Thread Tools
May 06, 2009, 03:43 AM
Registered User
Quote:
Originally Posted by MorenoEnNYC1
You just had to do it huh! Amazing! You can mount a serious piece photo of equipment on that...I see loads of mounting ideas in between those wide LG. What are your plans???

sigh...trying not to be swayed

AJ
Of course I had to do it . Versatility for heavy camera equipment is exactly what I had in mind for this setup.

The UC you see on it is ready to take a pan/tilt gimbal from photoshipone set up as Roll/Tilt. The Gimbal weighs a whopping 380g on its own (inc servos) but uses full sice digital servos so will be capable of handling any camera up to 2kg - as far as the Gimbal is concerned.

Camera for a start will be an Olympus E420 ESLR - probably with a pancake 25mm lense but heavier lenses also no problem. Camera trigger will be opperated via 3 position switch (off - focus - cont shoot).

All this will add a further ~1kg onto the MK - thus keeping it nice and agile and well within its comfort zone - which is where you want to be.

I may mod the frame to place the front two motors next to each other to give me more up tilt on the camera without the prop entering pictures. I may even get away mounting the SD9 underneath the camera tray to give enough up movement for tilt stabilisation during filming although I do think your X design is more suited to that type of work.

A
Sign up now
to remove ads between posts
May 06, 2009, 04:56 AM
Registered User
ddrake's Avatar
Quote:
Originally Posted by crashedagain80
Talking about 800 frames - here's a video of my Cayman 750 (octo) on a payload test ... it did rather well (see the video for the figures)

As usual please excuse the poor video quality - SD9 arrived today so things should be better in future.

Enjoy!

A
Mean machine!!!
Good job!

Considering the HVR-A1U ( link ) is just 690 grams (no battery) you have some serious potential there

Looking forward for footage from the SD9..
May 06, 2009, 05:14 AM
Tri-Quad-Hexa-Octo-copters!!
Alex,
Are you running those 2S 3200's in a series for 14V? Are the props counter-rotating on each arm? It looks like they might be, but I had to be sure. Where did you get the CF APC props?
Cheers,
Jim
Quote:
Originally Posted by crashedagain80
Talking about 800 frames - here's a video of my Cayman 750 (octo) on a payload test ... it did rather well (see the video for the figures)

As usual please excuse the poor video quality - SD9 arrived today so things should be better in future.

Enjoy!

A
May 06, 2009, 05:43 AM
Tri-Quad-Hexa-Octo-copters!!

XBee TX/RX


For all you data junkies out there Here is a neat thread with some info on building an XBee TX/RX. I like this part : "(PCB module for Futaba/Jr/Graupner/Multiplex, new receivers design (up to 12 channels), downlink dataloger system in study, ...) Here: https://www.rcgroups.com/forums/showthread.php?t=962320 Great future potential
Cheers,
Jim
May 06, 2009, 06:17 AM
Registered User
Thread OP
I believe the Draganflyer X6 also uses XBee for TX and RX and thus downlinks some info from the craft. Really like that little TX design they have.

ACT have 12 channel 2.4GHz modules already, but I beleive they've had to change the PPM pulsewidths to accomodate the additional channels, so they won't work off the shelf with current firmware. They certainly offer a 2.4GHz receiver with PPM output specifically for use in the multikopters.
May 06, 2009, 06:56 AM
Registered User
Thread OP
Interesting little discussion on how to adapt the FC 1.x to also have jitter free tilt AND roll (and shutter servo?) control here: http://forum.mikrokopter.de/topic-po...tml#post124580

Would only require addition of a small breakout with a HEF4017 chip, 4 wires (and some servo connectors), and a very small patch to the source code according to Holger.
May 06, 2009, 06:59 AM
Tinkering again!
MorenoEnNYC1's Avatar
Can't wait to see it all done! Will the mount be 'hard fixed' what sort of vibe dampening do you have in mind with 8 rotating props? I wonder if the heavier design alone will suppress most of the vibration...just my silly theory

Great job!

AJ

Quote:
Originally Posted by crashedagain80
Of course I had to do it . Versatility for heavy camera equipment is exactly what I had in mind for this setup.

The UC you see on it is ready to take a pan/tilt gimbal from photoshipone set up as Roll/Tilt. The Gimbal weighs a whopping 380g on its own (inc servos) but uses full sice digital servos so will be capable of handling any camera up to 2kg - as far as the Gimbal is concerned.

Camera for a start will be an Olympus E420 ESLR - probably with a pancake 25mm lense but heavier lenses also no problem. Camera trigger will be opperated via 3 position switch (off - focus - cont shoot).

All this will add a further ~1kg onto the MK - thus keeping it nice and agile and well within its comfort zone - which is where you want to be.

I may mod the frame to place the front two motors next to each other to give me more up tilt on the camera without the prop entering pictures. I may even get away mounting the SD9 underneath the camera tray to give enough up movement for tilt stabilisation during filming although I do think your X design is more suited to that type of work.

A
May 06, 2009, 08:40 AM
Registered User
@Jim - Yes the total voltage is 14.8V nominal. The props are counter rotating and using a bit more pitch on the bottom props seems to work really well - even yaw control at this weight was problem free!.

You can get the carbon props clones (APC and EPP) from Christian Linke's shop.

@AJ -Thanks - Will post some pictures of the complete system soon.

You're right somehow the heavy design certainly doesnt seem to give any more vibration than what I'm used to. Additionally the fact that the frame is made from wood means it filters out high frequency vibration fairly well. I intent to sandwich neoprene foam (apparently one of the filming industry standard for supression) between the UC and Mikrokopter and see how that goes.

The Gimbal concept is virtually identical to the one you use for tilt/roll stabilisation.

Alex
May 06, 2009, 08:50 AM
FPV Flyer
wavess's Avatar
@crashedagain80 Where did you get that cayman 750 frame? Is there an online shop with details for purchase? Ive been searching a bit and I only see 1 shop with the cayman 350
May 06, 2009, 09:33 AM
How high does it go?..duh
Pigs dont fly's Avatar
Quote:
Originally Posted by Arthur P.
Interesting little discussion on how to adapt the FC 1.x to also have jitter free tilt AND roll (and shutter servo?) control here: http://forum.mikrokopter.de/topic-po...tml#post124580

Would only require addition of a small breakout with a HEF4017 chip, 4 wires (and some servo connectors), and a very small patch to the source code according to Holger.
This is something we all need! Keep us posted Arthur

Edit...just got some HEF4017's
http://cgi.ebay.co.uk/HEF4017-4017-J...3%3A1|294%3A50


As per Tobias on the MK forum (Thanks )
Look at the circuit of the FC-ME and you see that you only need 4 wires from the HEF4017 to the FC1.* like this:

FC->HEF
J9->Pin15
J7->Pin14
Gnd->Pin13 and Pin8
+5V->Pin16

When the software is modified you have:

Servo->HEF
1->Pin2
2->Pin4
3->Pin7
4->Pin10
5->Pin1

S6->Pin5
S7->Pin6

and all should work, if not solder a capacitor to the VCC-GND of the HEF4017.
Last edited by Pigs dont fly; May 06, 2009 at 10:19 AM.
May 06, 2009, 10:39 AM
Tri-Quad-Hexa-Octo-copters!!
Here are the Decade counters at mouser: http://www.mouser.com/Search/Refine....eyword=HEF4017
Cheers,
Jim
May 06, 2009, 11:06 AM
Registered User
Quote:
Originally Posted by wavess
@crashedagain80 Where did you get that cayman 750 frame? Is there an online shop with details for purchase? Ive been searching a bit and I only see 1 shop with the cayman 350
As far as I know the 750's are only availlable on request - try contacting a guy called "Hammer" on the German forums - he might be able to help.

From what I heard the Q4 frames will also be sold in the HI shop soon so might be easiest to wait for that.
May 06, 2009, 03:13 PM
Registered User
Thread OP
Holger's suggestion makes perfect sense and partially upgrades FC1.x's to 2.x status with respect to the additional hopefully jitterfree servo channels. The diagram for the circuit is simply in the diagram form FC 2.x (ME), only missing a 100uF cap accross 5V to GND (and you might get away without, but why not put and additional 1g on the scale).

With two simple code changes in timer0.c you could use parameter8 to both turn this option on for FC1.x boards with an extarnal counter circuit connected as described above, and for both FC1.x and FC2.x you could already control the direction of the control (currently 0.74d only controls as if inverted is on):

First change which is needed for FC1.x but doesn't hinder the ME requires changing:

Code:
  #define MULTIPLYER 4
  static int16_t ServoNickOffset = (255 / 2) * MULTIPLYER; // initial value near center positon
  static int16_t ServoRollOffset = (255 / 2) * MULTIPLYER; // initial value near center positon

  if(PlatinenVersion < 20)
  {
    //---------------------------
    // Nick servo state machine
    //---------------------------
    if(!PulseOutput) // pulse output complete
to

Code:
  #define MULTIPLYER 4
  static int16_t ServoNickOffset = (255 / 2) * MULTIPLYER; // initial value near center positon
  static int16_t ServoRollOffset = (255 / 2) * MULTIPLYER; // initial value near center positon

  if((PlatinenVersion < 20) && (Parameter_UserParam8 < 128 ))
  {
    //---------------------------
    // Nick servo state machine
    //---------------------------
    if(!PulseOutput) // pulse output complete
This forces the source code to look at parameter 8 and if it is running on FC ME or if parameter8 has a a value of 128 (or higher), it would use the demultiplexing of two PWM channels for servo control. This would still allow other users to use the other bits of parameter8 for other purposes. And it could be used immediately without modification needed to the MKTool.

Second change is one line of code and removal of two commenting outs, using another bit from user parameter 8 by changing:

Code:
  //if(EE_Parameter.ServoRollCompInvert & 0x01)
  { // inverting movement of servo
    ServoRollValue += (int16_t)( ( (int32_t) 50 * MULTIPLYER * (IntegralRoll / 128L ) ) / (256L) );
  }
  /* else
  { // non inverting movement of servo
    ServoRollValue -= (int16_t)( ( (int32_t) 40 * MULTIPLYER * (IntegralRoll / 128L ) ) / (256L) );
  }
  */
to this:

Code:
  if(Parameter_UserParam8 & 0x40)
  { // inverting movement of servo if Parameter8 has 64 added to its value
    ServoRollValue += (int16_t)( ( (int32_t) 50 * MULTIPLYER * (IntegralRoll / 128L ) ) / (256L) );
  }
  /* */
  else
  { // non inverting movement of servo
    ServoRollValue -= (int16_t)( ( (int32_t) 40 * MULTIPLYER * (IntegralRoll / 128L ) ) / (256L) );
  }
  /* */
This would still allow for 32 decimal values (e.g. interpreted as the range from -16 to +16) to be used e.g. for some level of correction if someone would want to tinker even further with the code and introduce some ability to make the roll compensation more exact within a small range.

I got 2 of those chips (the DIL version, but hey, still a lot cheaper and lighter weight than an additional FC) already at the local electronics shop Still need to do some tinkering re framechanges and new camera mount, but this should be up and running pretty quickly. Will keep you posted once I-m ready to test (unless all you guys beat me to it)
Last edited by Arthur P.; Jun 25, 2009 at 09:25 AM.
May 06, 2009, 04:52 PM
Registered User
Quote:
Originally Posted by Arthur P.
Holger's suggestion makes perfect sense and partially upgrades FC1.x's to 2.x status with respect to the additional hopefully jitterfree servo channels. ...

First change which is needed for FC1.x but doesn't hinder the ME requires changing:
...
if(PlatinenVersion < 20)
{
//---------------------------
// Nick servo state machine
//---------------------------
if(!PulseOutput) // pulse output complete

to
...
if((PlatinenVersion < 20) && (Parameter_UserParam8 < 128 ))
{
//---------------------------
// Nick servo state machine
//---------------------------
if(!PulseOutput) // pulse output complete

This forces the source code to look at parameter 8 and if it is running on FC ME or if parameter8 has a a value of 128 (or higher), it would use the demultiplexing of two PWM channels for servo control. ...
Hey Arthur, very interesting. So, is the 'Nick servo state machine' the new routine for the counter? If yes, shouldn't the first line actually look like this?

if((PlatinenVersion < 20) || (Parameter_UserParam8 > 127 ))
May 06, 2009, 06:40 PM
Registered User
Thread OP
No, that's the old solution where the FC itsself is generating the servo out pulse. If parameter8 is 128 or larger (or actually if parameter8 AND 0x80 is true as I-m only using one bit there), then the logic of this if statement fails. I.e. it goes to the ELSE statement both if the FC is a version 2.x FC or this parameter8 bit is set. The section it then gets to is the one which uses the counter on the FC ME, but could use the same counter on 1.x if it is wired up to 5V, GND, J7 and J9 as described. That then results in the counter generating separate servo outputs.

Similarly the 2nd proposed change would use bit 6 to turn invert servo direction on or off for the roll servo (current code has that ON by default and no ability to control it other than by changing the source code and recompiling).

Currently the 0.74d code only anticipates a maximum of 5 servo outputs if I read it correctly (and maybe up to 8 in the future). For now I-m prototyping for 5 servos max. I believe the shutter servo still comes off of the FC so this actually still gets you 6 servos.

Unless I go for XBee or an ACT TX and RX, with the DX7 I only have 7 control channels. But I-m still interested in the additional channels. I could well see some modification of the firmware allowing multiplexing of commands over single channels. E.g. with a row of 7 switches, you could theoretically use a single channel to provide for 7 separate on/off commands, or you could split this into eg. 1 channel with 16 steps and 3 on/off controls.

I-m still dreaming of a multikopter which could be commanded to hold position and/or fly a particulr pattern and then perform a predetermined pattern of AP shots such as panoramas, 4x4 pattern, ascending pattern from predetermined altitudes, circle around target pattern, zoom in on target in discrete zoom steps, etc). Some of the patterns shouldn't need GPS per se, some would.

If you think about it, any more complex pattern and camera control quickly gets you to a) tilt servo and tilt correction (which we have but it's a bit jittery sometimes), b) roll correction (which is starting to be incorporated), c) shutter control (which we have but it is very jittery and all that jitter eats power), and d) zoom contrkol. Leaves 1 or 2 servos for e.g. changing some other camera setting, turning on or off a higher capacity flash or infrared lights, or releasing a couple of postcards advertising your capabilities to the onlookers


Quick Reply
Message:

Thread Tools