Worked out what I did wrong. I had added a piece of heat shrink around the sensor, 10mm diameter, squishing the foam a little. I thought this would help with prop was and wind. But it seems it was doing a few things, it was funnelling the air on to the sensor, i tested by blowing at the bottom of the board. With no heat shrink it moves 5-10, but with it moved all the way up to 20-30. But also, just sitting on the bench it was causing the readings to fluctuate. Something to do with the pressure on the foam i suppose. I went out and tested it for 40 minutes this morning, retuning alt hold. It was much better. Its a nice day here with little wind (for once) so il have to retest on a windy day to see how it handles. One other thing that I've been wondering: I loose compass calibration after each days flying. Im flying at the same spot, but each time i go back the motors won't start and i have to redo the calibration. Is this normal ?
You mean that you have to do the whole tumbling the copter calibration thing? That should not be happening. Sounds like there might be a problem with the navigation control board failing to preserve data in it's on-board flash memory. Andy.
Does the copter say "Error: Calibration"? If so, that's not the compass, it's the gyros. That's normal. They need to be calibrated after every power cycle. Just move the left stick up and to the left.
Good point -- except you cannot take off unless you do gyro calibration before each flight...and it sounds like Ryan is just having to calibrate each day. Andy.
Yea its the compass calibration, error, calibration. Which im having to do each day. I think i may have worked out what happened. I calibrated in the middle of the field. which is inside a velodrome, so lots of concrete and i assume metal under there. If i plug the battery in when I'm on the concete it doesn't work, but if i do it out on the grass it does. Il have to go back tommorow and see if it hold the compass calibration over night.
If you have an iPhone (or smart phone with a compass), you can get a "second opinion" on whether the magnetic field is distorted, by getting a reading for North in one place and comparing it with the reading you get in the Velodrome. You obviously need some distant landmark to use as a reference point. Andy.
Your iPhone has a compass in it, but if you have a real compass, I'd use that. You may need to calibrate the iPhone compass: http://support.apple.com/kb/TS2767 before you use it. Andy.
When doing a compass calibration, is it vital that when flipping the multi rotor it stays pouting exactly north? Does it make a noticeable difference if it wobbles through the rotation?
As long as the "wobble" is not too extreme, no. The real test is to calibrate the compass, fly, and then look at the GPX file which shows the results of the calibration. Andy.
Hi thats what i had done in the past, i tested it again with out the heat shrink that was clearly causing a few issues. It still moves around a lot, all the way up to 24. There was a bit of a slight draft in my office due to the old windows, the numbers didn't seem to match the draft. But i thought maybe the draft was causing it to move. So i got a large bit of heat shrink and hot glued it in place around the foam of the sesonor, making sure it didn't touch it. Then i got some tape and sealed the top of the heat shrink. Just to verify how much it was moving with no airflow. It still moved around a bit, while it was noticeable less and slower movements; it still crept up to 8-10 then back down. Here is a video of it with with jus the foam over the sensor: This has me thinking that there is defiantly an issue with the sensor.
When I was having altitude hold issues Holger suggested covering the sensor to rule out the issue. I put heat shrink tubing over the sensor and foam and it left a little 1/8 hole on the end of the heat shrink so that air could get to the sensor. The heat shrink tubing was about 3-4 inches long past the sensor so that the only light coming into the sensor would be from the dark bottom side of the hub where there really isn't any light. When I tested this in flight the AH was much worst and going up and down a lot and would just randomly drop. When I removed the tubing it went back to its consistent deviation of up and down but was no where near as bad as it was with the tubing. A new flight control board solved the problem.
Dave can you plug yours into the computer and just have a quick watch for me; to check how much your numbers move around? Il contact support again and see if i can sort something on warranty. Since its practically new.
I think the goal is to permit gentle airflow without sudden compression and rarefaction gusts because make the sensor think that it's suddenly going down or up -- so I've found that the best method is to use two pieces of self-adhesive Velcro. The part that's stuck on the sensor has a hole punched in it where the hole on the sensor is -- and the other piece of Velcro is left with its backing on so the adhesive is not exposed. Then I mesh the two pieces together -- the airflow then moves in and out laterally between the hooks and the loops of the Velcro. This seems to remove the gusts pretty well. Andy.
Would you suggest i pull off the foam and try this method even though its a 2.5 board? Also do you have a photo of it?
Oh..sorry, Ryan. I spaced that it's a 2.5. Duh. In which case, having refreshed my memory by re-reading the entire thread, I suspect you may have a rogue FC 2.5 board, I'm afraid. Andy
I have re-contaced support and I'm going to RMA the board. But il have to ship it back from NZ, so it will take a while I'm guessing. Does any one know of anyone selling a 2.5 board? Il need to buy one to use in the mean time. I also re-watched your MK videos Andy and observed the barometer reading and noticed mine is defiantly not the same as yours. jumping around around a lot more.
Hi Ryan: Yeah, bummer about being so far away....just had a house guest from Hikurangi -- long flight to Portland, OR. You'd expect to see *some* "jitter" from the sensors (well, more from the analog to digital converter), but it sounds like your board is a bit munted. Andy.