Build Club: HJ-Y3 Carbon Fiber Tricopter

[Hands0n]

Ooh, yes please to the PID values.

That build looks absolutely fantastic. I love what you did with the camera.

COG is a complete pain in the arris with this model. But you';ve shown that some careful positioning sorts so that out.

Have you some anything with the front arms locking mechanism? The suggestion earlier in this thread is not to lock in place but allow friction to do the job. It might have saved me a shattered arm plate if the arm could have moved on impact

Sent from my Nexus 5 using Tapatalk

[andythilo]

Thanks Danny. There was a bit of head scratching I have to say. Not much room so you need to get inventive. I also didn';t want to see props in the FPV so it had to be out front. My fronts are locked in but I am wondering if I fitted some nylon screws instead, they would shear on impact? Ordered some 9x5';s now to try, the 8x4.5s don';t have enough chuff for my liking. Slightly higher kV motors might have been better as well, 1200 range.

Here you go, screenshots + full CLI list. One thing I have set is tri_unarmed_servo = 0. This stops the servo from moving about when it';s unarmed.





Current settings:
looptime = 3500
emf_avoidance = 0
midrc = 1520
minthrottle = 1150
maxthrottle = 1990
mincommand = 1000
mincheck = 1100
maxcheck = 1900
deadband3d_low = 1406
deadband3d_high = 1514
neutral3d = 1460
deadband3d_throttle = 50
motor_pwm_rate = 400
servo_pwm_rate = 250
retarded_arm = 0
flaps_speed = 0
fixedwing_althold_dir = 1
reboot_character = 82
serial_baudrate = 115200
softserial_baudrate = 9600
softserial_1_inverted = 0
softserial_2_inverted = 0
gps_type = 0
gps_baudrate = 0
serialrx_type = 0
telemetry_provider = 0
telemetry_port = 0
telemetry_switch = 0
vbatscale = 110
currentscale = 400
currentoffset = 0
multiwiicurrentoutput = 0
vbatmaxcellvoltage = 42
vbatmincellvoltage = 33
power_adc_channel = 0
align_gyro = 0
align_acc = 0
align_mag = 0
align_board_roll = 0
align_board_pitch = 0
align_board_yaw = 0
yaw_control_direction = 1
acc_hardware = 0
max_angle_inclination = 500
moron_threshold = 32
gyro_lpf = 42
gyro_cmpf_factor = 600
gyro_cmpfm_factor = 250
pid_controller = 0
deadband = 0
yawdeadband = 0
alt_hold_throttle_neutral = 40
alt_hold_fast_change = 1
throttle_correction_value = 0
throttle_correction_angle = 800
rc_rate = 90
rc_expo = 65
thr_mid = 50
thr_expo = 0
roll_pitch_rate = 50
yaw_rate = 40
tpa_rate = 10
tpa_breakpoint = 1500
failsafe_delay = 10
failsafe_off_delay = 200
failsafe_throttle = 1200
failsafe_detect_threshold = 985
rssi_aux_channel = 0
yaw_direction = 1
tri_unarmed_servo = 0
gimbal_flags = 1
acc_lpf_factor = 4
accxy_deadband = 40
accz_deadband = 40
acc_unarmedcal = 1
acc_trim_pitch = -38
acc_trim_roll = -14
baro_tab_size = 21
baro_noise_lpf = 0.600
baro_cf_vel = 0.985
baro_cf_alt = 0.965
mag_declination = 0
gps_pos_p = 11
gps_pos_i = 0
gps_pos_d = 0
gps_posr_p = 20
gps_posr_i = 8
gps_posr_d = 45
gps_nav_p = 14
gps_nav_i = 20
gps_nav_d = 80
gps_wp_radius = 200
nav_controls_heading = 1
nav_speed_min = 100
nav_speed_max = 300
nav_slew_rate = 30
p_pitch = 50
i_pitch = 30
d_pitch = 30
p_roll = 50
i_roll = 30
d_roll = 30
p_yaw = 55
i_yaw = 60
d_yaw = 17
p_alt = 50
i_alt = 0
d_alt = 0
p_level = 90
i_level = 10
d_level = 100
p_vel = 120
i_vel = 45
d_vel = 1

[Hands0n]

Thanks for the full set of config Andy - that will come in very useful to compare with what I flew with. It wasn';t stock PID, I';d already tweaked basis hand-held tests of the thing.   

A great idea setting the servo to off when unarmed - it was annoying the hell out of me!  So much of the settings are undocumented, although it does seem to be improving slowly.  Just a massive lag between software/hardware and supporting essential documents.  I';m finding info in the Baseflight code itself, not that I want to be reverse engineering code simply because I can!!! LOL

Thats a great idea with the nylon screws for the locking posts.  It would hold the arm more securely against any likely drift - although the other lad suggests the grip at the hinge point is secure enough.  Both methods are worth checking out. I had the nylon screws shear on my 3D printed MHQ the other day when it had a bad landing ';cos of motor issues.  So if 3D printed ABS will shear a nylon screw I';m sure that carbon fibre will have no problem.

See what I just did then? Talked myself out of using screws and back in again

What motors are you running? I';ve got EMax 930KV which with the 10in props seems to work well, it lifts easily and hovers at bang on 50% throttle.  I think with 1200KV you';d probably not have had any issues with the 8in props.

Now I';m really looking forward to seeing what yours does.  Its such a cool design and with your work its going to look an absolute beast.   Every time I look across at mine one word comes to mind ... ."Klingon" ::)

[Udinbak]

Very nice setup andy, it';s given me some ideas for when I get round to fpv. I';ll be grabbing a little spycam to put on mine in the meantime though, although after sunday';s crash I';m thinking of fitting it away from the front (which was where I';d originally intended to mount a camera). It would be nothing but plastic shards and splayed circuits

On props: I was using 8045s with 1000kv motors and felt it was a little underpowered when pulling out of a drop, so I figured I';d go for 9047s (because they';re cheap and red). I can';t give a review until I get them, fix the arm and fly it, but handson seems spot on re: kv/prop size.

And I think I';ve figured out why mine flew off and dived into the concrete: It needed blue tail lights!
Problem Solved


[attach=1][attach=2][attach=3][attach=4]

[guest325]

Probably the best setup would be what I originally intended to use - Sunnysky 2212 1400kv on 8x4.5 props; now thinking about going around 2000kv on 6x4.5';s.
I think that with a smaller rotating disc it should handle breezes better and possibly be more agile, I';ll pass judgement when I get the build done (will be a while yet though - got 3 planes to build and maiden).

[andythilo]

Went to the field today to test out the Tri properly and I was VERY impressed. She flew really well, really stable, fast and cornered well. This was still on 8x4.5';s so on 9x5';s she should be even quicker. Without futher ado, a couple of vids, one 3rd person view, one FPV footage.

http://youtu.be/J3Jx9B-L2vY

http://youtu.be/Fs_ud9TPq2g

[Hands0n]

Andy that looks rather well sorted out even with the 8in props.    Fabulous.

[guest325]

Found this and thought it might be of interest:
http://youtu.be/I5N2FozjJz0

[Hands0n]

Nice find Darrell, thanks for sharing

Sent from my Nexus 5 using Tapatalk

[guest325]

Nice find Darrell, thanks for sharing

Sent from my Nexus 5 using Tapatalk

That';s what I thought, if it can be set up on a kk2 and fly like that it looks like a good ';un!

[Hands0n]

I';m convinced its a good ';un, that frame is so rigid when assembled. Interesting his comments about the servo spline. I';ll use the broken servo, heat up the metal spline to help form the spline in the nylon. Shouldn';t have to do this!

Sent from my Nexus 5 using Tapatalk

[guest325]

There are better ways of connecting the servo to the yaw block - WIP, playing about with a few ideas!
Have a look at Waynes conduit tricopter, I';ve subscribed I like his ideas!

[Trev2971]

nice work Andy thats lovely ...

really need to get mine build but got lots to play with at the mo ... got a blackout to build and then the tri maybe lol

[Udinbak]

I came across these earlier, I figured it would help provide numbers for more scientific prop size evaluation 

http://personal.osi.hu/fuzesisz/strc_eng/
http://adamone.rchomepage.com/calc_thrust.htm

[guest325]

I came across these earlier, I figured it would help provide numbers for more scientific prop size evaluation 

Nice find, thanks for sharing!

[Udinbak]

Just finished fixing the arms.

9"x4.7" props installed. Test hopped and hovered inside. Just waiting for a Crius All In One Pro v2 FC to arrive and I can fly it with relative confidence outside again. I';m not sure exactly what caused it to lose control, but since transmitter interference is a possibility, I';m hoping that the failsafe feature will solve it, along with a lipo in the Tx.

In case the info is useful for anyone else';s repairs:
Used JBWeld epoxy, wrapped the arms in card (from a 12 pack of cadbury chocolate rolls ::) ) one layer thick (would have preferred fabric but forgot to be bothered enough) and used matchsticks split into thin strips (encased in JBWeld) to fill the gap left by the missing arm fragment. Used the unbroken arms as a splint, with card spacer to maintain the gap between the arms caused by the card wrapped around the repair. tied the small ends of the arms with thread, and the large ends screwed together and tied with thread. Filled the screw holes with wax to prevent JBWeld from filling the holes (used existing holes as a mold, filled with melted wax, popped out the wax plug with a hex screwdriver tip). Opened up the screw holes while the JBWeld was still pliable. used standard epoxy in the somewhat threaded motor mounting screw holes to hold the screws in place. Peeled off the smooth, un-JBWelded side of card from around each joint once it had set (I left it 3 days to cure fully), leaving the card half it';s original thickness around the join. Will paint it black at some point



[attach=1][attach=2][attach=3][attach=4]

[update: I just took it out for a quick test flight with the kk fc, there does appear to be yaw issues, drifts to either side resulting in uncontrolled spin and inconsistent giro response. Looks like I won';t be flying until the crius AOIP arrives The LEDs are doing their job though, seems easy to see orientation in the dark, even as it pirouettes into the ground, which turned out to be a nice little test for the repair on the arms, they';ve survived a few minor crashes happily with the arms folding on impact ]

[update 2: Borrowed a kk2.0 in order to a) get my tricopter flying and b) figure out how the hell to get the kk2.0 to fly a tricopter :P After a LOT of trial and error, another servo repair, inexplicable motor failure followed by inexplicable motor unfailure (a day after I ordered a replacement), and blind random chance, I stumbled across the correct settings and it';s flying again ::) now if only I hadn';t drained my batteries figuring everything out... maiden-esque flight vid (also, crash vid) to follow...]
[attach=5]

[Udinbak]

Here';s the test flight:

http://youtu.be/4zBfVb3txv8

Also: The repair on the arms broke in transit, so I flew it anyway. I';ll have to reinforce it at some point.

[Udinbak]

And here';s my friend';s flight vid


http://youtu.be/gJBFL1XYlkk

[JoeB]

Why is everyone using 30A ESCs with these? Looking at the motor stats, on 3S with 10x4.7 props, max amps is < 15A so a 20A ESC would work? or not?

Has anyone measured their motor amps?


http://www.banggood.com/Wholesale-SunnySky-X2212-KV980-Brushless-Motor-p-60600.html

[Hands0n]

Why is everyone using 30A ESCs with these?

In my case, it is what I had laying around in my box ';o bits.

[JoeB]

Thanks. If i use 1250kV then i would need a 30A esc. Are these suitable motors? What is evereyone else using? Planning on using 9x5 props

http://www.banggood.com/Sunnysky-X2212-KV1250-Brushless-Motor-p-908721.html

[barneyg]

I';ve seen reports from people with the appropriate equipment that SunnySky have a tendency to Current Draw test their motors with quite flexible props.  This results in the current draw being quite low, if you use stiffer props even just basic plastic gemfans which are apparently stiffer than the test ones then your current draw will be higher than those listed.

This however is hearsay so take it with the pinch of salt it deserves.

[JoeB]

I have a watt meter and a spare 40A esc so i';ll check the motor/prop current pull before ordering the escs.

So, 980 or 1250 kv motors?

[Hands0n]

So, 980 or 1250 kv motors?

For this model? I';d go for 980KV on the basis that I am using 930KV Emax which are a fair equivalent of the SunnySky 980. I used 10x4.5 props that came with the motors. They';re a bit too much and it could easily do with 8x4.5 and still be "brisk"

[guest325]

I';m seriously thinking of building mine with 2204 2200kv with 6*4 or 6*4.5 props, I think that sort of combination could work very well;  should improve its stability in breezes as well.

[JoeB]

Dammit ive just bought the 980s :P

Let us know your results!

[Trev2971]

Danny did you ever get yours sorted?

[Hands0n]

Danny did you ever get yours sorted?

It sits here Trev, waiting for me to pay some attention to it.  I got stuck into some other projects. But I do intend to get back to this, espesh as I';ve a full new frame here to be scavenged for the broken part     Think its going to be my Spring project.

[Trev2971]

Been thinking about getting mine sorted as the weather has been pants ... looks a bit of a pita to set up reading around 

[Tim1981]

I have been slowly building this Tri Copter since December and finally finished it. Set up for fpv on 4S . Its running on a naze 32 acro. I could not test fly it properly today as the winds were 20mph , thanks a lot to Danny for the advice ! Here is a small line of sight video I made today and some pictures:


<iframe src="//player.vimeo.com/video/118401658" width="500" height="281" frameborder="0" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe> <p>HJ-Y3 Tri Copter Maiden from proquads.co.uk on Vimeo.</p>