HobbyKing.com New Products Flash Sale
Reply
Thread Tools
Old Aug 15, 2010, 12:37 PM
Registered User
Alexinparis's Avatar
Paris, France
Joined Jun 2009
756 Posts
Sure, it is possible to find a smarter way of averaging values.
I think 4 values is a good compromise between response time and smooth input. A bit shift is still possible with 4 values
I'm now sure the problem with some corona receivers is no tied to some odd values coming from receiver, so removing the highest and lowest values might nor help for this problem.

Quote:
Originally Posted by ddrake View Post
@Alex: do you plan to do a version to work with modded receivers that have all channels summed up in one sequence (as MK does) and using just one pin of the Arduino?
Am doing some coding on this (as I have 3 such receivers... ) but response is not as smooth as in MK code.
Some months ago, I suggested a code to adapt AeroQuad software to handle a single RX PPM pulse train. I and someone else tested it successfully.
As it is not one of my need (we can't get directly a PPM pulse train from a ASSAN RX), I didn't integrate it for the triwii but it's not so complicated to do.
http://aeroquad.com/showthread.php?2...er-pulse-train
Alexinparis is offline Find More Posts by Alexinparis
Last edited by Alexinparis; Aug 15, 2010 at 01:02 PM.
Reply With Quote
Sign up now
to remove ads between posts
Old Aug 15, 2010, 02:30 PM
Registered User
ddrake's Avatar
Milan
Joined Mar 2007
381 Posts
Quote:
Originally Posted by Alexinparis View Post
...
Some months ago, I suggested a code to adapt AeroQuad software to handle a single RX PPM pulse train. I and someone else tested it successfully.
As it is not one of my need (we can't get directly a PPM pulse train from a ASSAN RX), I didn't integrate it for the triwii but it's not so complicated to do.
http://aeroquad.com/showthread.php?2...er-pulse-train



Would it be too much trouble to integrate it in the TriWiiConf too?
Will do some tries while I wait for the ESCs...
Thank you again for your great work!

Regards
ddrake is offline Find More Posts by ddrake
Reply With Quote
Old Aug 16, 2010, 04:03 AM
Friend is a 4 letter word...
JussiH's Avatar
3rd stone from the sun
Joined May 2007
2,676 Posts
Quote:
Originally Posted by Alexinparis View Post

Some months ago, I suggested a code to adapt AeroQuad software to handle a single RX PPM pulse train. I and someone else tested it successfully.
As it is not one of my need (we can't get directly a PPM pulse train from a ASSAN RX), I didn't integrate it for the triwii but it's not so complicated to do.
http://aeroquad.com/showthread.php?2...er-pulse-train
I got Arduino Pro Mini today, so setting up to build it the "right" way....

On the PPM: Id really like to see a version that can use Spektrum satellite RX....sadly my programming skills are not up to make that code...apparantly the data is serial and it could be a very good alternative to PPM streaming!

Spektrum satellite RX data:

http://diydrones.ning.com/profiles/b...ogPost%3A64228

More on binding and data formats for Spektrum satellite RX...

http://www.rcgroups.com/forums/showthread.php?t=922566
JussiH is offline Find More Posts by JussiH
RCG Plus Member
Old Aug 16, 2010, 04:42 AM
Flightless crawler
kuba_em's Avatar
Poland
Joined Apr 2010
139 Posts
Quote:
Originally Posted by Alexinparis View Post
Some months ago, I suggested a code to adapt AeroQuad software to handle a single RX PPM pulse train. I and someone else tested it successfully.
As it is not one of my need (we can't get directly a PPM pulse train from a ASSAN RX), I didn't integrate it for the triwii but it's not so complicated to do.
Hi all,
This exactly what I want to do for my 7-channel Futaba/Robbe R6107SP 2.4GHz RASST Serial. I modified code of TriWiiCopter with use of procedures from arduino.cc forum. User "mem" shared your servodecode library (PPM pulse train from RX) and libraty with servo operate on timer2. I hope, that in original TriWiiCopter code will be sufficient to change "rc functions" and ISR(TIMER0_COMPA_vect) -> timer2 and some of variables.
At this moment I can't to test new pde file, if somebody want to try I can share.
(I apologize for my lousy english...)
Best regards,
Kuba
kuba_em is offline Find More Posts by kuba_em
Reply With Quote
Old Aug 16, 2010, 05:19 AM
Registered User
Alexinparis's Avatar
Paris, France
Joined Jun 2009
756 Posts
Quote:
On the PPM: Id really like to see a version that can use Spektrum satellite RX....sadly my programming skills are not up to make that code...apparantly the data is serial and it could be a very good alternative to PPM streaming!
It would be much more complicated with a spektrum satellite.
The little atmel 328p is not strong enough to handle in soft a 115200 baud serial stream + the triwii loop.
So the only way would be to reuse the hardware serial decoder, not very convenient and incompatible with a simultaneous use of GUI or LCD.

Quote:
This exactly what I want to do for my 7-channel Futaba/Robbe R6107SP 2.4GHz RASST Serial. I modified code of TriWiiCopter with use of procedures from arduino.cc forum. User "mem" shared your servodecode library (PPM pulse train from RX) and libraty with servo operate on timer2. I hope, that in original TriWiiCopter code will be sufficient to change "rc functions" and ISR(TIMER0_COMPA_vect) -> timer2 and some of variables.
At this moment I can't to test new pde file, if somebody want to try I can share.
I wrote a specific servo drive code based on TIMER0, especially because TIMER1 and TIMER2 are already used to drive the ESCs with analogWrite. So moving it for a TIMER2 code wont work.
Hopefully, the code I suggested to decode a PPM pulse train is compatible because it uses no timer, its just a variant of the current RC code with only one digital PIN.
Alexinparis is offline Find More Posts by Alexinparis
Reply With Quote
Old Aug 16, 2010, 06:47 AM
Friend is a 4 letter word...
JussiH's Avatar
3rd stone from the sun
Joined May 2007
2,676 Posts
Quote:
Originally Posted by Alexinparis View Post
It would be much more complicated with a spektrum satellite.
The little atmel 328p is not strong enough to handle in soft a 115200 baud serial stream + the triwii loop.
So the only way would be to reuse the hardware serial decoder, not very convenient and incompatible with a simultaneous use of GUI or LCD.

.
So that would need to use a second ProMini to decode the spektrum serial data, then...? I see that is not very convenient then....thanks for clearing that up!
JussiH is offline Find More Posts by JussiH
RCG Plus Member
Old Aug 16, 2010, 07:13 AM
Flightless crawler
kuba_em's Avatar
Poland
Joined Apr 2010
139 Posts
Quote:
Originally Posted by Alexinparis View Post
I wrote a specific servo drive code based on TIMER0, especially because TIMER1 and TIMER2 are already used to drive the ESCs with analogWrite. So moving it for a TIMER2 code wont work.
Hopefully, the code I suggested to decode a PPM pulse train is compatible because it uses no timer, its just a variant of the current RC code with only one digital PIN.
Thank you Alex for explanation, so now I'm sure that my code is worng and why.
I will try to check your suggestions from aeroquad.com.
kuba_em is offline Find More Posts by kuba_em
Reply With Quote
Old Aug 16, 2010, 07:52 AM
Registered User
ddrake's Avatar
Milan
Joined Mar 2007
381 Posts
Quote:
Originally Posted by Alexinparis View Post
It would be much more complicated with a spektrum satellite.
The little atmel 328p is not strong enough to handle in soft a 115200 baud serial stream + the triwii loop.
So the only way would be to reuse the hardware serial decoder, not very convenient and incompatible with a simultaneous use of GUI or LCD.
...
Which makes me wonder when will projects start to move from Arduino to more powerful boards.

Here are three interesting prototype boards:
http://www.xgamestation.com/view_product.php?id=51
Arduino code compatible, with VGA and Analog video out.
Have still to find out if it could manage video overlay so also act as OSD

http://mbed.org/handbook/Tour
nice, fat Cortex M3 processor on tiny board

http://www.coridiumcorp.com/catalog/...products_id=84
Coridium board, Cortex M3 based too.

I hope this isn't too much OT

Regards
ddrake is offline Find More Posts by ddrake
Reply With Quote
Old Aug 16, 2010, 03:31 PM
xxxxxxxxxxxxx
JUERGEN_'s Avatar
Deutschland
Joined Nov 2009
660 Posts
Wlalkera rx2610s

or,

a excelent base, the WALKERA RX2610S . . .with IDG650 and ISZ650 ATMEGA168



- http://www.rcgroups.com/forums/showp...&postcount=700


juergen
JUERGEN_ is offline Find More Posts by JUERGEN_
Last edited by JUERGEN_; Aug 16, 2010 at 03:38 PM.
Reply With Quote
Old Aug 16, 2010, 06:41 PM
five by five
sonicj's Avatar
Panama City Bch, FL
Joined Jan 2009
2,074 Posts
Quote:
Originally Posted by JUERGEN_ View Post
or,

a excelent base, the WALKERA RX2610S . . .with IDG650 and ISZ650 ATMEGA168
huh? this project already uses a more capable atmega 328...
sonicj is offline Find More Posts by sonicj
Reply With Quote
Old Aug 16, 2010, 07:17 PM
I think my wheels fell off.
Thumb_number's Avatar
Perth
Joined Oct 2009
894 Posts
I got my TriWii brain into my own frame on the weekend. All is not well. I need some help to determine if my problems are from using a 36Mhz Corona or if I have a bad / overloaded ESC.

From the video you see all the failures seem to be a cut-out of the rear motor. The ESC (Mystery 20A) on that motor does seem to be 'warmer' than the other two but not what I would call 'hot'. The ESC should not be working too hard and the front two ESC have never cut-out like this. The problem can happen in the first minute of a flight or after 5 mins, there is no real pattern.

When on the GUI, there seems to never be this issue (mind you, have never had the motors turning while on GUI, only looking for radio 'glitches'). The event seems to look very similar to Logiq's from post #222. Can this cut-out occur on only one channel? Same channel every time?

I have some new HK SS 18-20 ESC that I will change over and see if that is the problem, but if any of you guy's with more experience have any ideas, I would be grateful.

Thanks,
George

http://vimeo.com/14179855
Thumb_number is offline Find More Posts by Thumb_number
Reply With Quote
Old Aug 17, 2010, 02:33 AM
Registered User
Alexinparis's Avatar
Paris, France
Joined Jun 2009
756 Posts
Quote:
Originally Posted by Thumb_number View Post
I got my TriWii brain into my own frame on the weekend. All is not well. I need some help to determine if my problems are from using a 36Mhz Corona or if I have a bad / overloaded ESC.

From the video you see all the failures seem to be a cut-out of the rear motor. The ESC (Mystery 20A) on that motor does seem to be 'warmer' than the other two but not what I would call 'hot'. The ESC should not be working too hard and the front two ESC have never cut-out like this. The problem can happen in the first minute of a flight or after 5 mins, there is no real pattern.

When on the GUI, there seems to never be this issue (mind you, have never had the motors turning while on GUI, only looking for radio 'glitches'). The event seems to look very similar to Logiq's from post #222. Can this cut-out occur on only one channel? Same channel every time?

I have some new HK SS 18-20 ESC that I will change over and see if that is the problem, but if any of you guy's with more experience have any ideas, I would be grateful.

Thanks,
George

http://vimeo.com/14179855
Hi,
It looks like your problem is similar to Logiq's one.
Only one specific glitch could be the cause of a motor off command. It is not always detectable via GUI is the duration of the glitch is very small.
However, a motor off command should affect always the whole system, not only one motor. strange..
I'm working on a new release to prevent the glitches from shutting off the motors. It should help, I hope.
Alexinparis is offline Find More Posts by Alexinparis
Reply With Quote
Old Aug 17, 2010, 03:43 AM
five by five
sonicj's Avatar
Panama City Bch, FL
Joined Jan 2009
2,074 Posts
Quote:
Originally Posted by Thumb_number View Post
From the video you see all the failures seem to be a cut-out of the rear motor. The ESC (Mystery 20A) on that motor does seem to be 'warmer' than the other two but not what I would call 'hot'. The ESC should not be working too hard and the front two ESC have never cut-out like this. The problem can happen in the first minute of a flight or after 5 mins, there is no real pattern.
swap one of the front esc for the rear esc. if the problem is in the esc, it will follow the esc. do the same for the motor and rx.

-remove the servo connector and check that all wires were crimped correctly.
-remove heatshrink from esc to inspect wire terminations at the board. a stray strand of copper, wires not properly tinned, jacket pulled too far back, etc.
-inspect all solder joints with a loupe.

-sj
sonicj is offline Find More Posts by sonicj
Reply With Quote
Old Aug 17, 2010, 07:43 AM
I think my wheels fell off.
Thumb_number's Avatar
Perth
Joined Oct 2009
894 Posts
Quote:
Originally Posted by Alexinparis View Post
Hi,
It looks like your problem is similar to Logiq's one.
Only one specific glitch could be the cause of a motor off command. It is not always detectable via GUI is the duration of the glitch is very small.
However, a motor off command should affect always the whole system, not only one motor. strange..
I'm working on a new release to prevent the glitches from shutting off the motors. It should help, I hope.
Thank you Alex,

I thought I may have been immune to the Corona problem as the Rx has done at least 1.5 hours of work on a different frame with your Wii brain without issue. I do understand from the image posted showing the pulse structure the issue with corona's, so I encourage your work on that issue. I hope that it is not my issue but hope your work can remove it as a concern.

Quote:
Originally Posted by sonicj View Post
swap one of the front esc for the rear esc. if the problem is in the esc, it will follow the esc. do the same for the motor and rx.

-remove the servo connector and check that all wires were crimped correctly.
-remove heatshrink from esc to inspect wire terminations at the board. a stray strand of copper, wires not properly tinned, jacket pulled too far back, etc.
-inspect all solder joints with a loupe.

-sj
Thanks sj,

I have gone to lengths to check all or most possible cause's. New to Tri's but have been punished enough to check most mentioned with my fixed wing FPV adventures.
I will replace all ESC with new one's as a matter of course as I'm not overly impressed with the Mystery units anyway. I hope that will cure this problem.

Again, being new to tri's, I have to say again to Alex, what a great system, so cheap yet can get a 'noob' in the air and enjoying the multi-rotor experience with minimal fuss.

Cheers,

George
Thumb_number is offline Find More Posts by Thumb_number
Reply With Quote
Old Aug 17, 2010, 04:42 PM
Registered User
Alexinparis's Avatar
Paris, France
Joined Jun 2009
756 Posts
Hello,

Here is the first release with the integration of a Wii Nunchuk
This release is backward compatible with a WM+ only config and this should be the case for the others to come.

GUI with a gyro only config


GUI with a WM+ and Nunchuk config


The nunchuk connection to the WM+ is described here: (I will do a nice scheme later with the right nunchuk orientation)
http://aeroquad.com/showthread.php?1...ull=1#post2324

My prototype is in image attachment. Please do not joke about the wire arrangements

Here is the list of improvements:
  • if you want to reverse your servo, comment/uncomment the 2 lines
    #define SERVO_DIRECTION 1
    //#define SERVO_DIRECTION -1
  • the throttle is now calibrated to match the range 1100(min motor)-2000(max motor) for every brand of ESC. => the throttle stick range and the altitude control is better
  • RC input is smoother (50Hz now with the average of the 4 previous values, instead of 12.5Hz before)
  • motor off/on command must be maintained for approximately 0.5s. this prevents bad motor off command due to RC noise. Another small verification is done in RC loop. => better for corona receivers even if there remains some glitches
  • Nunchuk is not required and the Triwii can fly without it just as before => gyro only configuration can support the new soft.
    At the initialization, the triwii soft detects the presence of the nunchuk and adapt the I2C communication to the good configuration.
    The soft is optimized for the higher refresh rate of a gyro only config
  • nunchuk autostable mode is now supported.
    • it tries to maintain a fix attitude angle if your RC ROLL/PITCH sticks are in the range [1350;1650], so you can still use trims to adjust a perfect level
    • it is activated once the channel 5 is upper to 1700. => you need at least a 5 channels RC. The channel 5 must be connected to digital PIN6
    • the tricopter is moving in the GUI according to the artificial horizon calculated by accelerometers
    • the tricopter must be as flat as possible when you power it to "learn" the good horizon. But you can still calibrate it latter by radio stick command
    • a new param in the GUI defines the strength of accelerometers correction
    • note that autostable mode does not prevent the tricopter from flat drifting !
    • the autolevel stabilization is far from being perfect but is works and can prevent you from crashing if you loose the tricopter orientation
    • The nunchuk coming from ebay can also be wrong copies, it's just like the Wii motion plus.
  • calibration procedure is different and more precise
  • an auto calibration of PITH/ROLL gyro is done permanently if nunchuk is present. It cancels the gyro drift problems during a flight even is auto stable switch is not activated.

The versions are updated here:
http://radio-commande.com/internatio...copter-design/
Alexinparis is offline Find More Posts by Alexinparis
Reply With Quote
Reply


Thread Tools