Update June 2016 Edit
http://diydrones.com/forum/topics/vibration-finally-solved-the-frantz-strut?xg_source=activity A multirotor weighing more than 300 grams can only respond with significant thrust deltas in the 20 or less Hertz range (see other research I produced testing actual lift response rates on various size ships). Thus vibration noise is typically greater than 5 to 20 Hz, depending on the ship and the FCU is effectively responding to attitude changes occurring at less than 20 Hz. by http://diydrones.com/profile/ForrestFrantz
http://diydrones.com/forum/topics/vibration-isolation-and-dampening-of-apm-px4-for-version-2-9?xg_source=activity&id=705844%3ATopic%3A1102762&page=57#comments by http://diydrones.com/profile/ForrestFrantz
solution: Wirerope and DJI aluminum block Edit
Wire rope isolator Openpilot
- http://forums.openpilot.org/topic/6888-flexacopter-the-making-of/ Mount the Gimbal camera on a small square plate which in turn is mounted in wire-rope. Silicon tape on the platform won't necessarily create a mechanical low pass filter, the idea is not to eliminate all vibration, but to create a mechanical low pass filter.
- http://diydrones.com/forum/topics/vibration-finally-solved-the-frantz-strut Frantz strut
- Combining Quad frames flexacopter wirerope isolation makes for a low pass filter. High hz vibrations saturates the IMU and causes waves on the camera image. Wirerope is the most effective solution.
- http://www.infinity-hobby.com/main/product_info.php?products_id=10670 Circular wirerope encased in red plastic version.
- https://www.youtube.com/watch?v=cajoxGhFQck Wire rope vibration/rubber band combination, same system as used by flexacopter.
- DJI Wookong IMU mounted on small aluminium block idea. (open the imu unit and upload a picture here, plus dimensions, anybody can edit this page.) https://www.youtube.com/watch?v=MapiklR_9qg describes 3dr solo "swaying" motion compared to DJI, most likely the stability and higher inertia due to the aluminum block. An alternetive is to use Lathe#Epoxy_granite, instead of aluminum, its vibration absorption properties is 8x that of cast iron.
Posts on vibration Edit
Combing through the endless forum posts on preventing IMU vibration saturation provides the following:
https://groups.google.com/forum/#!topic/drones-discuss/fnWYM48pGys Vibration high hz noise causes crashes.
1 Copy the Quad frames flexacopter wire-rope and DJI Wookong IMU on aluminium block idea. The quad frame as a system should function as a low-pass filter, the idea is not to stop all vibration but to cut-off high frequency vibration preventing IMU saturation and distortion of video. Diydrones are aware of the fact that the wire-rope concept(http://www.flexacopter.com) solves the vibration issue but are reluctant to copy the idea. Experimentation showed that it is a specific thickness of foam under the IMU that cuts of high hz vibration, packs of vibration absorbing gells will not turn the quad frame into a low pass filter. Different types of material gell, foam or metal etc, will depending on frame enable the quad to function as a low pass filter. The resonant frequency of the IMU orientation sensor is in the higher frequency ranges, which results in gyro saturation if the quad frame acts as a high-pass filter due to resonant oscillation(like a bridge collapsing at its resonant frequency). Invensense will not disclose what this resonant frequency is without a NDA agreement, resulting in multiple users missing the frequency by luck and then thinking that their particular rubber,wood,gel patch-up worked the magic and then perplexed that the same idea won't work on a different frame.
Van Niekerk idea, (wirerope inside a plane will work better, this post by Van Niekerk is another example of not understanding the underlying engineering, but just so by luck happening to create a mechanical low pass filter.) Nitro or fuel palnes is to susspend the APM on the CG in the air. so for example use the ArduPilotMega lazer cut box and dril a hole on each side of the mounting plant then span a rubber band from each side of the box to the closest serface in the plane then if you tied it very tight and the APM board it level in the plane take tissue paper and support the APM box all around it in the airframe. it looks weird but it works very well - http://diydrones.com/profile/JohannVanNiekerk , http://diydrones.com/profiles/blogs/vibration-damper-for-gasnitro, http://diydrones.com/profiles/blogs/vibration-damper-for-gasnitro Chris Anderson and his developers knows exactly what the solution is , but it will involve copying Flexacopter, DJI Wookong and other people's ideas who would probably demand royalties, kick up a fuss etc... Which is frustrating, because we the end-users inoculated against Imaginary Property(HackPatents) , just want to get our stuff to work. Life is too short to reinvent the wheel!
Vibration plate Edit
Vortex flush Edit
http://diydrones.com/profiles/blogs/rubber-suspended-mounting-of-the-apm-board-not-a-good-idea?xg_source=activity&id=705844%3ABlogPost%3A838898&page=3#comments Idea with vibrations is to form a mechanical low pass filter, allow the low frequencies of the frame through and block the high frequencies
Key point that has been missed in the discussion thus far: Much of the vibration present in multi-copters doesn't come from the motors, but in fact comes from the "vortex shedding" of the prop-wash on the arms. So even if you rubber mount the motors, there will still be vibrations on the airframe. (there is a drone firm with a solution to this, they added a slotted circular pattern cnc'd into a plate just under the props.)
Just a little something I picked up while working with vibratory feeders: It's hard to tell how the dampers work before testing the full setup with all the parts and weight in it as they all affect the fundamental frequency of the said system. The problem is that if your motor rpm range happens to hit a natural frequency of the mount plate+motor+propeller "system" you'll end up worse off than with any dampers at all. You'll probably have to test the system, with different thicknesses/types of those rubber(?) grommets, to find the optimum setup. In other words you'll want the damping effect to be adjustable to avoid vibration by experimenting (this principle is basically used in reverse with vibratory feeders via an adjustable spring load to fine tune the fundamental frequency of the system to get what ever is being moved to move as efficiently as possible). As a general rule, in the case of damping the vibration at your motor mounts, you should have the fundamental frequency of the mount as low as possible -- i.e. the mount should be as "loose" as possible -- but still above any frequencies that might be caused by the movement of the frame (which is especially the case with damping the autopilot board). One thing also worth mentioning is the use of multiple stages of dampers that have different fundamental frequencies. That being said, the balancing of the motors/props is probably the easiest way to eliminate the worst of vibrations. Anyway: nice looking motor mounts! Are you planning to 3D print them, or? Edit: of course use of materials with good shock absorbing properties also help
http://diydrones.com/profile/EdKirk520 I use 12mm round booms (Multiwiicopter Scarab) I use a short piece of arrow shaft, about 3 inches long with 2 inches glued into the 12mm boom with aquarium silicone at the motor end. I epoxy a another short piece of 12mm boom to the arrow shaft to attach to the motor mount. Nothing scientific just trial and error, I see improvement especially with a camera mounted. The video quality is much improved and the amount of audible noise picked up by the camera is less.
I started using my "soft mount" on my camera Gimbal and now use it to soft mount my motors. It is somewhat difficult to make, alignment is critical, I made an assembly tool for alignment and the acquarium silicone has to be applied in two stages so that is will set up (like old school contact cement). If it doesn't get enough exposure to the air it takes a week to dry. When done correctly is seems soft and plyable but quite strong, has survived a number of crashes. The arrow shaft is quite strong, better then I expected, the layer of silicone is pretty thin, doesn't seem to take much.
There actually is a lot of prior art here. In the traditional helicopter comment we've been dealing with this for several years now especially when it comes to Flybarless controllers and especially the more advanced ones that have both gyros and accelerometers (for self balancing and unusual attitude recovery). There has been all kinds of crazy experimentation going on and much of it has just been trial and error without any research. Vibration isolation and dampening is an old field with much research behind it. One source that I've found to be an excellent source of basic theory is this website: http://www.farrat.com/vibration_theory.html . One of the things I've found and you'll find in any professional discussion on vibration isolation is proper dampening is essential . Think about it, if you took the shocks off your car and just had springs it would bounce many times upon encountering a bump. The shock absorbers provide a degree of resistance such that the springs don't just oscillate. A while back someone made a post diagraming their idea of using large o-rings to suspend the APM. Creative idea but absolutely would be worse than nothing because the o-rings will oscillate like a guitar string and at some resonant frequency would actually tend to vibrate even more. My experience with Flybarless controllers has shown that one needs to think simple here. Fancy vibration isolation and absorbing mechanism will rapidly become unwieldy and impractical. We can take some lessons from the Flybarless manufacturers. Everyone of them has done a lot of work to get their controllers to work well in a very high vibration environment which also has a broad range of frequencies to deal with (tail rotor, motor, main blades all at different frequencies). They've realized that an isolation system is essentially like a low-pass filter - we want to let low frequencies pass (the movement of the ship) but isolate and filter out the high frequencies from the motors and props. At the same time the APM must not be so loose that it flops around.
So, the best know method has been to use varying densities of foam mounting tape and, in some applications, metal plates to add mass (in order to reduce the resonant frequency). On my FC's ranging from APM1.0, APM2.0, HoverFly, XA, CopterControl etc. I've had excellent results using the foam mounting pads that come with the Skookum SK720 (available here: http://helidirect.com/skookum-anti-vibration-mounting-kit-p-22097.hdx?) and recently. Recently, due to the weight of the APM along with all the wires connecting to it, I've begun to use the pads that come with the MicroBeast which are a bit stiffer (available here: http://helidirect.com/beastx-special-foam-tape-p-22951.hdx). Remember, it's essentially a tuned circuit - weight of the APM vs springiness of the foam and the stiffness/resistence of the foam provides dampenings of oscillations. I use one one of these pads and cut it in half. One half goes on each end of the bottom of the APM (thinking lengthwise here). I've tried the approach of mounting the APM directly to the airframe using the mounting holes and even with rubber vibration mounts on each hole I've not been satisfied with tuning and performance. Mounting use foam mounting pads as I've described above is a proven mechanism and I've found that using that meting I can tune PID's using the "hand" method and, unless I've done something wrong it flies well and doesn't require precise PID tuning. In most cases the default PID values have flow pretty well right out of the box on airframes very different from 3DR's. One more thing, many guys have tried to "strap down" their flybarless controllers for security or thinking that it will "prevent vibration" but that completely defeats the function of the foam tape. Think about if you put steel cable around the springs of your car and "strapped" them down. It's pretty clear it won't work very well. Lastly, you can't just use any pads. Some have no springiness and not enough compression "travel" (i.e. how much it can be compress). That would be most "servo tape", so called "gyro tape", any sort of gel, etc. It needs to be mounting pads specifically designed for mounting flybarless controllers. While it's very likely you can buy similar stuff it all seems to not be quite right so I've stuck with the Skookum and MicroBeast stuff. Feel free to experiment if you have the time. I should also point out that my results on Flybarless stuff was facilitated by the fact that the SK720 has built-in vibration analysis - spectrum analysis which can be monitored on each axis. This proved VERY useful in checking the results of various mounting methods. Would be great to have the APM log the vibrations during flight and then have the Mission Planner chart out the spectrum analysis. This is essentially what Skookum and Mikado do - invaluable. You'd be surprised at how much of your tuning difficulties are related to vibration. I'm getting ready to leave on a multi-week trip but if photos would be helpful let me know and I'll try to post some. Cheers, Ron
From my experience: Urethane for high harmonics (motor rpm vibrations) Rubber for low harmonics (Prop to strut ratio, bent axles) Foam rubber for very low harmonics (not stiff frames) Delrin for engine mounts, very high kv (>1500).
Identify what vibration you have... Then use the appropriate material. If you're a 3DR frame for example, with 100% perfectly balanced motors and props, urethane is the way to go.
Where you want the dampeners placed depends on you frame design and the 'type' of APM mounting plate. Stiff frames and a thick mounting plate like the 3DR kit and dampeners closer to the vibration source work (you typically want to dampen at the source)... A more flexible frame or thinner mounting plates, it's ok to dampen right at the APM mounting holes.
]Monroe, your description of the Cadillac is not entirely accurate. In many cases, the ratio of sprung to unsprung weight is really not any different than any other car. The real reason they ride so smooth is simply that the spring and damping rates are very soft compared to other vehicles. As for the O-ring hammock style mount, they can work, and work well. But it depends on the design. Yes, if the orings are too weak, it can create control oscillation. Get it right, and it's fine. Simple as that. And this is much easier to do on a multi than a helicopter, since the control response on a multi is so much softer than a heli. The reason that the FBL controller manufacturers use foam tape, is not because it is the be-all end-all. It's because it works well enough, it's light and compact, and it's cheap. But that doesn't mean that you can't improve on the performance of the foam tape. Being light, compact, adequate performance and cheap are all good qualities. There's nothing wrong with that. It's a viable option. But, a really good system can allow a FC to perform even in vibration environments that would ground a foam taped mounted FC. It's easy to make the argument "Well, make sure your machine doesn't vibrate that badly in the first place." And that's fine, but it's not always practical. There are plenty of stories where FC's have failed *mid flight* due to some imperceptible bearing problem that occurred while in the air. I've read so many stories of people blindly replacing every bearing in their machine in search of the one bearing causing problems, even though it feels fine by hand. Another great example would be that of a mutli-copter losing blades on a motor mid-air. It happens. And it creates massive vibrations. This will shoot down other FC's mounted on foam tape. But a rubber suspended APM will still be able to get safely back on the ground.
Ron, Your simple analogy of the car spring idea is pretty much how I came to this conclusion about the rubber oring idea people are using would be worse than having nothing at all and hence the start of this blog. As I mentioned , any movement at all of the APM relative to the frame position or attitude will be disastrous. Monroe, On my copter control board because it is so small and light (about 1/8 that of the APM board) I made a sandwich of two types of rubber material and a metal plate which works as the "air anchor" fitted just under the CC board which is pretty much the idea you mentioned as well. . Robert another reason that big cars ride more smoothly is because of the longer wheel base and that the distance between the wheels to where the driver sits is further than in a small car. Car manufacturers design the seating deliberately so the drivers back is sitting exactly on the centre between the wheels so the ride is smoother. I suspect then that multi rotor copters with wider arms spans would also not be as nervous to fly because of the same? Wim. although your idea removes the vibrationr really well, my concern again is the possibility that because the motors as suspended then the motors can move slightly independently relative to the APM during flight. There is a lot of torsion when the motors are required to yaw or do quick maneuvers and this may cause them to move out of line relative the the center plate alignment which again will give the APM false readings and poor flight control. The only way your really going to be able to test this is during flight.
Comment by John Campen on April 20, 2012 at 7:08pm
Monroe I agree on that and it was also one idea I have on my list to try out. As it is I use double sided tap on my batteries to stop them sliding around in the battery box. Ive taken most of the stickiness off them so they don't stick, but enough still there to prevent them moving around. This two sided tap foam material looks like it would act as a vib dampener and with the battery working like the shock absorber because of its weight
Vibration negation Edit
http://diydrones.com/profiles/blogs/vibration-damper-for-gasnitro I had read about issues or concerns with vibrations and gas/nitro powered planes so I come up with a vibration damper assembly made up of 2 fiberglass plates separated by 4 foam ear plugs that I cut the tips off and glued to each piece. On the lower board I also mounted my receiver and the airspeed sensor.
I flew this in a 66” Decathlon power by an O.S. .70 four stroke and had no issues at all with vibration – at least as far as the APM is concerned. I tested the unit in stabilize mode and RTL both worked flawlessly! Late into my second flight while in manual mode I was doing a few touch and goes after the third one I was on the downwind leg and everything quit working the plane ended up rolling onto its back and nosed straight into the ground. The plane is a total loss however after stripping all of the parts and performing a few checks on the APM all of the really cool stuff is still working perfectly. The vibration mount had to help some but it does show the boards are fairly durable
The best way I have seen sofar to get rid of vibrations in a Nitro or fuel palnes is to susspend the APM on the CG in the air.! so for example use the ArduPilotMega lazer cut box and dril a hole on each side of the mounting plant then span a rubber band from each side of the box to the closest serface in the plane then if you tied it very tight and the APM board it level in the plane take tissue paper and support the APM box all around it in the airframe. it looks weird but it works very well - http://diydrones.com/profile/JohannVanNiekerk
Reduce engine vibration via mount Edit
Use the http://uavfactory.com/ Aerosonde type frame which allows the motor to be mounted at the back onto a vibration absorbing material.
http://www.diydrones.com/profiles/blogs/gel-as-imu-vibration-dampene this linke was removed by diydrones for some reason.
Silicon foam Edit
For me Silicone Foam works the best for eliminating any vibrations. I have my APM mounted inside of a semi enclosed servo box, the kind higher end digital servos come in, with a 0.25" thick piece of silicone rubber foam between it and the frame. I am using double sided foam tape to adhere the foam to the frame, and the APM box to the foam. Vibrations almost completely eliminated. AC2.025 is working great though. Yaw now resembles the yaw we had in RC2. Alt. hold is working excellently, 1-1.5 meter drift up and down. Loiter still overshooting a little. I have made some PID changes and will be flying again today. Great work!
On the front page there is a jpg link to a commercial quad drone that attaches the four arms of the quad to the drone enclosure with a rubber block.
Vibration absorbing blocks Edit
http://www.barnardmicrosystems.com/L4E_inview_gallery.htm mounts two engines on vibration absorbing blocks that are attached to each wing.
The blocks can be internally mounted on vibration shock absorbers.
Vibration isolation Edit
http://forums.openpilot.org/topic/6888-flexacopter-the-making-of/page__st__300__p__75848#entry75848 http://www.flexacopter.com uses wirerope and battery as absorbing element
PID issues Edit
Diydrones user PID settings, vibration http://diydrones.com/profiles/blogs/arducopter-dev-team-test?id=705844%3ABlogPost%3A427991&page=2#comments
Steal rod in rubber Edit
Grommets openpilot Edit
Vibration Reduction with Grommets and Camera Basket Started by comster, Today, 03:58 AM http://forums.openpilot.org/topic/10972-vibration-reduction-with-grommets-and-camera-basket/
http://www.droneflyers.com/2015/08/3dr-solo-is-a-nogo-for-now/ It is plastic, no wire rope, hence Gimbal problems.