete a mission even after some really bad crashes! That's a record that stands on its' own merit because we have flown thousands of SAR missions.

4. Matched systems - Now that Dean has had the opportunity to fly one, there are things being done to the Atto that maximizes not only the Atto, but the platform it is flying in. This fine tuning is complimenting the capabilities of each system while at the same time correcting some of the shortcomings of each.

Most certainly, it is not a panacea. However, for the widest range of payload and operating variables, it's an easy decision.

Gene
www.rpflightsystems.com

[Tom Harper]

Centex,

So what's the best section - Clark Y with some up elevon?

Tom

[CenTexFlyer]

Ancient Chinese secret! :)

It's actually a blended airfoil that took as long time to get the right flight characteristics on a tapered wing. I don't know what you'd call it, but it's in my computer as profile!

Gene

[icebear]

Um

I'm getting 40 to 60 minutes with a smaller battery and more traditional airframe.

It might not fly well in 20 knots but would I get the shot in those conditions?

More importantly can I mount and protect cameras on a flying wing better than the norm??

Is the market for your product in a robust device that flys in whatever or somebody looking at the forecast and spotting a reasonable day and going and getting the shots.

Just chucking it out there

G

Oh I would like to be able to sell a ready to go system that could stooge around for 30 minutes taking shots within LOS

Thats right I already have an airframe that can.

Interesting subject - I have no experience with flying wings, but I am looking for a simple, robust airframe that could reach 1 hr duration, manage moderate winds (up to 20 mph) and carry a payload of 8 oz and still be transportable and bot weigh a ton (i e less than 3 lb's).

That said, my simple SuperMiss polyhedral motor glider probably meets the above specs but it is not as durable and obviously not super efficient. It does fly well in winds up to 15 mph even though it cruises at 20-25 mph.

Will do lots of tests with different airframes this season and see if I can find something useful...

End of hijacking your thread Tom - I am following this thread with interest!

/icebear

[dmgoedde]

Dean,
How do we get that code?

BobI will e-mail out firmware updates to customers, and you burn it with a USB-to-Serial adapter (your micro-SD card coming with your Atto has a .pdf describing the in-field firmware update process and required hardware. I will send a broadcast e-mail of details when I publish the latest firmware (from yesterday). I killed the deadband stuff from the code, because what I was really looking for was the expoenential gains. I also cleaned up a couple things bugging me in the code. A complete "tech report" pdf will accompany the latest firmware.

It sucks badly to be stuck doing business and non-technology stuff. The last 2 days I said 'screw it' and spent both days at the field flying and working on code both days. I have a sunburn to prove it. Really fast development by coding and flying in iterations. Heaps of LOG data from each flight helps tremendously.

Example change: If you forget to set home position before flying, Atto used to look at GPS ground speed, and if it got above 25 km/h, then Atto said to itself "oops they forgot to set home position but are flying... I will now grab home". Problem is sometime upon first lock the GPS position is settling down and the reported ground speed (on occasion) can be above 25 km/h... I have only had this happen about 1 time / 100 flights, but it is a safety risk because motor can come on once you put plane to Auto briefly to set home... Atto has already blown past the initial stuff and thinks plane is flying. The change is that I now ignore GPS and instead look at altitude barometer, and if it shows > 20m AGL, THEN it auto-locks the home position if you forgot. This is much more robust.

I also fixed up some things in the gorund test setup code to make it work better when teaching elevon mixing. I also added to the ground setup procedure capture of the tiny offset in Pitot barometer, and you can put this in the SET file for much more precise airspeeds < 20 km/h.

[Tom Harper]

Dean,

Nice to be trimming a finished product.

It sure is fun to watch in the air. The bank angle is dramatic when it hits a WP.