OpenTX - Nightly builds for pre-2.0 testing

[Rob Thomson]

Yup.

This is not a normal button sequence to press.

Makes sense for it to be hidden.

Sent from my GT-I9300 using Tapatalk

[Daedalus66]

It's just hidden enough.

[Scott Page]

Hello friends,

I've been absent for many weeks going through a rough patch. I just downloaded Ctx and 3/29 build and will try to bring myself up to speed over the next week or so. Launched Ctx and see that "we are not in Kansas anymore".

[Scott Page]

The curve generation and model wizard are part of the model edit. That part of Companion has been heavily updated to handle the new eeprom format. The two functions are not yet re-implemented, but the plan is to bring them back. I can not guarantee that we will have any of them ready for 2.0. I am however working on a new Wizard and it will certainly be back in 2.1 (or whatever it is called).

I'm very relieved that curve generator will be returning. That was one of the first things I noticed missing. The curve generator is a huge selling point to people when I show them how easy it is to make a custom curve.

Model Wizard was another thing that crushed my soul when I saw it was not functioning. Very pleased it's not been thrown in the bin. The wizard is the most powerful selling point to people wanting to know "how hard" is it to use the Taranis. It also reduces the amount of support questions we get posed.

I noticed the length of the BMP's has been shortened to only 8 characters. I can't spell Messerschmitt Me 262 with only 8 characters.

Is there really a substantial savings by dropping the number of characters for the models - I was hoping we'd get more characters for both .wav files and .bmp.

I see a tremendous amount of new implementation that is very nice. I like that sticks is now called Input - that alone will drop the number of noobie questions by 30%. (I can always hope).

I question why the mixer continues to display the (Ixx) designation after an input has been assigned a name. Not a big deal of course - just seems a bit jumbled - but I bet there is a good reason.

I'd hoped for an additional IT function that would center using the trim settings (like ER9x did) while flying, but I see that won't happen. I'm getting accustomed to using extended trim and then the servo menu to center the trims - except when I forget to turn on extended trims ahead of time. The current implementation of IT cost me plenty early in the open9x evolution. Seems I can't hold my mode 2 aileron and elevator steady while flicking the momentary switch on the right side. I just made a mess of things that way.

c**p, I just read this and it seems like I don't like what I see. That's not the case. Lots of polish has been applied. I'm sure LUA will be a growth curve for me. I've not coded in 30 years.

Tomorrow I'll move past companion and burn the firmware to my simulator Taranis and put some sim time on this.

[jhsa]

You will like the new way to flash the radio with the bootloader

[bertrand35]

A quick answer to Scott:
- The wizards have been disabled, and are in the middle of a full rewrite
- I just fixed a small bug in companion that limited the length of bitmaps to 8 chars instead of 10

[bertrand35]

The Companion installer for Windows should now be automatically generated during the night.

[Thorvald]

I am slowly writing a wiki page for Lua scripts on OpenTX. It's here:
https://github.com/opentx/opentx/wiki/L ... -in-OpenTX

You will be able to download and run the classical snake game at the very end of the page ... in case you broke all your planes

Will it be possible to use telemetry values as inputs to Lua scripts? And will it be possible to use outputs from Lua scripts as input values in the mixer?

Hopefully. I'm also waiting for some of the Telemetry functions to be revealed so I can try to implement Head Speed for RPM. Haven't had a chance to browse through the source looking for the functions currently defined though.

EDIT: Just browsed through lua.cpp for the latest build and I see that we can use the "getValue" function to do what we need. Currently it's rather basic though and doesn't include all the Telemetry variables (easily fixed when time allows) but here are some of the ones already implemented:

altitude
latitude
longitude
pilot latitude
pilot longitude

Using Altitude as a sample we can easily add code to pull from all the other telemetry values. For example for RPM we could add another else clause (not tested, just a quick 'what if' post):


const char *what = luaL_checkstring(L, 1);
if (!strcmp(what, "rawrpm")) {
lua_pushnumber(L, double(frskyData.hub.rpm * 60));
return 1;
}


Though I see that the current next build still has the "blades" bug where it assumes a minimum of 2 blades (it's used for an rpm divisor) which means you displayed value in the telem screens is divided by a minimum of 2. Thought I got this resolved back in 2923...weird.

You can then go through all the other sensor hub values and create a clause for whatever value you want. Might play with this tonight.

[Scott Page]

A quick answer to Scott:
- The wizards have been disabled, and are in the middle of a full rewrite
- I just fixed a small bug in companion that limited the length of bitmaps to 8 chars instead of 10

Obviously just my opinion - but I think it's better to release a completed 2.0 rather than a mostly done 2.0 with missing features released in 2.1

Those who are familiar with FrSky are accustomed to waiting longer to get a better product.

Relieved to get 10 characters in .bmps. I was already feeling cramped with 10.

I love the ability to resize the icons because I'm now using a tablet for companion. A Dell Venue 8 Pro. It's running Windows 8.1pro. I think tablets will become more common for Companion once their potential is realized by more. It essentially gives me all the pie in the sky things people want in a transmitter the color touchscreen, 5 mpix camera, audio recording, video camera, bluetooth keyboard and mouse, etc.... and it fits into a transmitter case. I am not using 1.99 on the tablet yet because I'm on holiday for a week and was stupid and left it home. (of course I can also run photoshop and audio editors on the tablet).

New Preference setup is a big plus. Very tidy.

I like the sticky in customlogical switches - however I miss the toggle in special functions - unless I'm missing something this ads a step to setting up a toggle for a single switch as now it has to be made a sticky in logical switches before it can be used as a toggle in special functions. I like that the sticky for more complex logical combinations is very powerful, but the simplicity of a single switch has been lost.

I have not been able to crash the program yet. I'm gifted that way so we'll seem Given that I'm on holiday I won't be able to check the bootloader from a virgin computer until I return - but I know a person who still can't get Zadig to work and I'm wondering about suggesting he try using 1.99 with 2940. I'll wait until I get a suggestion here before I recommend it to him.

I wish I had put the 6 pos switch into my simulator Tx so I could check that out -- but there is plenty more to explore and play with.

Great work you guys.

[dvogonen]

I think it's better to release a completed 2.0 rather than a mostly done 2.0 with missing features released in 2.1

I think we all agree, but there is a limited amount of develoment time available and some of the new features, like Mikes bootloader, are huge and very necessary improvements that need to be released.

New Preference setup is a big plus. Very tidy.

Thanks! That one was mine.

I know a person who still can't get Zadig to work and I'm wondering about suggesting he try using 1.99 with 2940.

I don't really understand how that would help. Companion 1.99 still needs a working driver stack for Dfu-util. No change compared to 1.52.
But Zadig had an update a month or so ago that made it work better on many machines. Previously it crashed on e.g. certain Samsung Ultra books (I have one of those). Now it works just fine. Link on the project page of OpenTX (www.open-tx.org).

[MikeB]

The bootloader code I passed over was fully working. It only took me about half an hour to compile a version of r2940 with the bootloader in.
This still applies. The bootloader sources I'm using are now in a Github repository (https://github.com/MikeBland/OpenRcBootloader).

These are the sources I'm using for ersky9x on the SKY board, 9XR-PRO and Taranis, and for openTx (r2490) on Taranis. They include some fixes for use with MAC computers to solve the "unmounting" problem.

Mike.

[Scott Page]

I don't really understand how that would help. Companion 1.99 still needs a working driver stack for Dfu-util. No change compared to 1.52.
But Zadig had an update a month or so ago that made it work better on many machines. Previously it crashed on e.g. certain Samsung Ultra books (I have one of those). Now it works just fine. Link on the project page of OpenTX (http://www.open-tx.org).

I guess I misunderstand. I believed Ctx1.99 had a new bootloader that did away with zadig altogether. This is wrong? I now get the impression the bootloader is separate from Ctx and part of the firmware itself.

Studying Mikes link now...

[dvogonen]

Yes, the bootloader is a separate part of the firmware.
Once a firmware version with a bootloader has been installed, the bootloader can be used to flash new firmware files stored on the radio SD card. This means that the user only has to be able to copy files to the radio using the built in memory USB drivers. No more special USB drivers for DFU!

But all Taranis radios in the world have to be updated the old way one last time before it is possible to flash using the bootloader.

[MikeB]

The bootloader takes up the first 32K of the program memory, and is a completely separate program. It is built into the main program, so flashing with Zadig etc. flashes both the bootloader and the main program. The bootloader 'knows' not to overwrite itself.

The bootloader provides USB access to the SD card, and a second drive. This second drive used to contain just the EEPROM, but now also contains a file mapped to the flash memory.

The bootloader allows you to select a .bin file from the firmware sub-directory on the SD card and flash it. It does some small checks to see if
the .bin looks like a suitable file to flash.

Since the flash memory is also mapped to a file, it is also possible to write to it directly. For ersky9x, I have modified eepskye to do this. Simply power up the Tx in bootloader mode (hold both horizontal trims towards the centre while switching on), run eepskye, click on the "Flash Firmware to Tx" button, select the file to flash, and click OK. 3 or 4 seconds later the new file has been flashed! I assume Ctx will have the same capability. I've been using the bootloader on my SKY board, the 9XR-PRO and the Taranis for some time, eepskye flashes all of them.

In ersky9x, I also have "maintenance mode" (hold both horizontal trims away from the centre while switching on). This runs as part of the main program, and allows you to update the bootloader from a .bin file on the SD card. On the SKY board, it also allows you to update the co-processor from a .bin file on the SD card. I'm also working on having this be able to update SPort devices' firmware from the SD card. This would include the internal XJT module, I just need some more time to get it done!

Mike.

[Scott Page]

Working with march 31 build of CompanionTX 1.99 on windows 7. Mostly cosmetic things - a couple that might need some attention.. not sure.

Found that 10 characters is once again supported for .bmp files. YAY! I also like that using an arrow key allows one to browse through the images in the setup window. Will make finding defective .bmp files much easier.
Found the curve generator back in curves with some new features. YAY!

Found that choosing simulation colors in preferences does not change simulation color when simulation is chosen from EEPE window.
Found that choosing to simulate backlight on in simulation from model window does not work.
Found that SD card menu - quick click the ellipse [..] moves above the specified root in SD structure Path. Can get a person into trouble. probably should be restricted to the specified SD Structure path.
Found that when simulating TX from EEPE window or model window - the switch selection in Special functions is much different than switch selection in Companion special function window...


That's all I had time for. It's 1 AM - Hope to put OpenTX on simulatorTX tomorrow.

[bertrand35]

Why do you say that the switch selection is different? It shouldn't! Perhaps Companion still doesn't hide unavailable switches while the firmware does (i.e. you won't see the unconfigured logical switches or the multipos switches if you don't have any etc.)

Scott we have an issue open here:
https://github.com/opentx/opentx/issues/827

I thought I fixed all remaining bugs in this area, would you have a quick check so that I close the issue?

Thanks for your help

[dvogonen]

Found that choosing simulation colors in preferences does not change simulation color when simulation is chosen from EEPE window.

The setting works for the 9x. I guess that you were simulating Taranis. Since Taranis comes with only one background color (blue), the setting has no effect. The setting should naturally not be visible for the Taranis. I will file an issue report and fix the problem.

Found that choosing to simulate backlight on in simulation from model window does not work.

I guess that this is a bug. Turning backlight on and off works if the simulator is started from the model selection box. It does not work if the simulator is started from the Model Edit dialog. I will file an issue report.

[Scott Page]

Why do you say that the switch selection is different? It shouldn't! Perhaps Companion still doesn't hide unavailable switches while the firmware does (i.e. you won't see the unconfigured logical switches or the multipos switches if you don't have any etc.)

Scott we have an issue open here:
https://github.com/opentx/opentx/issues/827

I thought I fixed all remaining bugs in this area, would you have a quick check so that I close the issue?

Thanks for your help

See PM.

[Scott Page]

Using april 1 uploads of companion and firmware on windows.

In special functions the switch availability in transmitter and Tx simulator are much much more limited than in CompanionTX. These trim switches immediately follow the hard switches, . S11 - 36 are missing and also missing are all logical switches (L2 - L32) except L1. This is the same with the Transmitter.

Also -- in the simulator the buttons are really a pain. The buttons are way too sensitive and I end up in the next screen frequently with a single press of the ENT key .. or sometimes of the minus key. The short key press needs to allow for a longer time I think.

Additionally, in the simulator once in the LUA menu and after selection a LUA script .... long page press frequently won't back out of the screen back to Special functions.
Finally, the LUA screen won't wrap around to the first screen in the simulator but it does just fine in the Transmitter.
But if I just page through all the screens and don't stop on the LUA screen I can wrap around to the first screen

[HeliKris]

Hi,
FVAS- is now implemented. But the max. selectable voltage value for creating a low voltage warning is still 25.5 Volt.
Too low for a 12s battery voltage warning.

Thanks for your great work,
Kris

[bertrand35]

snapshot_27.png (1.72 KiB) Viewed 16440 times

[bertrand35]

Scott, in the firmware, the logical switches which are not defined (which are empty) are not shown in the list of switches when you configure a mix / expo / function etc. I will do the same on Companion side when I have time.

Also go in the Hardware Menu on your simulator and choose "Multipos Switch" for S1 to S3 and you will see 18 new switches in the switches list. The ability of hiding those unuseful choices in the list is a feature, not a bug!!!

[jhsa]

A very good idea indeed.

Sent from my GT-I9195 using Tapatalk

[jonlowe]

A quick answer to Scott:
- The wizards have been disabled, and are in the middle of a full rewrite
- I just fixed a small bug in companion that limited the length of bitmaps to 8 chars instead of 10

Obviously just my opinion - but I think it's better to release a completed 2.0 rather than a mostly done 2.0 with missing features released in 2.1

Those who are familiar with FrSky are accustomed to waiting longer to get a better product.

(snip)

+1. Maybe release a "stable beta" for power users. A lot of people haven't updated firmware at all from what FrSky put on the system. We will still need to get them over that hump, with Zadig, in order to get the boot loader installed. These non-power users are the ones who most need templates. A lot of setup questions go away once they realize templates are there.

Jon

Sent from my Nexus 5 using Tapatalk

[dvogonen]

Based on the feedback this far Bertrand has put the Curve Wizard back into 2.0 and I just completed the Model Wizard. An incomplete Model Wizard was committed to next yesterday and the finished version should be available in tomorrow's nightly build.

[Scott Page]

Scott, in the firmware, the logical switches which are not defined (which are empty) are not shown in the list of switches when you configure a mix / expo / function etc. I will do the same on Companion side when I have time.

Also go in the Hardware Menu on your simulator and choose "Multipos Switch" for S1 to S3 and you will see 18 new switches in the switches list. The ability of hiding those unuseful choices in the list is a feature, not a bug!!!

That makes so much sense! Great idea.

I just quickly reviewed the model Wizard. All I can say is WOW. That will become the best friend of many people I'm sure. The artwork even looks friendly. I know it's incomplete so don't know if it's time to mention it needs ability to set two separate servos for elevator. But then again - not sure if those who NEED a wizard will be using two separate servos for elevator on standard tail - or perhaps in my haste to try every combination if I missed it. But then again, I also have a biplane with 4 separate servos for aileron used as brakes (top down, bottom up) - so I guess that a wizard can't consider every contingency.

[dvogonen]

Martin Hotar made the nice Wizard graphics. He also made the very nice default icon theme for Companion 2. I think that the cartoon style graphics brings a friendly and inviting feeling.

My thought was that it is simple to set up a second, third or fourth servo once the first has been set up for you by the wizard. I wanted to keep the wizard as simple to use as possible. That is why I did not add every imaginable servo configuration.

[Scott Page]

Martin Hotar made the nice Wizard graphics. He also made the very nice default icon theme for Companion 2. I think that the cartoon style graphics brings a friendly and inviting feeling.

My thought was that it is simple to set up a second, third or fourth servo once the first has been set up for you by the wizard. I wanted to keep the wizard as simple to use as possible. That is why I did not add every imaginable servo configuration.

I certainly support that concept. It seems to me that by the time a person is using two elevator servos, four aileron servos, or more than 8 channels they probably are ready to use the mixer. Really great stuff there.

[rdeanchurch]

The new model wizard is simply amazing.

Thanks to all resposible.

If you are flying models that need multiple servos per surface, I would think you darn well better be understanding mixes and limits....

[bertrand35]

A quick answer to Scott:
- The wizards have been disabled, and are in the middle of a full rewrite
- I just fixed a small bug in companion that limited the length of bitmaps to 8 chars instead of 10

Obviously just my opinion - but I think it's better to release a completed 2.0 rather than a mostly done 2.0 with missing features released in 2.1

Those who are familiar with FrSky are accustomed to waiting longer to get a better product.

(snip)

+1. Maybe release a "stable beta" for power users. A lot of people haven't updated firmware at all from what FrSky put on the system. We will still need to get them over that hump, with Zadig, in order to get the boot loader installed. These non-power users are the ones who most need templates. A lot of setup questions go away once they realize templates are there.

Jon

We are discussing here around the Nightly Builds. They are available only for testing, not for flying, and I am really happy about the feedback. Many users help to find issues "in the corners" (at least on Taranis) and this is good!

As soon as we are ready (I mean no new bug found during 1 or 2 weeks, no hurry), we will make available a Release Candidate 1. Something like 1.99 RC1. Then if needed there will be a 1.99 RC2 and finally an OpenTX 2.0 release.

Bertrand