I recently built a Sky Hero Spy 750 Y6. It is using Y6B parameters (AC 3.2) which after having it in the air a number of times is obviously not correct. The motors are KDE 2814XF-515. It seems there is a "RC FEEL" slider, does that affect the PID settings? Right now it feels like driving a Greyhound bus, and I only land in Loiter because it lumbers too much at the slightest hint of wind gusts.
My first flights showed the props were too small. In general, if the copter stability decreases greatly when wind increases even slightly, what should be changed to help alleviate wildly rolling back and forth? I am not ready to do Auto Tune yet and to be honest am a bit nervous considering the size/weight difference between this and my Y6B, so would like to get it a bit better before attempting Auto Tune. Would increasing Stabilize Roll/Pitch help?
Current settings:
Replies
DG,
I find it interesting that both my Y6B and OctoQuad use a smaller prop on top and a larger on the lower when you have you're configured in the opposite but my PWM readings support this configuration. I also noted that Markus' article indicates that you should use a larger on top and smaller on the bottom with a higher degree of pitch.
I'm trying to understand the dynamics of the difference between your setup and mine. With weather improving, I plan to recalibrate my birds and get them back into the air. Like to share RCOUT with you and see if there is something I'm missing. I'm also planning to swap out props to see the PWM differences. :-)
I double checked my post as I have a habit of wixing my mords up. The smaller props are on top and larger bottom. The pwm is very closely matched, but flight performance is marginal.
The Tarot props are not going to be used anymore once this is figured out now that the weather is improving. I've concluded there is no way these props are ever going to work, and are 75% of the problems with this Y6.
Today I used the Gemfan 14x5.5 top and bottom. It introduces yaw on takeoff but straightens out once in the air. Interestingly current draw dropped with this configuration. Those props, although cheap, are still stiffer than the Tarots. There is a high pwm gap, which is why I'm about ready to try T-Motor BW 14x6 props for the bottom and see what happens.
All props are well balanced. The Gemfans were the worst; the hubs on a few were so bad I had to remove material from one side and add to the other. No amount of sanding could balance them horizontally so I resorted to tape on the leading edge. I'm done wasting money on low quality props.
The vibration levels are very good during a quick hover test today:
Markus suggested swapping the props (put Tarot 15x5.5 on top and Gemfan 14x5.5 bottom), which was done and the results were predictably bad, both in pwm mismatch and flight characteristics. The theory may have merit, but IMO there are so many variables involved with a Y6 it isn't possible for a one-size-fits-all solution.
Rob L said he would use higher pitch on the bottom but never small diameter. Go figure.
The bottom line is, for TM style props the choices are slim to attain the desired effect of spinning the bottom props slightly faster, but also having smaller diameter and higher pitch.
Quality props makes a world of difference. I resorted to going back to APC props and have had good results thus far. As for the mix up on which sizes you have... Ouch! I hope people didn't go with the choice of smaller on bottom and larger on top or they will have some issues but the data doesn't lye when comparing PWM readings in RCOUT for each arm. I also think it's very important to check vibration levels especially Z-axis as a badly balance propeller can cause a vertical flight acceleration/instability if it's severely off balance.
There are people using smaller props on the bottom without compensating with higher pitch. I tried and got poor results all around. Then again, those Tarots are just really bad props in higher thrust configurations. Virtually every test I've done with them is meaningless. They caused the Z axis vibration. One person suggested the arms were bending.....uh, well, not with 30mm tubes, and these Sky Hero frames are very strong. It was harmonics from the props flexing.
With the same props on top/bottom there's a divergence, but depending on all the other variables they may or may not work.
I found a good deal for unused Xoar PJP-N (matched set) 15x5 props and also 3 PJN (single prop version of matched set) 14x7 CCW. Now I can start to test the ECalc hypothesis, plus increase the axial distance between the props with the use of the prop adapters. The bottom props should spin slower and may still be too fast, but it's a good starting point. Xoar offers about every size/pitch one could ask for and the prices are reasonable.
I will find the settings Leonard recommended to at least get it in the air to Auto Tune.
Prop choice is a HUGE factor. My advice right now is you get what you pay for.
Before discussing AT, what motors/props do you have and what is the AUW of the copter?
I have fairly large (54mm diameter) rctimer x6 motors. Currently have 15x5.2 CF props mounted, but also have a set of 16x5.5 APC nylon ones.
Weight is 3.6kg without batteries, 5.0kg with (6Sx5000 x 2)
I haven't installed gimbal or FPV gear yet, that will add a bit.
Wow, your 750 is a bomber!
If you have a log, look at the RCOU and compare top to bottom motors (1 & 2, 3 & 4 etc.) pwm. If the spread is too high the copter becomes unstable. I had to put smaller props on the top to bring the pwm closer and make the copter flyable. Higher pitch is desirable on the bottom if there's a choice.
When I use Tarot 15x5.5 top and bottom, any wind that comes along causes a lot of oscillation, and so moderately fast descents. Look at the Z axis......the copter acted like it wanted to lose control. The second group of spikes were really bad:
On a smaller Y6 frame the effect of the flimsly props isn't felt as much, but once you start getting into the 400-600+ thrust range, having good props is a must. You didn't mention what brand the CF props are, and I assume they are not TM style based on the pitch of 5.2.
I think another thing that negatively influences AT on my Spy is the rubber battery tray isolators. Note how they move back and forth easily. As AT pitches and rolls quickly, I think the accelerometers pick that up, and yours having 2x stacked may make it worse.
Anyway, here's the thread where I posted about my 750 and Rob/Leonard gave some tips on getting it off the ground. Honestly, I think if I had the correct prop combination off the start, it would have been easier, but in any case the PID's do need adjusting before performing AT. The flight logs tell the story.
http://diydrones.com/forum/topics/after-doing-the-autotune-the-dron...
I'd be interested in your results, so please post as you go. As a side note, do your motors get hot? Where did you mount the ESC's?
BTW, Rob L said AC3.3 will be awesome for tuning problems.
EDIT: I've been conversing with Markus at ECalc concerning an article they wrote about coaxials. It's interesting, but finding the right prop combo is the problem. I see they updated the article with the subject of "Ring Vortex State". Lots of variables going on with a Y6, but sorted out they are a great bird to fly. http://ecalc.ch/calcinclude/help/xcoptercalchelp.htm#hints
Yeah, it's big..
The props on it now are RCT foldable, which got lots of attention in the balancing and polishing department. One of the reasons I chose this frame was because of how the front arms folded back, I'm hoping to make the foldable props work to help with transportability. Vibrations are good.
The bottom PWM values are slightly lower than the top, but that was observed from its only flight to- date, hovering 1-2 M for a few minutes, at which point I decided the PIDs needed some attention before continuing.
ESCs are in the arms. I'm fully prepared to need to move them if heat becomes an issue; we'll see..
There isn't as much room in this thing as I had hoped.
I'm experimenting with a 3d-printed anti-vibration mount for the pixhawk on this build.
(photo attached)
The mount itself is printed in flexible TPE (ninjaflex), and the square in the center is ABS, which adheres to TPE. pixhawk is attached to the ABS section.
DG - I was curious where you ended up with PIDS on this frame? I'm in the same boat - Just finishing up a 750 Y6, and I started with the 3DR PIDs, which, like you said, are clearly off.
This is the largest frame I've built, and my first Y. My usual tuning procedures don't work on something this large, and I don't have have confidence in letting auto-tune do its thing based on how unstable it is currently.