Thread Tools
Old Jul 09, 2012, 04:31 AM
savaga is offline
Find More Posts by savaga
Registered User
Is it possible to use PPMSUM on this board?
Sign up now
to remove ads between posts
Old Jul 09, 2012, 05:09 AM
Rober-t is offline
Find More Posts by Rober-t
Registered User
as you can see in the nice manual quadframes made it's possible.
Old Jul 09, 2012, 03:35 PM
Quadframes is offline
Find More Posts by Quadframes
Registered User
Quadframes's Avatar
Quote:
Originally Posted by bluelightning View Post
Thanks for your advice. I have used the 5V from the I2C plug because, If I power the AIOP with the USB (during the dev phase) there is no voltage ouput from the S1-S3 plug, so the GPS is not powered by this way.

Attached the successful test with the GPS EM406
Hi yes I know that

The only way to get power to the S1-S3 is if you power via the extended power holes as seen in this picture Then your GPS will have power and that way you dont have to use I2c to power it
( YOU MUST REMOVE THE YELLOW PIN) when powering this way and use a UBEC
Last edited by Quadframes; Jul 09, 2012 at 03:50 PM.
Old Jul 10, 2012, 05:44 AM
Cesco is offline
Find More Posts by Cesco
Registered User
Cesco's Avatar
Working great so far except ..

The Z acc value calibrates to 256, but reversing the board will give a -780 reading. This is obviously wrong, the calibration value should be 512, not 256. This also makes a very twitchy board at 90 deg angles. Where do i change that?

The baro did knock me off my chair. How is that accuracy possible?

Edit:
i did change in sensors.c, function ACC_init

#if defined(FREEIMUv043)
acc_1G = 512;
#else
#if defined(CRIUS_AIO_PRO_V1)
acc_1G = 512;
#else
acc_1G = 255;
#endif
#endif

Works. Upside reading 512, reversed board reading -530.
But the twitchieness at facing 90 deg up, down or sideways stays. Must be another issue.
Last edited by Cesco; Jul 10, 2012 at 05:59 AM.
Old Jul 10, 2012, 07:10 AM
Rober-t is offline
Find More Posts by Rober-t
Registered User
Cesco great! I ordered my board today and am looking forward to playing with it soon.

issue with the board at 90:

if you calculate the angle out of the accelero vector you'll have to use sin/cos at some point.
cos at 90 will be around zero. this very small value will lead to very unusual behavior in the measurement of the angle.

there are possibilities to avoid the cos zero crossing by using cosine direction matrix or quaternions for calculating the orientation, unfortunately our chip will not be able to calculate that in time. so just don't fly at 90 tilt or use heading hold for that.
Old Jul 10, 2012, 08:13 AM
Cesco is offline
Find More Posts by Cesco
Registered User
Cesco's Avatar
I think it would be a good idea to have A8 as input for PPM Sum instead of one of the serial ports. Gnd, 5V, A8 is in a row, perfectly lined up to connect my frsky receiver with one standard 3 pin servo cable. No soldering.

btw. I do use a nokia usb cable. The normal mini-usb would not fit.
Last edited by Cesco; Jul 10, 2012 at 08:31 AM.
Old Jul 10, 2012, 08:19 AM
bluelightning is offline
Find More Posts by bluelightning
VTOL UAV designer
bluelightning's Avatar

ArduplaneNG v2.4 now on AIOP v1.0


Hello,
I have just done a successful porting of the ArduPlane v2.4 to the AIOP v1, so it is now called ArduPlaneNG v2.40.
All things are working well...
I have used:
- a Mediatek MT3329 GPS
connected to RX2,TX2 from the S1-S3 plug and powered by the I2C plug (+5Vdc, ground)
- a airspeed sensor MVPX7002DP
connected to A0 from the A0-A5 plug and powered by the I2C plug (+5Vdc, ground)
- a turnigy receiver 9X8Cv2
This has been fully tested with the AP mission planner 1.1.96 mavlink 1.0
soon tested in flight on a Easyglider... Stay tuned...
see photo attached:
Last edited by bluelightning; Jul 10, 2012 at 02:54 PM.
Old Jul 10, 2012, 08:38 AM
Orphee is offline
Find More Posts by Orphee
Registered User
Quote:
Originally Posted by Rober-t View Post
Cesco great! I ordered my board today and am looking forward to playing with it soon.

issue with the board at 90:

if you calculate the angle out of the accelero vector you'll have to use sin/cos at some point.
cos at 90 will be around zero. this very small value will lead to very unusual behavior in the measurement of the angle.

there are possibilities to avoid the cos zero crossing by using cosine direction matrix or quaternions for calculating the orientation, unfortunately our chip will not be able to calculate that in time. so just don't fly at 90 tilt or use heading hold for that.
So currently the code for MultiWii with AIO Pro given by Quadframes (thanks to him) is not stable ?
still need adjustment ?

Thank you
Old Jul 10, 2012, 09:07 AM
Rober-t is offline
Find More Posts by Rober-t
Registered User
Orphee, i would not say it's unstable, but it becomes inaccurate around 90 nick or roll in stable mode, as you should use cosine for that calculation. It's a common problem in PID controllers that only use the cartesian coordinate system for describing the current position in space (there is no top or bottom of the copter)

you never fly around 90 in stable mode, common angles go up to maybe 60 and that is already accro-stable :P

there is no way with cartesian coordinates to make it accurate around 90 AND 0 or it's a hack (f.e. resetting the acc vector close to 90 (swapping axes) / using sin on gyro and cos on acc etc.) BUT you'll always have more than one acceleration if you come to 90 inflight, so the 90 will be inaccurate as long as you look at the accelero itself. You'll have to use gyro integral as well and merge it to the acc value etc etc etc

code is good so far, don't worry
Last edited by Rober-t; Jul 10, 2012 at 09:19 AM.
Old Jul 10, 2012, 09:13 AM
Kipkool is offline
Find More Posts by Kipkool
Old Flying Bastard
Kipkool's Avatar
This is a mpu6050 chip inside, we fly with it for month without problem with many boards, I have doubt the default code is not correct.
But as AccZ is not used for now in mwc, perhaps that's why nobody corrected it before ?

Just flied the board quickly with default code, default pid, everything looks smooth and leveled.

Just a big bad point here: The micro usb detached from the board when I've plugued it for the 4th time. Bad soldering. And the 2 gnd pad at the side of the usb plug broke at the same time ). Not really solid so be care.

I have to use the ftdi port for now.
Old Jul 10, 2012, 09:24 AM
hpcre is offline
Find More Posts by hpcre
Registered User
i'll be buying a board soon and want to buy a GPS with it, which GPS board have you guys tried and found well worth the money ? Mediatek MT3329 ? or the U-Blox-06 or other ?
Old Jul 10, 2012, 10:32 AM
Orphee is offline
Find More Posts by Orphee
Registered User
Quote:
Originally Posted by Rober-t View Post
Orphee, i would not say it's unstable, but it becomes inaccurate around 90 nick or roll in stable mode, as you should use cosine for that calculation. It's a common problem in PID controllers that only use the cartesian coordinate system for describing the current position in space (there is no top or bottom of the copter)

you never fly around 90 in stable mode, common angles go up to maybe 60 and that is already accro-stable :P

there is no way with cartesian coordinates to make it accurate around 90 AND 0 or it's a hack (f.e. resetting the acc vector close to 90 (swapping axes) / using sin on gyro and cos on acc etc.) BUT you'll always have more than one acceleration if you come to 90 inflight, so the 90 will be inaccurate as long as you look at the accelero itself. You'll have to use gyro integral as well and merge it to the acc value etc etc etc

code is good so far, don't worry
Ok thank you.
Il will try it with 2.1 RC 964
Old Jul 10, 2012, 03:21 PM
Kipkool is offline
Find More Posts by Kipkool
Old Flying Bastard
Kipkool's Avatar
Just a quick test of Crius pro with mwc 2.1 level model this afternoon, in a strong wind.

This mpu6050 thing is rock stable in all conditions here.


Hunter Vtail 500 - ready to rumble (4 min 41 sec)
Old Jul 10, 2012, 04:27 PM
nebstress is offline
Find More Posts by nebstress
Advanced Monkey
nebstress's Avatar
Quote:
Originally Posted by Kipkool View Post
Just a quick test of Crius pro with mwc 2.1 level model this afternoon, in a strong wind.

This mpu6050 thing is rock stable in all conditions here.


http://www.youtube.com/watch?v=bJApP...&feature=g-upl
Nice video & great V tail!
Last edited by nebstress; Jul 10, 2012 at 04:35 PM.
Old Jul 10, 2012, 05:17 PM
Toumal is offline
Find More Posts by Toumal
Registered User
Hey everyone,

I've written up a Howto for the Crius AIO for MultiWii 2.1 on fpvcentral.net - special emphasis on using a u-blox GPS and configuring that for the higher update and baud rates.

@Quadframes, I've credited you for your own tutorial writeup. Feel free to incorporate stuff from my howto if you like!


--> http://fpvcentral.net/2012/07/crius-...ltiwii-part-1/









Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Mini-HowTo Arduino and HobbyKing Multi-Rotor Control Board V3.0 or KKmulticontroller v5.5 pprioverde Multirotor Talk 31 Apr 14, 2016 02:58 AM
Discussion New updated Crius Multiwii SE flight controller SGsoar Multirotor Electronics 72 May 15, 2013 09:36 PM
Discussion Jim, any plans for more multi rotor flight controller? chris24g Xtreme Power Systems 12 Jun 25, 2012 09:11 PM
Sold One New, one used ArduPilot 2560 APM1.0 Flight Controlls/autopilots wingspinner FPV Equipment (FS/W) 3 May 28, 2012 01:55 AM
Question Quad Build Using HobbyKing Multi-Rotor Control Board V3.0 (Atmega328 PA) SJackson Multirotor Talk 27 Apr 30, 2012 09:51 PM