I've been having a little bit of weird GPS activity that might be contributing to what was suggested as the NAV and GPS boards needing to be higher than any lipo wiring. Since I raised my battery plate with 60mm its conceivable that the Lipo wiring is now higher causing sparatic issues. My question is this does it matter if the Smart OSD board is the bottom board or if its above the Navigation board? If the GPS and Navigation boards are sensitive to metal I'm also thinking that having the Smart OSD board right below the GPS board might be a bad idea because of metal so I moved the OSD board to the bottom. I tripped up on the plastic standoff's I had just for testing as I only had longer metal ones (trying to keep metal away from the GPS). I haven't been able to test it yet as its been raining all day. Does anyone have their OSD board on the bottom? Does it matter for some reason that I don't know? Here's how I reconfigured the boards and a side perspective of how the Navigation board is above any lipo wires.
The stack I use is: GPS board (top) Smart OSD Navigation Control Board. I've not seen any problems or unreliability -- I only use one standoff between each board, but do have a 25mm stand-off right at the bottom to elevate the entire stack up from the boom. Andy.
Dave very unlikely that you motor wires in the tube are the issue. My stack is GPS-OSD-NAVI 25mm above the boom with external wiring.
I do the exact opposite as Andy and Gary. I am not sure if there is really a recommended solution I put NAV under GPS and the OSD on the bottom. My thought process was that we are trying to shield the NAV board from as much EM influence so it seemed to me to make sense to have it directly inside the hat. Another thought was that since the magnetometer is on top then it made sense to have it shielded from EM caused by the OSD. Then again I flew for the second day after upgrading to FC 2.2 and loading 0.90j not doing 2.0 until issues get resolved. I got TBE for the first time. It wasn't as nuts as others have had just a little gyrating around. I lowered the GPS Yaw to 80 which I thought was supposed to mitigate it so maybe I need to go down to 70, I don't. I hate automation anyway just thought I would try it. AH is a little better although I think I have a little overshoot going on because it sometimes oscilates at first, meaning the P value might be too high. The joys of tuning PID's again Looking at your picture Dave, Your NAV board completely exposed. I would put up it in the hat with the GPS. I think that is why DJI has their stupid dome on a pole, the Magnetometer is in the GPS thing with them. Another thing I do when calibrating the compass is that I turn on all electronic equipment like the Video TX's etc, so I can normalize the magnetometer to as many electric fields as possible. This is my theory, that I adopted from aviation when they calibrate the Whiskey compass to all the aircraft avionics. Another one of my unreasonable requests to the world of MR autopilot manufacturers is that it would seem reasonable that they would have an algorithm that would normalize the compass with the GPS as the A/C moves around it seems like that would be a straight forward thing to do. By our GPX files, the system knows when the EM field is getting jacked up, so I wonder they don't incorporate a soft calibration scheme in real time.
Hi Gary I never thought the motor wires were causing the problem. Ziggy thought that having the height of the LIPO wires above the nagivation board causes issues. From Ziggy "Your compass is looking at a really low magnetic field trying to separate it from the noise around it. When the lipos are in the way of the compass and the north pole every time you turn it affects the compass. The Lipo acts as a shield blocking true magnetic north. this is why DJI puts theirs way up on a stick And MK puts their lipo's below the Navi."
Hi Shaun That's the OSD board on the bottom now, the navigation board is tucked up inside the hat. The reason why I did this was sorta along the lines of your thinking but a little different. I was thinking that the actual OSD board could be giving off magnetic fields to the GPS board so i thought it would be good to have the OSD board as far away as possible to the nav and GPS.
I know he was just making an observation on pics I posted. I posted my "before" pictures and he said he has seen Cinestars similar to mine having this issue. In his opinion the compass looks for low magnetic fields (below its level) and if the lipo wires are below the compass it can cause interference issues. He also suggested cleaning up the lipo wires which I also have done. Here's essentially what he has said. I too have now seen two systems that I have had troubles with, that didn't show issue in 90j , Both I was able to sort out by rerouting the lipo wires. From your pictures your lipo wires are all over the place, I have never been a fan of the GPS/Navi near the lipos. or out on the Arm. I would say your pictures just make my case stronger. Earlier versions did not use the Compass the same way and also we had less issues. High Current makes EMF Your compass is looking at a really low magnetic field trying to separate it from the noise around it. When the lipos are in the way of the compass and the north pole every time you turn it affects the compass. The Lipo acts as a shield blocking true magnetic north this is why DJI puts theirs way up on a stick And MK puts their lipo's below the Navi. Twisting lipo wires doesn't really help much but keeping the wires together and short with no big bends or loops is the key. Reducing the number of possible noisy connections to help a lot. Tie wrap anyplace that the wires do not stay touched. Less Connectors the better, I use two one for each lipo. It can really be amazing how a small change can help. You may notice, even if you are flying ok that every other motor is hot or uses more current.I'd like to see the test with the compass sitting on your lipos when you power it up like in the video's.. The lipos really make a difference. If 90j is working for you stick with it, but my guess is your problem is still happening, it just compensates better for some reason. I'd like to know what Range your Earth Mag is at during flight. Even on 90j and Yaw it in a circle to get min and max.. Really I'd like to know this data for my research. Knowing your Earth Mag %'s is really important, it can tell you if you have a good calibration or not, and if the Current in your system is giving you problems. Trust me I feel your pain. I've spent a lot of time with this one, ever since we went to bigger motors and differnt frames. These are just the recommendations I can tell you that works for me, Good luck and let us know. Zig
Which will cause the most EMI, I wonder? The Smart OSD, or the more-or-less square waves at high current flowing along the motor wires? Based on some research I did, you can significantly reduce the EMI from the motor wires if you tightly braid them -- but that shortens them and you then have to run them along the outside of the boom. But I think we can easily resolve some of the (quite reasonable) speculation. If I get a chance, I'll do some experiments with the copter and recording the FPV video with SOSD. Here's what I plan to do -- and it might be worth other folks doing the same thing if you have the FPV-Japan DVR or some other means of recording the FPV. 1. Calibrate the compass. 2. Get airborne. 3. Point the copter at Magnetic North, East, South, and West based on the SOSD. 4. Land. 5. Compare the terrain features visible in the FPV and correlate it with ground features. 6. Compute the accuracy of the compass by comparing the SOSD Magnetic Heading with the known ground features. (Note: Google Maps shows True North to the top of the screen, so you will need to look up your declination and factor that in. http://www.ngdc.noaa.gov/geomag-web/#declination (Enter zip code, click on Get Location, then click on Calculate.) If you don't have a DVR to record the FPV video, you can work "the other way around." When airborne, point the copter to the N, E, S, and W based on land features and record the actual value shown in the SOSD magnetic heading. Andy.
Oh yeah that is your OSD, Nevermind. On that note you would think that if the OSD would be a "static" EM field. When I mean static the EM field is not changing or that the change is negligible. By that assumption if the OSD were on during compass calibration then the magnetometer is being normalized to that field. The motors on the other hand have an EM field that increases as more current goes through the wire the magnetic field increases and makes a bigger balloon. By tightly braiding the wires what you are doing is canceling out that magnetic field poles because the motors are three phase. The principle is that the if I have two parallel wires with current flowing in the opposite direction that the magnetic poles cancel each other out. If I were to put a third wire in the middle of those two in theory the third wire would not be generating any current. Andy's test of knowing the actual magnetic direction of the copter and comparing would be a good indicator of how much influence are causing. I am going to try that out.
I'm with Andy too. My Navi on the bottom, then Smart OSD, then GPS. One stand-off each. I've had no problems for a year with PH until 2.0. I've now reverted back to .90, much better. Of course I haven't tried the latest 2.0 z or whatever? lol Spent way too much time and energy trying to get 2.0 to work. Not to hijack but: I have a real slow update on my Xbee, takes 1-2 minutes to update wirelessly. Using "buffered " setting and not direct on the Smart and Xbee 3DR Radio set. Do you guys have real slow update to MKtools with the Smart OSD?
Gary: The current wisdom is NEVER to use the XBee or BT for firmware updates. The communications protocol used for doing the update does not appear to be resilient enough to handle timeouts, errors, retries etc. and there is a real risk that the update will fail and will "brick" the board whose firmware you're trying to update. I've also had failures connecting via the SOSD -- the best is to unplug the small ribbon cable twixt the NC and SOSD and jack into the NC directly. I only had one instance of a failed FC upgrade when I jacked in via the NC, so that's the "safest" bar actually jacking into the FC directly. Andy.
I don't do wireless firmware updates. I jack direct to the Navi board for firmware updates. I was just chatting about MKTools with the Xbee attached through the Smart OSD, which I use for FC setting changes and to check things intermittently. Seems it takes forever to connect wirelessly.
Gary The line of thinking is that the GPS and Navigation boards look for signals below its centerline. Some of us have raised the battery plate pretty high (50mm to 60mm) and the height of the lipo cables could be inducing magnetic forces. There's really no proof of this just a theory so that is why I raised it.
Update: I went flying Sunday with the new configuration of the raised GPS and Navigation boards. Interesting feedback. I didn't notice any tighter pattern for the copter (probably because its consumer GPS grade) but I did notice that it was much better in the wind - very planted in 15-18 mph winds. My copter always seemed to spaz out in higher winds trying to correct. That seemed to be gone. Also when I was in PH, it went straighter when I was overriding the PH with the joysticks trying to fly straight. NO GPS or PH glitches neither (this was the reason why I tried the experiment of raising the boards).
90J. I had a couple instances where the GPS just spazzed out. Luckily enough I just killed PH and it was good. I was thinking lipos interference like Ziggy suggested might have been causing it.
Yup, agreed I am guilty of that buffoonery. I almost bricked the FC two weeks ago trying to do an update through the OSD. I knew better but did it anyway because I lost my FTDI. Luckily Andy knew a trick to un%*#@ it.