Separate names with a comma.
Discussion in 'Cinestar 8' started by Andy Johnson-Laird, Oct 7, 2012.
...is that some new age eastern saying????
I think I'm going to have a shoulder patch made: "Advolet, ergo vibratque." (It sounds better in Latin.)
Adam- That seems bizarre that it's consistently the boom at the 6 o'clock.
Tabb- I'm not too concerned about it since the vibration doesn't seem to be making it past the isolators, just thought I'd let Andy know he's not alone.
Andy- Does this work?
Hi Nick: Yes, the video link works. I'm having a hard time seeing the vibration though -- there's some video artifacting on playback and I think it's masking the problem.... what's the time code where you see the vibration please?
Looks like 0:04 right before the wheel goes off the screen.
Hmm....I certainly see the image go fuzzy at 0:04 -- wow, that really is vibration. Looks like a natural resonance of something, doesn't it.
Just for grins, I'd wedge a small lipo in the landing gear above the wheel (and cable tie it in), just to alter the mass and see if the vibration still happens -- the mass of the lipo should change the natural frequency of the landing gear -- which looks like it's acting like a tuning fork.
Today I flew the Kit C8 frameset that provided the smooth FPV video in posting #27 of this thread (the posting numbers are just to the left of the Reply buttons -- hover over the Reply button to turn them from light grey (aka "invisible") to dark grey (aka "visible").
I attached the 3-Axis gimbal, but didn't have it powered up. It had a Sony CX760 on it, along for the ride -- I just wanted to see what having the payload did for the "ride quality" of the frameset.
I used two FPV cameras. One looking straight ahead from Boom #1 and slightly down, and another one looking straight down. (Shooters keep asking me, "Can you hover over X" and I cannot see what I'm overhead with only one FPV camera. The video switch is on an RC channel so I can flip back and forth easily.)
As you can see, with the gimbal/camera as payload, there are resonant vibrations. To my eye, they are at the natural frequency of the gimbal suspended by the O-rings in the vibration isolators -- you can see more clearly when the movie gets to the "pointing down" camera and you can see one of the landing gear.
So now the question becomes: How can I get rid of that resonance?
I'll be shooting with the gimbal powered up tomorrow and post some footage of that just to see if the Radians, when active will null out that resonance.
Also attached is the GPX file from this flight -- the video is taken from the middle part of the flight. I still have the mystery as to why motor #8 is pulling such a low current. After my shoot tomorrow, I might swap out motor #8 just to see if changing the motor makes a difference. The BL Temperature seems nominal.
Flight Time: 2012-10-17T22:45:39Z
Flight Duration: 4 minutes 21 seconds
Average Current (Amps):
5.7 8.0 6.0 10.5 7.8 8.9 6.8 3.5 0.0 0.0 0.0 0.0 Total=57.0
cw=26.2 ccw=30.8 cw/ccw=0.9
Copter has a counter clockwise moment.
Average Temperature (Celsius):
43.5 44.0 46.8 68.0 80.3 92.8 56.7 51.4 0.0 0.0 0.0 0.0 (edit: debugged)
Maximum Temperature (Celsius):
48.0 54.0 56.0 84.0 86.0 102.0 65.0 57.0 0.0 0.0 0.0 0.0
It's probably what you said before. It probably measures the current incorrectly. It's same as ESC 5 on mine.
It looks like ESC 6 got too hot. Which is odd since ESC 4 reported the highest current average. It's likely that the accuracy of the current measurement is rather low.
I can see you've tweaked your program, William.
I'm curious about the "counter clockwise moment" -- how are you calculating that -- just based on the disparity between CW and CCW current?
I'm also starting to wonder about the accuracy of the current sensors on the ESC's -- I wonder if there is a way of testing them with the props off and measuring the actual current draw of a motor by interposing a test harness between the pig-tail wires on the ESC and the motor leads?
That's always the perils of digital readouts -- you see 9.3 and it's easy to infer that means the sensor is accurate to 0.1 Amps....
After looking at a lot of different peoples GPX files that the average currents tended to have a saw pattern (eg. high then low then high, low, high, etc...) Here are some examples from files I got off of forums:
7.3 5.5 10.6 5.5 8.0 5.0
10.0 9.8 10.3 9.3 8.9 7.6 8.8 8.0 6.2 4.3 6.8 6.7
7.0 5.9 8.3 8.4 7.6 6.1
4.5 2.6 4.2 1.5 1.5 1.5 3.6 2.6
5.7 8.0 6.0 10.5 7.8 8.9 6.8 3.5
See they all have that pattern. I finally realized what that is showing is that all the motors running in one direction pull more amps than all the motors running in another direction. So why would that be? Lets assume that it's not just coincidence. In which case the motors running at higher current are pulling more current because they are running at a higher speed. And the higher speed is set by the flight controller on purpose. Why? Well what would running all the clockwise motors faster then all the counter clockwise motors do? It would make the cinestar turn counter clockwise (I think. It's been a while since I took dynamics. Might be the other direction.). But the cinestar isn't usually turning especially when your just hovering it as I was. So if it's not turning there must be opposite moment canceling it out. So what I think is happening is that the cinestar naturally wants to rotate one way and flight controller is pushing back to keep it from rotating by raising speed of all the motors in one direction and lowering the speed of all the motors in the other direction. So by comparing the clockwise currents with the counter clockwise currents we can deduce which direction the cinestar wants to turn.
So in your case the counter clockwise spinning motors are drawing more current (and thus turning faster) then the counter clockwise motors. This would impart a clockwise moment on your cinestar which is there to counter act the natural tendency of your cinestar to rotate counter clockwise.
At least that's my guess. If I'm right and you zero out the gyro yaw control parameters, your cinestar will rotate counter clockwise. It probably wouldn't be a good idea to do that though.
See if clamping one of these over the boom works:
Mine all come out about the same. Check to make sure all your props are level. If all are level then your copter shouldn't try and rotate at all.
Interesting analysis, William. I can't fault your syllogism. I agree with the premise that having all CW motors running faster would impart a CCW moment too (I could see why it would impart a CW moment so it would have to be CCW).
One thing to bear in mind is that we're sampling at one sample per second, so there's plenty of time for things to change in between -- although, that said, I can't get my head around why the sample would be biased towards what we're seeing...you'd think that it would average out over numerous folks GPX files.
I've got one of those autoranging clamp meter -- I'll give it whirl (metaphorically speaking).
Right. Ideally cw/ccw should equal 1.0 in a hover. In reality there will always be error in the measurements so good luck ever getting a 1.0, but you should be somewhat close. If your getting 1.3 you should try adjusting your motors. It's not just saving some power, but more importantly it's keeping you from overheating. If your cinestar is pulling 64 amps the current averages ideally would look like this:
8.0 8.0 8.0 8.0 8.0 8.0 8.0 8.0
But if your motors are not spot on it could look like this:
9.0 7.0 9.0 7.0 9.0 7.0 9.0 7.0
All those 9.0s are going to overheat your ESCs and your flight is over whereas you might have gotten away with 8.0.
Hmmm.... Clamping over boom might not work. Carbon fiber is conductive. I'm not sure what that would do to the flux lines. Maybe It will have no effect, but try clamping over just the wires just in case.
I wondered about that, but don't those clamp on meters work on the magnetic field induced by the current in the wires? I don't know whether carbon fiber attenuates magnetic fields -- certainly it attenuates RF. The real problem is that I don't have enough slack wire around which to do the clamping (I braided my wires in the boom so that has the side effect of shortening them too).
I had a landing that transferred into a "tip over" this weekend, and props 2 and 3 experienced what you might call a SHOCK STOP. Essentially the ground stopped the props but not after they buried themselves 3-4" into the sod at the golf course green. Is this a "shock stop"?
What can you tell me about a shock stop for a brushless QC CS8 motor? Does it bend the shaft or bearing damage? I'm getting noise/vibration that I can hear just after poweroff from a landing now on prop 3. I'm debating that it is either dirt induced from dirt flying up, but i'm nervous it may be more severe and now suffering from a situation like your close-encounter-of-the-arboreal-kind.
I'm hoping air/wd40 may do the trick, but I'd like to follow your insight incase it falls short for the fix. I hope this isn't the case as I've seen many others have tip overs....
PS, guidance on applying wd40 properly to a qc3328 motor?
Good question....I suspect that if you got the props that far into the ground it might have been quite kind to the bearings....not so much a shock as a rude awakening....
But the ultimate metric is whether the motors sound sweet when you test them individually with MK Tool.
You can check the bearing for run-out (aka wobble) by looking at the blade tips if you have a brand new prop on there -- verify that the prop is true by placing it on a flat surface before you put it on the motor, then put it on the motor and hand-sping the motor and measure the height of the prop blades with the prop rotated at 0, 90, 180 and you'll be able to tell whether the prop shaft is leaning over. If it is, you can replace just that piece. Hmmm. Went looking for the QC-3328 propeller mount on QC's web site, but I cannot find it. I know they sell them as I've ordered them in the past.
I'm just working on a posting that will deal with which bearings to order and how to replace them.... I'll post the thread in a few minutes.
I would not use WD-40. While it's a lubricant, I suspect it's not the right one for the job.
I suspect this might be more suitable: http://www.bocabearings.com/bearing-inventory/13931/highspeedoil-lbt-lightning-lube given the temperature range it can handle.
Hope this helps
Have you checked that all the screws in the frame are tight? I had some vibes in a boom and turned out to be some loose screws in the frame..
Is there any reason to remove the zip ties and put tape on instead? I've just gone through balancing mine and I feel like the ties would stand a better chance of staying in place. I could be wrong though.
As a note, 5 of my motors were very easy to dial in. The other three too a lot more work and didn't get quite as nice, but had a very big improvement.
Just FYI my motors are AXi 4120/20's with FreeFly motor controllers.