Page 1 of 3

M4 swarm

PostPosted: Mon Feb 16, 2015 1:26 am
by r0sewhite
I just wanted to share my latest project with you. The University asked me for some small copter for swarm experiments. Since they want to load some additional electronic, I found my Leora frame to be a little too small while Rachel is too big with a span of 35cm.

So I developed a new frame with a 240mm mtm span. It looks pretty much like a streched Leora but is different in some details. 2.5mm was not stiff enough for the longer arms so I took 3.0mm plates which gives rock solid arms. I expect the frame itself to be almost as indestructable as Leora where we only sold one spare arm after more than 300 sold units.

Besides the 30x30mm mounting holes I have added a standard 45x45mm hole pattern which allowed me to mount a simple battery cage which is also good as landing gear. As an add-on I milled a simple top plate which may be useful as a carrying platform for electronics or simply as a protection cage for the M4.

The copters are equipped with M4, Jörg's expansion board, 12A KISS ESCs and Tiger MN1806 motors. The first indoor hover was scary: The copter was only oscillating slightly but after only 15 seconds hovering the motors were hot enough to almost light a cigarette at them. :shock:

After some PID tuning now the copters fly smooth and agile and the motor graphs look perfect. Indoor flight time with a 2S 1250mAh lipo (AUW: 312g) is 14-15 minutes. I did choose the PID controller for Quatos because the guys will surely change the payload many times and surely don't want to calculate masses everytime.

I didn't plan to release this frame but if anyone is interested in them for a special project, just let me know.






Re: M4 swarm

PostPosted: Mon Feb 16, 2015 4:57 am
by Mullet
Looks awesome! I'd be interested in a frame like this.

Re: M4 swarm

PostPosted: Mon Feb 16, 2015 6:58 am
by aBUGSworstnightmare
Hi Tilman,

great to see 'the whole bunch' next to each other! 8-)

Would you mind posting a detail shot of how you've mounted the GPS antenna? Would give an idea to some of the people here how to do it.


P.S,: MN1806 + KISS is quite common for 250size FPV quads! What about sharing your PIDs?

Re: M4 swarm

PostPosted: Mon Feb 16, 2015 8:12 am
by kinderkram
Cool & clean build, Tilman!
Looks like it will make a good reference build for the M4 Wiki. ;)

250ers are clearly a trend atm. though rather used for FPV racers. 312gr AUW leaves some room for payload and 15mins flight time for the said lipo is great. Did you also try 3s? Are you using 6" props?

btw: why didn´t you use the 35mm ceramic patch antennas? The extra weight should´t be a problem here...

Re: M4 swarm

PostPosted: Mon Feb 16, 2015 11:02 am
by r0sewhite
@Mullet: My offer was rather meant for other projects like this, in other words: lots of frames and not only one. Let's wait and see if some others would like to have one, too and then I will make some. The smallest number is always 8 frames.

@Joerg: Of course I can share my PIDs. I started with your set which was much better than the defaults. Motors still got warm but not hot anymore. I first decreased angle and rate D until my motor graphs were as tight as possible and then I was able to increase angle P to over 300. A higher angle I value finally gave me a snappy ACC.

Code: Select all
CTRL_TLT_ANG_D                        3000
CTRL_TLT_ANG_DM                         900
CTRL_TLT_ANG_F                        0.25
CTRL_TLT_ANG_I            0.00700000021607
CTRL_TLT_ANG_IM                         360
CTRL_TLT_ANG_OM                        1200
CTRL_TLT_ANG_P                         320
CTRL_TLT_ANG_PM                        1200
CTRL_TLT_RTE_D                        6000
CTRL_TLT_RTE_DM                        9999
CTRL_TLT_RTE_F                        0.25
CTRL_TLT_RTE_I                           0
CTRL_TLT_RTE_IM                        9999
CTRL_TLT_RTE_OM                        3000
CTRL_TLT_RTE_P                           0
CTRL_TLT_RTE_PM                        9999

Since GPS is only for playing and perfect reception is not required, I just milled a small carrier board from CF. I found some M2x5mm nylon spacers to put between the M4 and the GPS carrier. The space between your expansion board and the M4 is a little more than 5mm but an extra washer helped me to also put spacers between the M4 and your board so I could tighten the screws (pictures do not show spacers between boards. I've got the nylon washers after I made these pictures).



@Norbert: GPS reception is not so critical. I just added the small patch antenna for them so that they can play with GPS navigation but usually they will fly indoor and put some boards on top. It just didn't make sense to use a bigger patch antenna.

I've also tried a 3S 1000mAh lipo which is almost the same size (only 1mm thicker) but even with the 2S lipo I've got a great thrust/weight ratio so it just doesn't make sense to go for 3S. But it's good to know that they can upgrade to 3S if they want to add more than the planned 100g. I think they can use 2S as long as the throttle scale is below the 0.7 default value.

Re: M4 swarm

PostPosted: Mon Feb 16, 2015 7:46 pm
by JussiH
Nice ones, Tilman....Leora L?

I know you said its not important to your client, but nevertheless, I am curious as to what kind of HACC and VACC you can acheive with that antenna setup?

I found that on my Leora with a 25mm antenna, that a 52mm groundplane was insufficient and I had to increase it to 65 to get good reception and Hacc below 1.


Re: M4 swarm

PostPosted: Mon Feb 16, 2015 8:32 pm
by r0sewhite
Hi Jussi,

no problem, outdoor flights including mag calibration and nav PID tuning is planned for the next 2 days and then I can give you feedback.

Btw: I call the frame Saphira ;)

Re: M4 swarm

PostPosted: Mon Feb 16, 2015 8:54 pm
by kinderkram
"Safari with Saphira" - would make a nice title for a through-the-bushes FPV flight video. :D

I garantee those students will waste no time and go outside for autonomous flights.

First lesson: how to create a proper ground plane. ... PS-Antenna

btw: I guess AQ is "project enough" for a group buy? ;)

Re: M4 swarm

PostPosted: Mon Feb 16, 2015 10:31 pm
by JussiH

Quatos will let you get away with quite a lot of error before things go bad...I guarantee you that it will still fly much better than PID, even with a 100 gram error ;)

So if you build a good profile profile and tune the craft well to begin with, adding 100 grams of balanced payload is not gonna make much difference to the flight characteristics on such a small and agile copter.

Of course the right way is to do it properly with profiles for each payload, but it IS an ADAPTIVE controller and it does just that: Adapt :D

Re: M4 swarm

PostPosted: Mon Feb 16, 2015 11:03 pm
by r0sewhite
Jussi, you may be right. I discussed that with Peter a while ago and he agreed that for now the PID controller ist the safer option because we have lots of experience with it.

I have built an 11th copter so we can go on with testing and I will surely switch their copters to Quatos when we feel that it's the better choice for them but there is no need to do it now. I can do it anytime.