I have a new build here, 500mm frame with landing gear and 9x4 bi-blades, 4S2P 8200 mah LiIon, about 1.3kg AUW. I haven’t adjusted any tuning, just the basic parameter set for 9" airscrew.
Could anyone have a look to see if it looks safe to proceed with Autotuning, or if it would be recommended to go straight to notch filter log and setup?
I did not see any visible twitching/oscillating and seemed to respond decently to my control twitches for pitch and roll but this is much bigger than my last build so I’m wanting to be extra cautious.
VIBE Log does show quite a bit of vibrations (I think, unsure what to expect for this build) in X Y and Z.
The vibration levels are way too high. What Frame is it? This must be addressed 1st, the Notch Filter is not a fix for high mechanical vibration. You have some (not terrible) output oscillation also. Drop the Rate Pitch/Roll PID’s and set these:
INS_ACCEL_FILTER,10
INS_LOG_BAT_MASK,1
INS_LOG_BAT_OPT,4
PSC_ACCZ_I,0.36
PSC_ACCZ_P,0.18
That should tame the output oscillation and prepare for configuring the notch filter. But, the Vibe’s need to be reduced. Props balanced? How is the FC mounted?
FC is mounted on double stack of soft grommets over 3mm screws in the center. It does come with a plate for floating grommet-only mount but the plate will not take a 30x30 so I would have to make an entirely new one.
Props seem balanced on a quick hang on wire test but I do not have a proper balancer.
What would you suggest for the Rate Pitch/Roll PIDs?
I will also set those others and re-check the props. They are using a sizing bushing so I am a bit worried about that.
That has worked well for me but on a very rigid frame. There is a lot to shake around on that frame so your props will have to be well balanced. Invest in a balancer, doesn’t have to be fancy. I have used the same cheap mag balancer for years.
Leave the D-term where it’s at and drop P&I to .10 D may have to come down but if the Vibe’s and oscillation are reduced Auto Tune will take care of that.
To answer your 1st question, yes the notch filter should be configured before Auto Tune.
Those long skid struts aren’t helping. I use a similar frame (but not identical, and possibly more rigid) and have considered shortening them or just using some short vertical pieces at each corner.
You might consider removing the underslung battery contraption and just strapping the battery on top, closer to the prop plane. Contrary to popular belief, low slung weight doesn’t do much to stabilize a mulitrotor and can contribute to oscillation.
I’m liking the struts but if I move battery to the top I could shorten a bit. Because of how the arms mount I don’t really want to mess with something from the corners (and well I went with this frame specifically so I didn’t have to mess with custom landing gear on my other one lol)
Yeah I am not a fan of the battery mount in general, but having the top free gives me more room for mounting (I do not have a GPS pedestal so at this point it would be partially blocked by battery). If I can come up with a good way to situate and mount on top I will though, especially since that stupid battery mount just clips on and can also slide. Plus it would give me more room for undermounting other items I wanted. I was hoping to eventually vacu-form a canopy for more protection and light weatherproofing however, and battery on top would rule that out.
Looking at it, I may be able to remove the battery mount and attach directly to the rails below, maybe that would help a bit.
The brackets/clamps that capture each arm on the corners of the frame seem spaced very closely together, making for potentially flexible mounts. If you have access to a 3D printer, consider making some more rigid mounts for the arms. That may solve many of your vibration issues.
Carbon fiber drills reasonably well with a sharp bit and some backing. You could also just space the inner clamps a little farther inboard (carefully, as not to change any critical angles). And, for that matter, maybe create the 30x30 stack mount that you seem to be lacking. Use a little water as coolant and to keep the dust at bay (don’t breathe composite dust!).
I don’t know for sure, but your frame looks like a clone of sorts of the Holybro X500. I’ve used that frame as well as the Holybro S500 a decent amount, and the legs do “wobble” a decent amount unfortunately. A little while back @Allister showed me that HolyBro had a “leg brace” 3D print on their site - on this page X500 V2 Kits – Holybro very near the bottom. If you can access 3D print, the brace helps significantly in terms of rigidity. I haven’t done any back to back comparisons with tuning and such, but I’m sure it would help at least a bit there as well. I’m not sure how long the legs are on your frame. The X500 is much taller than the S500, so when I was flying the S500, I shortened the leg brace and it works well since all the other geometry is the same. Between the link and the attached files, I hope you find something useful. The braces just attach with a zip tie through the clip and around the brace portion, though I’d imagine you could glue them as well since they do slide around a little bit on the legs. S500 Landing Gear Leg Brace.stl (268.9 KB) X500orS500 Leg Brace Clip.stl (296.3 KB)
And to add to what @dpapworth said there is a TPU version of the leg brace. I’ve printed these in flexible PLA and so far so good. Obviously not as rigid as the regular PLA versions but still keeps the landing gear from flopping around. My hope is that in a crash the flexible version will give enough that any damage isn’t transferred to the main body and kept in the landing gear (or at least parts that I can 3D print like the brackets). Thankfully my students haven’t tested this theory for me … yet.
They are, that was my very first concern. From the image I thought I could just drill some extra mounting holes to space it back further, but there is a big cutout just behind and goes almost all the way to the 30x30 mounting. Honestly the mounts themselves are quite ridgid, however the top/bottom plates are thin and flimsy CF and that is where it seems to flex. Same goes for the landing gear, it bolts to just the bottom plate with no other supports. For that I’m thinking of cutting a bit bigger piece of CF sheet to double up and spread the forces. I was also thinking of adding some standoffs between top and bottom plate around gear mounts to give it a bit more.
I swear if I had easy (read: cheap) access to a CNC I would just cut new top and bottom plates.
Oh that is great! I will check those out. Unfortunately I do not have access to a 3D Printer really (rather getting my buddy to do it takes forever lol). I will keep those in mind though!
As an experiment I changed props to 8x5x2 HQ Props. I also noticed when removing the 9x4 Master Airscrew that the reducer bushing was not in fact snug, I had maybe 0.3-0.5mm of movement on the shaft so I don’t think I can reliably center enough to even balance (or ensure they are balanced/center on the shaft).
I did run another test flight. I re-ran the Initial Parameters to set for 8", I did not adjust any of the PIDs or Rates just so I had a fresh baseline, but I did however enable the INS_ACCEL and INS_LOG you recommended.
If you could have a look that would be great. It looks like the X/Y/Z VIBE are significantly lower and hopefully make it a bit more workable. I’m not seeing as specific peaks in the FFT as my other build but maybe still masked by vibrations?
The Master Airscrew props need their own prop adapter rings, they are a slightly different size to the common props available Adapter Rings Multifunctional x4?
Hey I am in fact using the adapters that came with the props. The next size smaller won’t even fit over the threads with decent pressure but maybe I will have to try to brute force it?