The polar parameters might make a difference, especially the K and CD0 parameters. The wiki has a guide to estimating these. However limitations in the logic using airspeed are probably more significant. Are you using the TECS synthetic airspeed i.e. is parameter TECS_SYNAIRSPEED set to 1?
Re. detecting thermals from powered flight, this is in testing currently, I hope to have a beta out for general testing soon. It will work roughly as you describe but it will use TECS_SINK_MIN and TRIM_THROTTLE rather than add an extra parameter.
Thank you for the suggestions Samuel Tabor, I will look it over and see if I can estimate the K and CD0 parameters. I do not have the plane accessible to check at this moment, but I have never touched any parameters related to TECS or airspeed. Whatever the default settings of TECS_SYNAIRSPEED are is what my gliders and power plane is set to.
I assume you will be letting us know here when the next SOAR update is out. We are in really good thermal season now and I will be happy to update my glider and test whatever you come up with !!!
@Samuel_Tabor Thanks, my fault. The version is Arduplane V4.1.0dev (9e1f03c9). I found SOAR_THML_BANK so I assume I have the latest beta software found under: https://firmware.ardupilot.org/Plane/beta/MatekF765-Wing/arduplane.apj feb 23.
Getting out of the sink with optimal speed, as explained in your great video would be an improvement for sure. I am doing this now by hand mostly in FBWA. As we are restricted to LOS I have to leave the thermal searching for the next one.
Hello, I have put the latest version 4.1 Dev, and I find some changes, among them I cannot find the SOAR_ENABLE_CH Parameter, and I do not know if the operating philosophy has changed.
Is this Parameter no longer necessary?
On the other hand and since I am trying to test soaring with a Volantex ASW-28 Sailboat, the same as the one I think you use, you could pass me the .Param, in this way it would make the task much easier.
Thank you very much for all that you are contributing.
Comment and also, that currently I am not going to mount the Pitot had, and in this ASW-28, I have mounted a small FC Omnibus F4 Pro.
If all goes well and I hope so, I have another Sailboat prepared with Soaring, a Reichard Sirius, in which I do have the ARSPD mounted.
Yes you can find my parameter file for the ASW28 here. Iāve also designed a few 3D printed parts that Iāll add, the most useful is probably a 3D printed tailfin including pitot mount.
SOAR_ENABLE_CH has been replaced by setting the corresponding RCX_OPTION.
Thank you very much for everything, I already have to entertain myself, and what about the Pitot I am thinking about how to implement, I will study your system.
Hello Samuel, yesterday I was flying the ASW 28, with the parameters you sent me, although the result has not been optimal, I was very happy with what I achieved.
It was a mild thermal day, but enough to get a respectable height.
The first thing to note is that when he found a thermal, he did nothing, he continued his course as if nothing happened.
Given the behavior, I decided to help, and as soon as I heard that the Vario was marking thermal, it was I myself who started the maneuver of turning to meet the thermal, and immediately I noticed in the Vario that it rose rapidly.
During the ascent I limited myself to staying in the area, and letting myself be carried away by the wind, and the Vario kept marking ascent.
Ardupilot was in charge at all times to control Pits, that is, the depth was controlled by Ardupilot in all the turns, with which it did not lose height but gained.
With all this I come to the conclusion that the parameter that you have in soaring of SOAR_VSPEED = 1, is too high and does not reach this value of VSPEED, and that is why I cannot detect the thermal.
The latter, or that it really does not work, as it should work.
On the next flight day, I will change the SOAR_VSPEED = 0.7 which is as it was previously.
Iām going to try to send you my .log file of the reported flight.
I have been experimenting with SOAR further, and am having some success with it. It is very close to actually working as well as a human, not yet, but there is so much potential !!! SOAR still seems to not quite find the highest area of lift in a circle sometimes, and I think I have figured it out
When a plane changes bank angle to enter a thermal, or encounters turbulence while in a thermal, the nose pitches up and down a bit. This can cause errors in vertical speed measurement I see this and adjust for it while I am flying, and SOAR needs to do the same thing. There needs to be a PITCH CORRECTION factor when measuring vertical speed. So if the nose if the plane is up 2 degrees, subtract X amount from the vertical speed. If the nose is down at the moment of measurement, add X amount to the vertical speed.
Every planes vertical speed would be affected by pitch by a different amount, so find a good resonable baseline number, and then let us vary it by a setting that would be something like:
SOAR_VSPEED_PITCH_COMPENSATION
This would result in much more accurate measurement of lift throughout the circle and allow SOAR to much better and consistently move the plane to the area of higher lift.
Thanks for sending the log, a couple of things to note.
automatically changing to THERMAL mode is only supported from FBWB, CRUISE or AUTO modes. You were in FBWA mode.
yes it looks like a lower value of SOAR_VSPEED would work better for you. You can even fly with a value of zero, which means āthermal as long as you think you can maintain altitudeā
you are flying without an airspeed sensor which is still experimental. In fact I have spotted and fixed a minor issue in your log that could lead to use of the motor in THERMAL mode
Thanks for the report. In theory the pitch compensation should be less of an issue since ArduPilot is providing stabilization. If you have a SD card log (.BIN or .LOG) I can take a look. You can send me privately too if you prefer.
Iām building my first airplane with a FC. My plane is an original FMS Fox sailplane using a Matek F765-wing FC. I was curious if anyone else is using a Fox, and if so what parameters they use.
Hello Samuel, the truth is that I think the operating method is not clear.
In the Ardupilot Soarin Manual, there is talk of a switch, to control Soaring, but in the current version, it is dispensed with. On the other hand, in your configuration you do not have any switch for this function, you only have an RC1 channel, for Soaring Enabled.
In Ardupilot flight modes, you use channel 8 for modes 1 and 2 Manual, 3 and 4 FBWA, and 5 and 6 Auto.
All this is quite complicated to understand, what is the operating method you use.
As I can understand, you use the RC1 channel, to get the Soaring operational, and with the flight modes you switch between Manual, FBWA and Auto.
Of these Flight Modes, only Auto Mode can make Soaring go into automatic.
It would be interesting and it would help me a lot, if you explain your operating method of flight, since the launch of the sailboat.
Nor do I understand why we have in this version, a Thermal flight mode, or what it can be used for, if you donāt use it with what you carry.
Anyway, itās all a big mess.
Thanks for your explanation, and we will continue working on Soaring.
Iāll see if I can make the wiki a bit clearer. To answer some of your specific queries/confusions:
The switch is still mostly the same, its the parameters to set it up that have changed. For example, I use channel 5. In the stable version, I would have set SOAR_ENABLE_CH = 5. In the latest version, I set RC5_OPTION = 88, or you can select āSoaring Enableā in the drop-down menu in Mission Planner.
The switch now has a middle position that allows more functionality as described on the wiki.
THERMAL mode contains the logic for exploiting thermals. In youāre in THERMAL mode, this logic is running to try to find the centre of lift. You can switch into this mode like any other mode using flight mode RC channel or from the GCS.
Typically yes I fly in AUTO mode with a WP plan, but donāt have to do this. You can fly in CRUISE or FBWB mode, in which case soaring can automatically change to THERMAL mode,or FBWA mode, in which case you will need to switch into THERMAL mode using the flight mode RC channel or GCS.
it might be worth adding that SITL simulation allows to extensively familiarize with the SOAR options / settings and their interactions right there on your desktop, before using these features for the first time in a real-life setting. that helped me a lot to decrease the stress-level and focus actualy flying
Hello again Samuel, Iāll see if I understand.
To simplify, I will only use the flight modes that Ardupilot offers, in my case channel 5, with which I will configure Mode 1 = Manual, Mode 2 = FBWB, Mode 3 = Cruise, Mode 4 = Thermal, Mode 5 = Auto and finally Mode 6 = RTL, very important for me the latter too.
I will launch the sailboat normally in Manual Mode.
Once I have reached a sufficient height, I will switch to FBWB Mode, I will see what it does when I enter Soar, when it detects a SOAR_VSPEED of more than 0.5 m / s.
I understand that if this occurs, the system will automatically go to Thermal Mode, without the need to do anything else.
In any case, and although this does not occur, because it does not reach SOAR_VSPEED of 0.5 m / s, if I personally decide to warn a Thermica devil through the Vario, I decide to switch to Thermal Mode, what should I do?
What is not clear to me, is that the dodo Thermal does, voluntarily, if in itself it already has the condition of FBWB, CRUISE or AUTO Mode, which are absolutely necessary, to do Thermal flight.
That is my great doubt ??.
I see this and adjust for it while I am flying, and SOAR needs to do the same thing. There needs to be a PITCH CORRECTION factor when measuring vertical speed. So if the nose if the plane is up 2 degrees, subtract X amount from the vertical speed. If the nose is down at the moment of measurement, add X amount to the vertical speed.