Separate names with a comma.
Discussion in 'Radian' started by Brad Meier, Sep 19, 2012.
The servo slop seems to be the same on all servos.. Really none at all.
If you put your finger on the mid-point of the tilt servo belt and push toward the other part of the belt, how far does it deflect?
On my gimbal, I found that if I steady a small steel rule against the big pulley it allows me to measure the deflection.
I see it as about 2 mm.
I'm just curious as to what the "correct" deflection should be.
Yeah mines as tight as I can get it.. about 2-3mm
That rules slack in the drive belt then. So where on earth can the tilt jitter be coming from?
I'm pretty sure, just from the tests you've done, that you've done your fair share of waggling the gimbal in the tilt axis on the ground and nothing else is loose. I suspect we'll need to defer to Tabb's superior knowledge here.
Sorry I can't suggest anything else to try.
Hey there, I´m currently also experimenting with the gain settings, starting with a NEX5n. I also have a sony cx730 and a 5DIII, but somehow I love the weight of the NEX, so if I can get it working properly I might use it most of the time.
My pan-axis isn´t yet balanced because I don´t have the longer booms. What I found though is, that my pan is way faster and responsive in one direction than in the other, same with tilt. I´m using Spectrum DX6i and Satellite. Before the Spectrum, I used the Graupner MX-16 (without the radians though) and pan was the same speed in both directions. Unfortunately the Graupner MX-16 interfered with the Graupner MX-20 of the cinestar, so I decided to change.
Now I start thinking that there might be a connection with gain settings and the effectivity of the radian with the radio/receiver used. Anybody tried different systems? S-Bus, PPM, Satellite?
What are you guys using?
I used Dx6i/Satellite with the beta sensors and then cut over to S-bus for the production Radian sensors. I've not seen that asymmetrical slew rate in the pan axis that you're experiencing with either system. What do you have your pan gain set to please? Also, is the pan Radian servo still set to the default slew rate of 60?
I'm just trying to think what would cause the asymmetry....
thanks a lot for your reply.
I changed the slew rate to different values from 20 to 60, but the problem remains.
I did a lot of gain testing today from 70/130/150 (pan/roll/tilt) all the way down to 30/65/75 and found a setting of 40/70/80 worked best with very little wind.
But when I went to another place with more wind (approx 10mph), those settings resulted in pretty shaky footage.
So I guess there´s more testing to come. If somebody wants to comment on the footage and possible improvement of the settings, please don´t hesitate ;-)
I need to note though that I´m using the NEX5n with the 16mm 2.8 pancake lens that doesn´t have any stabilizer at all. But since the Canon wideangles don´t have stabilization either I think this shouldn´t be a problem.
However, the asymmetry remains - not only roll, but also tilt - no matter how I change the settings.
Here are two videos of the testflights:
<iframe src="http://player.vimeo.com/video/50880728" width="500" height="375" frameborder="0" webkitAllowFullScreen mozallowfullscreen allowFullScreen></iframe>
<iframe src="http://player.vimeo.com/video/50880727" width="500" height="281" frameborder="0"></iframe>
Edit: how do you guys embed vimeo videos?
Your footage looks very much similar as mine at the moment. I'm beginning to think our biggest problem is vibration isolations in CS frames. Especially CS6 frame has only one dampener on one side which is kicking back and causes jitter when gains are at acceptable level. I already have replaced stiffer rubber o-rings which helped a little but not eliminating the problem yet.
Could you check one thing: If you connect up the Radian software via the Radian interface to the roll or tilt sensor, when you operate your RC controls, do you see a corresponding asymmetry in the Slew bar on the Diagnostics window? I'm just curious to see whether the asymmetry is coming from the input to the system from the RC controller, or whether it's being generate "downstream" from there.
Those gain settings of 30/65/75 seem very low -- they might work in still air because the Radian's don't have to do much compensation. The 70/130/150 seem closer to what I understand is "normal" (for some value of normal -- meaning there really isn't a normal!)
Also, to embed video, click on the little projection screen icon up the top of the window in which you are preparing your message. That allows you to insert a link to Vimeo, Youtube etc.
Thanks for your feedback, Kari!
I worked with Gelpads before when I used an MK Okto frame with their camera gimbal. They do a pretty good job against vibration.
What I´d like to know is - how good can the system get with perfect settings, a very experienced pilot and a camera without stabilisation system at various wind speeds, so that I know what to watch out for. It would be great to have some sort of standard test to compare our flying with those of the guys from freefly who really know what they are doing. For example: from a stable position fly 60ft forward, then back, then 60ft to either side - up to an altitude of 60ft and down again - all with smooth flying and a target timeframe of 60seconds (to all have similar speeds).
Thanks again, Andy!
I´ll try to check the diagnostics window on Monday.
I´ll upload some footage of the 70/130/150 right now to give you an idea of how it looks like - will post as soon as it´s online - vimeo is getting slower these days ;-)
Thanks for the tip with the projection screen!!!
OK, here we go
The roll oscillation is quite high frequency, isn't it? It seems a bit faster than the natural frequency of the gimbal as whole. Especially at 0:16 in your video.
I'm not sure whether that's incipient oscillation in roll, or whether something is mechanically loose on the gimbal. Can you check the deflection of the roll servo belt -- at it's mid-point I think it should be around 2-3mm if you deflect it between the large and small pulley. Is there any "play" in the motion of the roll servo pulley and the large roll pulley.
Thanks Andy, I´ll doublecheck everything once again and keep y´all posted!
Okay, back again. Didn´t have a chance to fly the bird yet, but I checked everything yesterday.
One thing that might sure have had an effect was the pan servo pulley - it was pretty lose - so everything might have been shaking up from here on. Pan was the worst shake.
The roll belt was the tightest of all, but nevertheless I tightened them all, didn´t expect to need it that tight - 2-3mm was a good tip to get an idea.
Haven´t yet had time to check the asymmetry.
Did some more testing in a studio environment (btw: a nice test stand is one of the background systems used for photography - I use a different one, but this one should work fine too and is only 59.99 US$).
Still no flight yet. However, my gains are now much higher (70/130/130), hopefully they will stay as they are after the first flights.
I forgot the Spektrum DX6i at home and tried my Graupner MX-16 with GR-24. The asymmetry was gone. However, one problem I found with the pan: one direction is very smooth, but the other direction is stuttering - in one direction I could put the gains much higher.
In general I prefer the Graupner Radio to the Spektrum, much better feel. Maybe I´ll give it a try again. I had some problems with it though while I was flying the gimbal with a MK FlightControl (the two radios had interferences and the cinestar crashed - hopefully that´s different, when using the radians).
Yeah, that's a good floor standing solution. I use a couple of lab stands (purchased from any high school science supply outfit) and a Bogen boom -- but you could probably also use a simple broom handle.
That is very odd -- if there's no asymmetry in the transmitter/receiver (and I think you've proven that point), what are the other possibilities?
1. If you remove the pan servo belt, is there any change in resistance to rotation?
2. Without the belt on, does the pan servo show asymettry in its motion?
I guess what I'm asking is whether there is any chance that it could be mechanical?
Hmm. I flew with an MX-20/GR24 (copter) and Spektrum DX6i/AR6210 Satellite (gimbal) and didn't have any interference at all. Where did you have the Graupner receiver mounted -- could it have come too close to the Spektrum Rx?
Thanks a lot again for your help - so far it has done quite a lot!
I´ll try to check with the pan belt removed. Not sure if I´d be able to detect an asymmetry of the small servo pulley. Perhaps I can replace two of the servos once I have time.
I didn´t have problems with the Satellite and the Graupner, but with two Graupners, I´ve written about it in this thread. But it was "pre-radian" using two flight controls.
Thanks again, I´ll keep you updated
I've used the trick of sticking a 10 cm strip of Post-It material on the small pulley to amplify its motion so I could see what it was doing better.
Yeah, I could see how two Graupner's with telemetry might interfere with each other, especially when the receivers are close.
Too bad, I was hoping I could try the MX-16 again with the radians. I´ll better go back to the spektrum route then. Thanks for the input.