Photohigher skyline rsgs

nicwilke

nicwilke

Active Member
Macsgrafs said:
How can continuos servos (as supplied witht he AV200) have a centre if the pot has been taken out & replaced with 2 x 2.2K resistors???????
Click to expand...
You are centring the position that the skyline registers the servos at, not the other way around. If the skyline is level, it indicates the servos to be at a certain position. If this is off, it will effect levelling. I'd suggest removing belt and roll gear, and making the servos stop while tweaking the servo position settings. Make them stop!!!. Gear up again, and then do an auto tune.
When the manual tilt (or roll) is applied to the equation, the leveling algorithm is more complex because now the angle/s need to be stabilized whilst the skline is off level. Anyways, heres my latest test of the skyline, and I concur the drift on horizontal on firmware 1.5 beta needs a little tweak. Brendon, well done on this version is bloody close, and fully usable.
I'd also suggest to improve the roll drift on panning that people reduce the roll integral gains to zero.

[video]www.youtube.com/watch?v=Hpdq0N06uLo&hd=1[/video]
Mind the little skipping, it was filmed in 60fps and I forgot to conform it to PAL. Dah!!!
 
Last edited by a moderator:

V

vip

Member
I suggest to search this thread, they indeed do have a center. I had the same problem you are having. You will see when you do the steps.. all of a sudden it will pause then depending on your input ..plus or minus will effect which way it turns.

I am far from an expert but It worked for me , try it..
 

nicwilke

nicwilke

Active Member
vip said:
You will see when you do the steps.. all of a sudden it will pause then depending on your input ..plus or minus will effect which way it turns.

I am far from an expert but It worked for me , try it..
Click to expand...
Exactly rite mate, dial out the movement. It comes to a stop.
 

GeorgeM

GeorgeM

Member
Hi Ross... Yes indeed.... A 360 servo has a neutral point....the point where it stops turning.
The best servos for stabilization normally have this point ( called deadband ) which is very narrow.... the best I've seen is 0.4uS !!
The worst ones...as much as a Full FAT ... 8uS !!
You must accurately find this centre point, for each servo, and memorize it in the PH software.
Beware... some servo's centre neutral point changes a bit with 1) different temperature, and 2) different voltage.
But.. this is usually taken care of by the algorithm of the stabilizing device.

:)

gm>>
 

T

tombrown1

Member
nicwilke said:
You are centring the position that the skyline registers the servos at, not the other way around. If the skyline is level, it indicates the servos to be at a certain position. If this is off, it will effect levelling. I'd suggest removing belt and roll gear, and making the servos stop while tweaking the servo position settings. Make them stop!!!. Gear up again, and then do an auto tune.
When the manual tilt (or roll) is applied to the equation, the leveling algorithm is more complex because now the angle/s need to be stabilized whilst the skline is off level. Anyways, heres my latest test of the skyline, and I concur the drift on horizontal on firmware 1.5 beta needs a little tweak. Brendon, well done on this version is bloody close, and fully usable.
I'd also suggest to improve the roll drift on panning that people reduce the roll integral gains to zero.

[video]www.youtube.com/watch?v=Hpdq0N06uLo&hd=1[/video]
Mind the little skipping, it was filmed in 60fps and I forgot to conform it to PAL. Dah!!!
Click to expand...

OK, this is very depressing. That horizon looks awful. Somebody must make something that keeps a level horizon when you do a slow 360 pan. I've been waiting with baited breath for 1.5, and if that's the best it can do then it is a total failure and unuseable unless you plan on keeping the sky out of all of your shots.

I've spent so much time with this thing it's unbelievable. Maybe I'll try Radian or Picloc. Any suggestions for alternatives until PH gets there s*** together?

TB
 

nicwilke

nicwilke

Active Member
Cool off Tombrown1, its only a beta mate. Its testing, and In my opinion, its really good and close. I have my roll frame quite firm, perhaps too firm. See from 57 seconds, the pan was quick and slowed down, and I think it may have even been a little 'stuck' in position. There's no point jumping ship, the skyline is great. We have to also acknowledge that smooth video comes from smooth flight. I'll fix up the gimbal tightness this evening and have another go, as its intermittent, so probably got little to do with the skyline, and more to do with the tightness of the gimbal. Im using an AV130 with no pan, my pans are with the hexa, nothing else. Radian users are complaining about drift, but most fix it in post because they are busy with getting on with the job. Stay with it, they are so close with the skyline.
 

Macsgrafs

Macsgrafs

Active Member
Nic, great video by thew way...very impressive.

Thanks guys, your I/P is appreciated....

OK so let me get this right.

1)take off the drive wheel on the roll servo
2)take off the belt on the tilt servo
3)tray should now be swinging freely
4)connect RSGS to the CPU
5)make sure tray is level
6)use transmitter to dial out any servo rotation
7)connect servos back to drive belt & gear wheel
8)use auto tune in the software?

Ross
 



T

tombrown1

Member
nicwilke said:
Cool off Tombrown1, its only a beta mate. Its testing, and In my opinion, its really good and close. I have my roll frame quite firm, perhaps too firm. See from 57 seconds, the pan was quick and slowed down, and I think it may have even been a little 'stuck' in position. There's no point jumping ship, the skyline is great. We have to also acknowledge that smooth video comes from smooth flight. I'll fix up the gimbal tightness this evening and have another go, as its intermittent, so probably got little to do with the skyline, and more to do with the tightness of the gimbal. Im using an AV130 with no pan, my pans are with the hexa, nothing else. Radian users are complaining about drift, but most fix it in post because they are busy with getting on with the job. Stay with it, they are so close with the skyline.
Click to expand...



Thanks for the reply Nic. Hopefully this can be worked out before the release. It seems like this thing was already beta though.
 

nicwilke

nicwilke

Active Member
Yeah mate, PH has already said it was released a little early. Hang in there for them to tweak this up, I'm sure you'll be happy.
 


Efliernz

Efliernz

Pete
I have an intermittent problem. 1.2.2 firmware as supplied - late November order so it is the latest edition. I calibrated it and it works on my diy mount. Not perfect but it works ( well my mount works well... it is just the rsgs that is not quite there.
Now suddenly it goes through its power-on-test (already sitting level) and in the end, it tilts slightly down and flashes red (as though it has the wrong direction).

I have repowered lots of times, made sure it's level before powering up.

It has worked. I will try to recalibrate it but I shouldn't have to.

Pete
 

V

vinnygogo

Member
Today I tried the new 1.5. Beta version. I did firmware upgrade, Autotune after xyz level.tilt stab was great, roll was great(on the Bench) but then I picked up the 360 mount, and did a pan. That went TERRIBLE! When turning the gimball, the roll just completely rolls to one side before even doing 180 degrees. And it then realllllllllllly slowly goes back to level, or turning back opposite way will help.

I thought this was fixed, but this was way better in 1.2.2 or is it possible a setting? Tried various gains, but could not get it out. If anyone could help please let me knowasap, otherwise I will move back to 1.2 tomorrow morning, because of video shoots planned for Friday and Saturday.

Thanks in advance!
 

P

photohigher-b

Brendon
vinnygogo said:
Today I tried the new 1.5. Beta version. I did firmware upgrade, Autotune after xyz level.tilt stab was great, roll was great(on the Bench) but then I picked up the 360 mount, and did a pan. That went TERRIBLE! When turning the gimball, the roll just completely rolls to one side before even doing 180 degrees. And it then realllllllllllly slowly goes back to level, or turning back opposite way will help.

I thought this was fixed, but this was way better in 1.2.2 or is it possible a setting? Tried various gains, but could not get it out. If anyone could help please let me knowasap, otherwise I will move back to 1.2 tomorrow morning, because of video shoots planned for Friday and Saturday.

Thanks in advance!
Click to expand...

This happens when the skyline is tilted when the set level xy is done. Make sure you level the skyline not the camera tray.
 

Efliernz

Efliernz

Pete
nicwilke said:
Yeah mate, PH has already said it was released a little early. Hang in there for them to tweak this up, I'm sure you'll be happy.
Click to expand...

I hope they do something soon.
I'm about to rewire a roll servo that got its wires shredded after the "Autotune" rolled and rolled and rolled... I am feeling it is time to sort out feedback pots and let my Hoverfly pro have another go... HF isn't perfect but once it worked, it worked. If this rsgs isn't ready for market, I'll have my NZ$450 back please. Yes I am feeling frustrated.

Pete
 
Last edited by a moderator:

P

photohigher-b

Brendon
I've uploaded the new firmware to the website as version 1.6 (Not to be confused with 1.1.6)

This is mostly the same as 1.5 with the roll issue fixed. There was a bug in the accelerometer correction when the pan was as angles -90 to -180. Unnoticeable when panning quickly but quite bad when holding in that range or panning slowing.

Also the is no self-test when starting up with the AUX in the off position.
 

nicwilke

nicwilke

Active Member
Efliernz said:
I hope they do something soon.
I'm about to rewire a roll servo that got its wires shredded after the "Autotune" rolled and rolled and rolled... I am feeling it is time to sort out feedback pots and let my Hoverfly pro have another go... HF isn't perfect but once it worked, it worked. If this rsgs isn't ready for market, I'll have my NZ$450 back please. Yes I am feeling frustrated.

Pete
Click to expand...
Theres so many factors Pete that go into gimbal tuning, and no one gimbal is the same. Even my AV130 would behave differently thatn anyone elses AV130 due to things like instal, camera weight, COG, previous crashes and bent/warped cages.... the list goes on. I had an issue where my tilt rolled forwards over the correct amount, and twisted my cables. I saw the Skyline was slipping on the mount bolt, and then in turn was not getting a grip on the tray axis. This is not PH's fault, its mine. Calm down mate, the skyline is way good.
 

Efliernz

Efliernz

Pete
I understand that there are variables Nick - I have been building + flying cameras for 6 years. What I find frustrating is that after I got mine working, strange things keep stopping it when I haven't changed anything. 1 week ago my mount worked well. Now it wont start. Shredding wires while calibrating isn't normal.
 

nicwilke

nicwilke

Active Member
No disrespect to your 6 years experience Pete, but you mentioned you had a nasty crash before Christmas with pitch oscillations or something? Perhaps there was something that happened in that crash to caused the skyline (wiring) or the servos to behave differently. I see you rebuilt the gimbal, looks good. A week ago it worked, now it doesnt? I had the same thing, and found that the startup skip/movement needs to be free to do its little quick thing unrestricted. My gimbal has some wires that were catching on the back of my tray that caused the skyline to fail out. Only suggestions mate. Got me wondering.
 

You must log in or register to reply here.
Top