trainer jitter

[gfnk357]

I just got an ersky9x board set up in my new radio so I thought I'd try setting up a trainer system with my er9x radio as student and the ersky9x radio as teacher. I connected the two with a generic double ended stereo cable and I was able to get it all set up without an issue but when the student radio is in control the servos don't move as smoothly, like there's not as much resolution as they move. When the teacher radio is in control the servos move smoothly. Does anyone have suggestions why this might be happening?

[preston91]

have you done the PPM mode? If not, look here http://openrcforums.com/wiki/index.php/ ... for_the_9x should fix it and easy to do

[ShowMaster]

Is the trainer 9x using the stock rf module or a frsky DJT?
What firmware does the trainer 9x use, stock or er9x?
I ask this becsuse even though you're not transmitting from it the stock module will load the ppm line with the stock firmware and the 1k resistor mod may help. A test would be to remove it.
If your using a Frsky DJT module in the trainer, again not transmitting, it does not load the ppm output do no 1 K resistor is needed.
If your using er9x in the trainer tx there is a menu option that will isolate the ppm out from the module loading of any module I think so again no 1K resistor mod needed.
Some answers to my questions may help narrow down a fix without the resistor.
First just try removing the trainer module to see if it's a ppm loading issue and go from there.


ShowMaster
Sent from my iPod touch using Tapatalk

[gfnk357]

The student radio uses er9x and the teacher radio uses ersky9x. On the student radio I had just left in the spektrum module that I had in there from an earlier flight, the teacher radio had a flysky transmitter in it. Removing the module from the student radio was one of the first things I tried and there was no difference. I've known about the module loading issues but I've used this radio for simulators for quite a while with no problems. I just left the radio turned off when the trainer cord was plugged in to the computer and the ppm out signal was sent with no problems. What is the menu option that isolates ppm out from the module loading, is it the enable ppmsim?

I also tried a spektrum module in the teacher radio to see if it was an issue with flysky and found a separate issue that I was able to solve. I use a hacked module from a hz champ so it's using the DSM2 protocol instead of ppm. The teacher radio won't read a DSM2 signal coming into the trainer port but if I changed the student radio to a model using ppm then the ppm signal gets sent to the teacher radio which then changes it to DSM2 and allows the trainer function to work correctly.

[MikeB]

I've just done an investigation of this. The problem is 'phasing' between the trainer inputs, the tX outputs to the Tx module and the Tx/Rx link and Rx outputs. These are all asynchronous to each other.
Currently I believe, the teacher radio waits for a complete PPM frame of channels from the student before updating the values to send out. This is then sent, starting at the beginning of the next frame to send to the Tx module. When the receiver receives this data, it updates the servo output. Each of these steps can cause a delay resulting in the jerky movement you describe.
If you move the student stick quickly from end to end, the servo works fine. If you move the student stick very slowly, the servo does follow it, so it isn't a resolution problem.
I got improved output by changing the student to be only 4 channels, with a frame period of 16mS. There was very little difference in smoothness between teacher and student then.
I'll look into seeing if this can be improved, time permitting.

Mike.

[ShowMaster]

Can you try another brand tx as the student radio? Maybe at your field if you don't have one?
My thoughts are to determine if it's a ersky9x issue or er9x issue now?
I've not used the trainer option on my sky board yet but I do have many student tx options to try. JR and Spektrum TXs as well as stock 9x and er9x TXs.
Please repost your ersky9x and er9x FW versions and maybe myself and others will find time to sort this out.
The trainer feature is an important option to have working well at any large flying field.



ShowMaster
Sent from my iPod touch using Tapatalk

[MikeB]

Just looked closely at the code processing the PPM in from the student. It looks like it misses every other input frame (both er9x and ersky9x). I've just put in a fix I'm testing, the student now looks much smoother.

Will be in the next revision.

Mike.

[ShowMaster]

Good news Mike on finding a posable error and working on a fix. I'll hold off testing setup.



ShowMaster
Sent from my iPod touch using Tapatalk

[gfnk357]

Mike, thanks for checking into it. I'll try again once the new revisions come out.

[gfnk357]

Mike,
did the trainer fix get in to the new revision? I just tried my radios and I still get rough movement if I'm using the ersky9x radio as the master and er9x as the student. If I switch them around so that the er9x is master and the ersky9x is student then it works smoothly.

[MikeB]

The fix for ersky9x should be in the revision I released earlier today.

Mike.

[gfnk357]

okay, do you know of any way to send a ppm signal from the computer to the trainer jack? I'd like to figure out if it's the ersky9x radio or the er9x radio that's the problem but I don't know anyone else that I can borrow a radio from to try hooking up as a ppm in test. I did try open9x and I got the same problems with the trainer signal being rough so it may be that the signal is just bad from my er9x but I'd like to be able to try another input source to confirm.

[Kilrah]

There are USB to PPM interfaces around, but they cost more than a 9x

[gfnk357]

Well.... I decided to just go ahead and make my own ppm source so I could verify if the problem was the signal that the er9x was generating or a problem with the ersky9x reading the signal. I found a joystick at a flea market, wired it up to my arduino uno board and wrote some code to generate a ppm signal based on the joystick positions




and ended up with this...



After testing it with both the ersky 9x and the er9x, there is still a problem with the ppm in for the ersky9x (at least for mine). I don't think that it's as rough as it was before but it is noticeably worse than my er9x. With the er9x as master, you can't tell a difference when the master or student is in control. With the ersky9x the servos move normally when the master is in control but move rougher when the student is in control, this happens with both the er9x and the joystick as student.

[MikeB]

In theory, both er9x and ersky9x use the same code so they should perform about the same. I'll have a look and see if some difference has crept in. It may just be the ARM processor is so much faster it causes the effect you describe. Good testing!

Mike.

[gfnk357]

Thanks for checking into it again, I'm sure you have a big list of other things to program for the board so I appreciate you taking the time. I did try open9x and both firmwares have the same issue with the servos moving roughly with the ppm in.

[ShowMaster]

Just wondering if using another brand TX as the trainer source has the same issue?



ShowMaster
Sent from my iPod touch using Tapatalk

[Kilrah]

Well, the reason for making his arduino-joystick was to test a different kind of device, and it showed same results...

[kaos]

I would love to have that PPM-USB interface so I can use joy sitcks to fly (including throttle and rudder control). How much does it cost to make it and how hard it is? any diagram/program can be shared?

[ShowMaster]

Well, the reason for making his arduino-joystick was to test a different kind of device, and it showed same results...

I understand but is it the same 100%?
I have other brands of TXs I'll try just to be 100% but can't until this weekend.
I'm sure you're correct in your findings but I thought I tried my PCM-10 when I first got the sky board and it seemed OK. I'll try what I own with the latest sky9x and open sky9x first chance just to see what happens.



ShowMaster
Sent from my iPod touch using Tapatalk

[gfnk357]

Showmaster,
The signal I generated was an 8 channel ppm signal with a 22.5 millisecond refresh rate, 300 microseconds between pulses, and a minimum pusle width of 700 microseconds and a maximum pulse width of 1700 microseconds which I believe is the same signal properties as the original firmware. I'm thinking it's a firmware issue because both the joystick and the ersky9x work perfectly as trainers with my er9x radio as master but both my er9x and joystick have problems with the ersky9x as master. You do have a good point that it still may not be a problem with the firmware, although if that's the case it looks like I'll have a hardware bug to track down. If you're able to run another radio through the ppm-in to test your ersky9x that would be great, it would help narrow down whether the issue is a general problem or just something specific to my radio.

[gfnk357]

I would love to have that PPM-USB interface so I can use joy sitcks to fly (including throttle and rudder control). How much does it cost to make it and how hard it is? any diagram/program can be shared?

It's actually pretty easy to do and I don't have a problem sharing the code. Total cost was $10 for the joystick (used) and $20 for the arduino uno. I actually found the code for a futaba ppm generator online and then modified it for the th9x specs, although it worked with the original futaba 6 channel signal anyway. If you want to pm me your email address I can send you what I have for the code so far.

As far as level of difficulty it really is pretty easy, the Arduino UNO board has sockets you can plug wires into so it's just putting the correct wires into the correct sockets and then programming the board. Programming the board is easy but if you need to change the code it would help to have some programming skill or know someone that does. I can see an issues with the movement range of the pots, every joystick could be different and there's no calibration routine in the code right now. I'm going to try working on that over the weekend to see if I can get something to work. I also don't know if ersky9x or the other firmwares can calibrate the ppm-in signal, if it can't then you would definately want the calibration routine in the code before flying with it since center points might be off as well as endpoints.

[Kilrah]

Just some clarification, I think Kaos understood you had a USB to PPM board, but AFAIK you actually dismantled the joystick , threw the electronics out, and connected the pots directly to the arduino's A/D ports, right?

[gfnk357]

Kilrah, that's correct. When I was looking online trying to figure out how to build it I did come across a ready made cable with circuitry that you would plug your USB joystick into, and that would generate a ppm signal for the tx. It was powered through the trainer cable though, and when I went back to look I couldn't find it, but there are ready made solutions out there.

[kaos]

fgnk357:
could you give me a link of that USB to PPM converter cable? I have been looking for that for a while. thx.

[gfnk357]

I did manage to find it again and it turns out I was a little off on what it does. You connect your joystick to the computer and then run another usb cable from the computer into your transmitter. Software comes with the cable to assign the joystick axes and buttons to the channels for your transmitter. The cable is the CompuFly usb to ppm adapter cable, there's versions with a futaba cable end and a JR cable end (I think you would need JR).

The site below sells them, I'm also including some links for the RC groups forum about it and the google code page
http://www.readymaderc.com/store/index. ... cts_id=285
http://www.rcgroups.com/forums/showthread.php?t=1167074
http://code.google.com/p/joystick-to-ppm/

[kaos]

Thx for the info. that is a good start to chase it down for me. I am looking for something like your DIY thing which can be connected to joy stick directly and to the trainer port without needing computer. Use a computer to set it up is fine, just want it be able to connect and fly without attaching a computer. If you ever come across something like that please let me know. Thx again.

[gfnk357]

I actually just came across something called a ForceFly that does just what you want. I wouldn't be surprised if more things like this start popping up in the future.
http://www.nghobbies.com/cart/index.php ... cts_id=486

[kaos]

Wow, that is exactly what I am looking for and more - the feed back sensation. but for 279.00 it is a little , no , a lot steep for me at this moment.

looks like it also has a 3D cam on there too.
thx a lot. I will keep my eye on it, when the price drops to a point I can afford it , or I can't stop my urge , I will get one like that.

I am still working on my FPV system, ground test is done. Once the placement and wiring is done of the plane I will be able to maiden my FPV with goggles. Hope by the time I am used to flying FPV the price will be half of that.

Thx for the info.
Don't want to side track this thread. If I have more question relating to that I will start a new thread under FPV.

[MikeB]

Thanks for checking into it again, I'm sure you have a big list of other things to program for the board so I appreciate you taking the time. I did try open9x and both firmwares have the same issue with the servos moving roughly with the ppm in.

Long time with this, but I have just spotted a bug where the ARM code compiles differently to the AVR code, due to having 32 bit integers instead of 16-bit integers.
This is probably causing this roughness. I've notified Bertrand so the fix should get into both ersky9x and open9x.
I'll try to commit an update later today.

Mike.