@Bo: Sounds like you've fixed your problems with the upgrade -- that's good. I hear you with the weather problems....it's been awful in Portland, Oregon too. Rain in the valleys, snow in the mountains. Good for skiing, bad for flying. Andy.
Hi again, One day Ill move to Portland - loving the mountains and the skiing.. Here we have snowing at sea level... I´m not saying that I fix the problem. I´m just saying that I don´t know if the upgrade will do me any good. I can´t check it be course of bad weather. Did you see my Radian settings? Cheers Bo
Hi Bo: Two things: 1. If it's not too much trouble would you be kind enough to change your user name from flybypix to your first and last name, please? 2. I did see your settings....It will be interesting to see if the the new firmware improves the situation -- it reportedly fixes some twitching with PWM. Let me know what you find when the snow stops! Thanks Andy
Yes, it reportedly reduces some twitching motion with certain PWM setups. Check out this thread. Andy
There are some other features that might be of interest, Pavlos. Here's the release note for 1.03. Andy.
Hi Andy, Did you get a response from Tabb on the small jitteriness oscillations in the roll axis? (Post #203) It still looks like many who have commented in this thread have the same problems, so an acknowledgement by FFS would be a good start! I see that FFS moderators only comment on others' successful outcomes. Cheers Matt
Hi Matt, Sorry Andy emailed me about this post and I forgot to follow up. The FF moderators are not deleting any posts other then spammers. On post like this I am never quite sure exactly what to write that would be most helpful for you. You are correct that any slop in the system (be it servo slop, belt, pulley, etc) can contribute to a feedback oscillation in the Radian / CineStar Gimbal. A noise filter is a possibility but adding a noise filter to a system like this is somewhat counterproductive. We want the Radian to pickup the tiniest most minute movements possible and correct for them. Where the problem lies is the use of hobby type servos. We are asking them to do a task they were never intended for and asking them to do it to an unbelievable level of precision. We are faced with a set of constraints that are opposing each other and make tuning the entire system complex and time consuming. We do not want any vibration in the image....BUT a loose mounting attachment will result in an inability to tune the Radian Gains high enough to be satisfactory. We want super smooth axis actuation but it also needs to be blazing fast. My best advice is to work through the system in a methodical fashion. Swap your roll servo for another servo and see if the results are any better. There should be no vibration (that would affect the radian) getting through to the CineStar 3-axis mount provided you are using the CineStar isolators. I think John is looking into remote gain at the moment, but certainly if you use the tuning method of hanging the multi rotor gain should not need to change frequently. Best Tabb
Hi Matt, we are doing our best to keep up, we are a small company that is growing quickly. Our goal is to have all our customers have successful outcomes. It typically takes a lot longer to write back an in depth response on the difficult posts then to type "nice job - looks great" which is why you might see that more from me. I am not quite the forum - ninja yet that Andy is Tabb
EDIT: I can see that Tabb has responded, but I think my posting is still relevant (well, I hope it is)! Oh. Sorry. No. I was hoping that he would have a chance to respond rather than tell me what his response would be. Let me try and rebuild the thread so I have an accurate picture. You mentioned in post #203, that you were having " juddering of the Cinestar 3 axis gimbal with Radians" -- "EXACTLY the same issues as Brad." In post #175 Brad said: I have my tilt running at 147 right now and it's dead on. Tilt for me doesn't seem to have much issue from 70 up to 150.. At lower numbers it does ten to lag behind but it's not hard movements. Pan is pretty much the same situation and I think I have that around 130. Rotation however still have a tad bit of wobble that I haven't been able to eliminate. It's set at 135 right now. Canon 7d and heavier lens.In post #190 Brad said: Andy, I was running 70 at the beginning of this thread and never had a problem but I noticed recently in one of my videos that it seemed to drift around in that axis very slightly. After upping the other two gains I decided that id try the pan as well and I havent had an issue with oscillations.Let me ask a few questions to help me get my head around your setup: 1. Given your "exactly the same issues" comment, am I correct that you're having issues with just the roll axis? 2. So may we start with some fact gathering because I don't have a clear picture of what your set up is beyond that it's a 3-Axis system. I'm inferring that you're running firmware 1.03, right? 3. Can you confirm that you're using Freefly servos please? 4. Can you tell me how you have the roll Radian sensor connected? What goes into each port, please? 5. Also can you tell me which receiver you are using and how it connects to the roll Radian sensor. 6. Can you reproduce the problem with the roll axis on the ground when you have the Cinestar suspended by the battery plate and it's free to turn (you may have to put your hand on a boom to simulate the flight dynamics as it's likely to pan slowly). 7. If you can reproduce the problem on the ground, could you connect up the Radian software to the roll sensor, do a read operation and take a screen shot of the diagnostic and configuration window please? 8. Do you see any positional "jittering" in the horizontal slider for the Slew control, please? You know, I hadn't realized that. Speaking only for myself, I suspect it's because if I cannot diagnose a problem remotely just using words, video and still images, I run out of things to suggest, so I tend to fall silent rather than offering platitudes that, were I on the receiving end, would not be helpful or would only serve to piss me off. I am probably applying to myself the adage that it's better to keep my mouth shut and be thought a fool than open it and prove the point. On the other hand, if by some lucky break, I help someone to diagnose what the problem is, there's a sense of relief and that triggers a comment. But, as I say, that's an interesting observation. I'm not quite sure what constructive or helpful comment I could make for unsuccessful outcomes, but I'll see what I can come up with.... Oh....finally, and not at the same level of priority as the above, would you be kind enough to change your user name to be your real first and last name. The reasons for this and details of how to do it are at http://forum.freeflysystems.com/index.php?threads/real-names.497/ Thanks Andy.
Hi Andy, thank you for your response. I've placed my answers inline in RED below. QUOTE: Andy Johnson-Laird Let me ask a few questions to help me get my head around your setup: 1. Given your "exactly the same issues" comment, am I correct that you're having issues with just the roll axis? Correct. Only the Roll Axis. The tilt has the odd jitter but we resolve that via gain adjustment, sacrificing correction speed for smoothness. 2. So may we start with some fact gathering because I don't have a clear picture of what your set up is beyond that it's a 3-Axis system. I'm inferring that you're running firmware 1.03, right? Yep, Using 1.03. 3. Can you confirm that you're using Freefly servos please? Yep, Using New FF servos in a new CS Gimbal. Belts move 3-4mm with a firm press. Incidentally, at this tension, the head of the servo belt pinion tilts as the pressure is so great. But we have found that this is required to reduce the jitter to the small amounts we are getting. 4. Can you tell me how you have the roll Radian sensor connected? What goes into each port, please? Radian sensor is connected as per the instructions. And double side taped to its mount point on the rear of the CS Gimbal. It's daisy chained from the Pan Radian with the lead going into port 1 of the roll radian. The port 3 output of the roll Radian goes to the tilt Radian, and the port 4 of the Roll radian goes to the roll servo. We have it correctly set-up. 5. Also can you tell me which receiver you are using and how it connects to the roll Radian sensor. Receiver is an FRSky S-Bus receiver. 6. Can you reproduce the problem with the roll axis on the ground when you have the Cinestar suspended by the battery plate and it's free to turn (you may have to put your hand on a boom to simulate the flight dynamics as it's likely to pan slowly). The problem does not happen: 1) On the ground 2) In the air with Radian off and belts connected 3) In the air with belts removed 1/2/3 above have buttery smooth footage, tho unstabilized of course. 7. If you can reproduce the problem on the ground, could you connect up the Radian software to the roll sensor, do a read operation and take a screen shot of the diagnostic and configuration window please? The problem can only be reproduced in the air with the full system connected and powered on for a normal flight. I'm on a shoot at the moment and when I find time I'll post a screen cap. FWIW, We are shooting with an FS700, and even with stabilisation set to Active which is the highest, we still get roll oscillations. 8. Do you see any positional "jittering" in the horizontal slider for the Slew control, please? I'll look when we get back. I assume you may be thinking this is a Rx driven issue? Thanks Andy. Can you let me know the update rate of the Radian output PWM? I'm considering hooking up some direct drive motors to eliminate the possible backlash across the system. Thanks again Matt
Hi Matt: Yes. I think there may be noise in the system that's causing the problem -- I can't explain why it's smooth with Radians disabled. So the pivotal question is why you're seeing twitching in roll when the Radian system is active. The FS700 is a large enough mass that it would suggest that it's not mechanical play in the system but that the Radians are affirmatively moving the gimbal. I'm afraid I don't know the update rate of the Radians. Let me see if I can find out. Andy.
AT the end you have generally higher Tilt gain then its Roll and the smaller gain around 40-60 is pan... Right??
The general ranges I've seen people using are: Pan: 50-80 Roll: 130-150 Tilt: 150-220 It almost seems as though the relationship is that the gain is inverse the mass in motion and Pan <= Roll <= Tilt. But that's just me searching for patterns. These are just general ranges as I recall them, so don't be surprised if your settings are outside these ranges -- the settings have to be determined experientially. Andy.
Today I moved back to 1.02, and with the old settings, but the servojitter made me turn back to 1.03 again and tryed the same settings I ran with 1.02. Now all works fine..! Figure that. Its like if the radians needed a reload of the newest fw?! Well mostly the roll atleast
Hmmm. Interesting Sedric. Well, glad you have fixed it, but odd that it needed a firmware refresh.... Andy.
Completely agreed on that Andy. After the "normal" update the roll was behaving strange, had to put ALOT of gain on it, and the middle was jerky. Now its like with the 1.02, same gains as with the 1.02, but without the jitter from the 1.02
It adds one more thing to the "if not else works try this" list, doesn't it? Keep us posted if the troubles returns, please. Thanks Andy.