TX16s with USB connection as controller

hsamuels

New member
I agree on the assessment of the Windows USB Joystick interface. The first 8 channels are proportional, and then there are a whole pile of buttons, which we would normally assign to 2-pos switches. You can assign the sticks, sliders, pots, or >2 position switches to the proportional channels. You could also assign a 2-pos switch to a proportional channel, but you won't make any friends doing that. And you may have a calibration problem.

I've assigned the 6-pos switch to a proportional channel. Even though it's impossible to set it to 50%, after a couple of presses of most of the buttons (1 and 6 alone would probably suffice) all of the buttons generate the correct values.

The Interlink controller has the nice feature that you can hold the momentary contact up/down button to scroll up and down through the list of models or airfields. I believe the controller reads the state of the switch, and then generates a bunch of up or down presses to send out the USB port. There's isn't an input device on the TX16s that can do that. (Light bulb turns on!) Can a LUA script or special function do that for a 3-pos switch? I'm fairly new to OpenTx, but it seems like it should be possible.

So far, it seems we've established that more than one of us has seen the 64% trim problem after calibration in RF8. And we've also seen examples of not having the problem in RF8. But we don't know what causes the issue, or more importantly, how we should set up the radio to avoid it.

And so far, it hasn't been observed in RF9. But that doesn't mean it won't happen.
 

Flapper

Active member
I finally got my setup to go wonky and give me either 36% centers or 64% centers. Still need more time to verify that it is reproducible. It is strongly looking to be Windows. Seems to be too much change on the input and reconfiguring of the "gamepad" device confuses things. And when Windows is confused, RealFlight is confused. I think the order of battle will be:
1. See if I can get the Windows gamepad config to go "wonky" on me consistently after trying different TX configs, without a restart. Also test deletes/reinstalls of the device without a restart, and see that can consistently upset things.
2. Do a driver delete and windows restart, before plugging in the new TX config. See if that gets me consistently well behaved Windows gamepad behavior.
3. Do some comparisons between an un-calibrated gamepad and RF vs doing the Win Calibration first, then RF.

The other "bad" thing I do - run it all through a USB hub, which is notorious for not being well behaved with some things. I'll do some testing with and without that, also.
 

Bill Stuntz

Member
Dumb question: Could it have anything to do with USB2/USB3 compatibility? I don't rember exactly what the problems were, but I had problems with my Interlink Elite when I plugged it into a USB3 port on my HP Z230 workstation. Moving it to a USB2 port solved whatever those problems were.
 

hsamuels

New member
Dumb question: Could it have anything to do with USB2/USB3 compatibility? I don't rember exactly what the problems were, but I had problems with my Interlink Elite when I plugged it into a USB3 port on my HP Z230 workstation. Moving it to a USB2 port solved whatever those problems were.
Look at the footer on your own post! I don't know if USB 3 could be the issues, but my computer has only USB 3 ports so I can't try a USB 2 port.
 

hsamuels

New member
I've noted that my WS2000 works perfectly. I also have an Rx2Sim that always worked perfectly, but I haven't tried it recently. It shares the 8-channel limitation. Could it be an issue with how Windows deals with more than 8 channels? Admittedly the joystick test window indicates proper operation after calibration - all 8 proportional axes are smooth, apparently centered, and there is no mixing. And as many buttons as you can come up with seem to light the fake LEDs.

But the TX16s only has two 2-pos switches. The rest are 3-pos, plus the 6-pos radio button. They light up the joystick test window LEDs without a problem. But how does the Windows driver really deal with the middle position? Perhaps this is why the WS2000 and Rx2Sim have only 8 channels - it prevents users from trying exactly what we are doing here.
 

Flapper

Active member
I've probably worn out my power switches on the computer and the TX completely, with trying out a bazillion combinations today. On my system, it all leads back to the Windows gamepad configuration. IF I start with a given setup on the TX and then change it, even if I delete the device in Windows, and reinstall, Windows just does not behave normally in reading/calibrating. And of course, neither does RealFlight. I have to delete the device, restart Windows, THEN reinstall, and all proceeds as expected. In addition, if the device is not calibrated in Windows first, RF calibration is always bad. But at least I can exit out of RF, calibrate in Win, then go back in to RF and it will calibrate correctly without a restart.

So, on an existing badly behaved system:
1. With the TX plugged in and in USB mode, go to Devices and Printers, and uninstall the gamepad device representing the TX.
2. Unplug the TX.
3. Do a full shutdown/restart cycle of Windows.
4. Plug in the TX and enable the USB Joystick. (if you already have the "Devices and Printers" open, give it a few seconds to completely self configure).
5. In "Devices and Printers", go into the device configuration, and calibrate it.
6. Make sure there are no profiles stored in your RF Radio Profiles folder.
7. Then open RF, and the new controller. Pick one of the stock transmitter types, rather than gamepad or joystick.
8. Calibrate it again there.
9. Map the controls as needed in the main Edit screen.

Why does it behave fine in RF 9.5 but not in RF 8.5, as previously reported? You've got me! Clearly some interplay between Win and RF. We know Win 10 changed USB drivers about a year ago. Perhaps something there is optimized, that is not anymore in the older version. But at least on this one system, making the install as clean as possible seems to resolve the problem.
 

hsamuels

New member
Thank you for sacrificing your switches to the cause!

In all of my failed instances, the Windows calibration appears to be working fine. According to the joystick test window, the bars appear to be centered, and the min and max limits on the sticks appear to make the bars go from 0 to 100.

I'll try your procedure for clearing the badly behaved system. Maybe I can even get two models on the radio, one behaving badly, and one not. That may provide some extra clues.
 

BrokeDad

Well-known member
A couple of us found that when using the Interlink DX in RF8 was fine as long as you only calibrated it once. Doing it a second time made it all fubar. Just thought I'd remind folks about that. I don't think it was ever resolved.
 

marcushh777

Well-known member
A couple of us found that when using the Interlink DX in RF8 was fine as long as you only calibrated it once. Doing it a second time made it all fubar. Just thought I'd remind folks about that. I don't think it was ever resolved.
I believe taking the option to 'reset to defaults' and then calibrating will straighten out the fubar.

marcus
 

N8LBV

Member
I currently cannot get past the calibration in RF 8.5 putting center at 64%.
No matter how many times I remove the controller from windows and shutdown or restart the computer
and re-plug the controller and re-calibrate in windows.
Calibration in windows always looks fine. and is truly at the numeric midpoint between 0 and 2048 (1024).
I have been unable to make it work in RF 8.5 period since I have been trying.

And it consistently always works perfect with no fuss in RF 9.5
 

N8LBV

Member
I believe taking the option to 'reset to defaults' and then calibrating will straighten out the fubar.

marcus
No, not here.
RF 8.5-
There is no "reset defaults" option within the controller profile editor or the calibration screen.
Resetting defaults from the RF 8.5 launcher had no influence on this problem.

Resetting default in Windows USB controller and re-calibrating there also has no influence on the problem.

And removing the device/uninstalling from windows and restarting windows has no influence on the problem.


Repeat: no sign of this problem in RF 9.5
If it were a windows/driver /windows/calibration issue I should be seeing the problem in RF 9.5 as well
 

BrokeDad

Well-known member
I kept 7.5 installed just because I have so many aircraft in it. I deleted RF8 once I had RF9. Unfortunately RF8 is now a legacy product that will no longer be supported with fixes or upgrades. The Interlink Elite worked fine in RF8 btw .. it was just my new DX that was difficult.
 

BrokeDad

Well-known member
I believe they will still give you tech support for RF8 but there won't be any further development/fixes of the software.
 

N8LBV

Member
I believe they will still give you tech support for RF8 but there won't be any further development/fixes of the software.
A fix would be nice.
And isn't it pretty normal for a company to issue fixes for bugs like this for current supported software?

Rather serious it seems.
I'd be satisfied with a work-around that works versus a software update.
Might be a "simple" matter of backporting what works in RF9.5 to this if that is possible.
Once whatever is different in 9.5 from 8.5 causing it to work there but not here is identified.
 

marcushh777

Well-known member
A fix would be nice.
And isn't it pretty normal for a company to issue fixes for bugs like this for current supported software?
Under 'normal' circumstances, yes. But, this situation isn't normal. Horizon Hobby isn't a software house. Sure, they bought out KnifeEdge, but Horizon Hobby isn't really motivated to update the software (and probably won't). On the other hand, they may farm it out to what's left of KnigeEdge, or some other software house; I wouldn't hold my breath.

marcus
 

N8LBV

Member
This from my perspective appears to be a bug or issue in RF 8.5
I am opening a support issue with Horizon Hobby to see if I can get any help with it there.
It's 100% repeatable on my end and I have not yet found a fix from everything I have tried.

If you think I have missed anything I should have tried please let me know.
Or if you want me to repeat any tests in specific order also please let me know and I will do that.
More data is better in this case if we have it.

Seems worthwhile to me.

I think it would be good to be able to use your OpenTX radio of choice and other controllers without this issue
causing the problem that it is. on RF 8.5
 

N8LBV

Member
Short response: Everything is working now, and all proportional channels are 50% when the controls are centered.

Longer response:

I had set up 10 inputs, and then assigned each input to a channel on the mixer screen, 100% for each input. But the order in the mixer wasn't always the same as the order on the input screen. The monitor showed what was expected from the mixer screen.

I set the order of the inputs to what I might need for RF - reset and gear on 'buttons' with everything else on the first 8 proportional channels. Then I deleted all of the mixes. Then I edited each mix by entering and exiting immediately, and the mixer screen looks almost exactly like the input screen (I gave the inputs names). Screenshots attached.

Flapper, trims 64.txt is a 'bad' setup. Everything looks great; you can assign various switches and sticks to things in RF. After you calibrate, the sticks are at 64%. Kindly let me know if you see the same issue with calibration. It really would be nice to know the root cause here.

Thank you again for all your help,

Howard
 

N8LBV

Member
Short response: Everything is working now, and all proportional channels are 50% when the controls are centered.

Longer response:

I had set up 10 inputs, and then assigned each input to a channel on the mixer screen, 100% for each input. But the order in the mixer wasn't always the same as the order on the input screen. The monitor showed what was expected from the mixer screen.

I set the order of the inputs to what I might need for RF - reset and gear on 'buttons' with everything else on the first 8 proportional channels. Then I deleted all of the mixes. Then I edited each mix by entering and exiting immediately, and the mixer screen looks almost exactly like the input screen (I gave the inputs names). Screenshots attached.

Flapper, trims 64.txt is a 'bad' setup. Everything looks great; you can assign various switches and sticks to things in RF. After you calibrate, the sticks are at 64%. Kindly let me know if you see the same issue with calibration. It really would be nice to know the root cause here.

Thank you again for all your help,

Howard
I am having extreme EXTREME difficulty following this post.
Says you got it working but I don't understand what you did.
At all.
I feel stupid.

All of my channels inputs and mixes are set to 100% and in same order
on INPUT and MIXES screens (in OPENTX)
First 4 channels are AETR

No matter what I do or what I try I am getting the center at 64% problem.

All channels at midpoint are indeed at midpoint in windows calibration screen.

And it always sets midpoint to 50% in Realflight 9.5

From this data alone it's rather clear that the problem is not in the TX or in Windows calibration.

I am only using six channels to test
All proprotional.

1234 AETR

5 is mode (3-way switch)
6 is throttle hold. (2-way switch)
 
Top