Help!!! Do I need to rebuild this Quad? *DRIFTS*

Started by rohan.khairnar, March 06, 2015, 03:37:18 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

rohan.khairnar

My quadcopter build:
X525 V3 Frame (HK)
APM 2.6
Ublox GPS module with compass
EMAX Simon Series 20A ESC x4
Turnigy SK3 2826 1240KV brushless motors x4 (HK)
Power Distribution Board
Cellcore 3300mah 3S 25C LiPo Battery
Turnigy 9x Transmitter with Rx.
1045 Propellors x4
Approx weight: 1.4kgs

I have got my quad in the air, landed it a few times and crashed it the rest of times. During takeoff the quadcopter will lift a little from one side, I fix this manually by giving little push to the yaw and raise the throttle and that's how it gets in the air. But even when in the air the quad will keep drifting to one side. Then I have to push/pull it back to level. I don't know where the problem is. Say, if it keeps drifting forward then after every while I manually pull it backward. All flight tests were done in 'Stabilize' mode.
I have:
Re-calibrated all the ESCs
Calibrated the accelerometer, compass(of GPS module)
Changed the orientation of APM(forwad direction)

Last crash:
While the quad was in air(really high) and drifting as always I suddenly lost control of it and it kept drifting. Then it flipped(I assume the battery went off) and fell on the ground(lucking through a tree). That damaged one of the four motor mounts and battery as well. I have decided to dump the battery and have already got one from rcbazaar. I assume signal loss to be the reason of this crash.
I'm awaiting for the motor mount, till then I still keep jotting down the PID setting values and try to understand what could be the problem. This is my first quad and I really wish for a smooth flight.
Any suggestions what could be the problem? I don't want to get bankrupt changing propellers. ???

Recent Changes:
Battery: Wolfpack 20C 3S 4200mah (rcbazaar)
Also upgrading T9x to FrSky DJT combo.

I have been suggested that the APM needs to placed perfectly in center. Awaiting flight test as I haven't received the Motor mount, do I need to make any changes? Buy new parts? Replace any?? Or rebuild? ???  :(


//I intend to implement autonomous flight on this quad using FPV by implementing image processing(tracking algorithm) but I'm stuck here itself since a month. The tests were without FPV cam. Still for the record I have bought boscam camera and video telementary set.
And without FPV my Quad weighs around 1.4kgs.//

parichya.gautam



One thing i would like to suggest to you is dont install the latest firmware, install a one step down one, it solved half of my issues :D


and regarding the drift, i have used the gps so wouldn't know the correction for the issue, but yea while lifting off, i did use an opposite throttle to get the quad up ( HAPPENED A LOT OF TIMES NOW )

ALSO... have you done autotune ? It helps by a huge margin.
Passion > Financial Constraints

To Contact:
FB: https://www.facebook.com/parichya
Mob: 8375066843
Mail:
-------------------------------------------------------
My Products On Sale Catalogue: ( !!! LOWEST PRICE GUARANTEE ON ALL ITEMS !!! ) https://app.box.com/s/84m5t5ljrwat54rmrldjg74dtie28g44
-------------------------------------
Product Request Form: https://docs.google.com/forms/d/1-OccRCaAQux8hxpuTw_xI4k3Ll9QW9JyOr0XDijzPXA/viewform?us

rohan.khairnar

Well yes, I did upgrade to the latest firmware. I did not know we can downgrade it.
Also, even I am using GPS.
Autotune yes I have gone through few video tutorials but couldn't get my quad in the air that well so didn't try it. I have seen people having no problem at all with takeoff. A straight vertical takeoff is all I wish for.
Thanks for the suggestion @parichay.gautam I will try that :)

Swapnil

I also had this problem initially. Recalibrating the radio inputs worked for me. If that doesn't work, try auto-trim.

rohan.khairnar

Quote from: Swapnil on March 06, 2015, 08:05:07 PM
I also had this problem initially. Recalibrating the radio inputs worked for me. If that doesn't work, try auto-trim.
recalibrated a lot of times. Also i am now upgrading T9x to FrSky DJT combo. Hoping it to be a good upgrade.

akhilzid


rohan.khairnar

Quote from: akhilzid on March 06, 2015, 10:57:24 PM
what about CG? is that balanced?
i assumed that APM will balance out the CG. But now i have been suggested to keep the APM centered. Awating motor mount will test further.

akhilzid

Quote from: rohan.khairnar on March 07, 2015, 02:09:00 AM
i assumed that APM will balance out the CG.

To get the best performance you should balance CG by distributing equal weight to all motors, else you can always expect drift(proportional to weight) to the heavier side(depends upon motor power and quad weight).

miginstruments

one more thing that I would suggest you is go for lower kv motors. for your size of quad a 980kv sunnysky motor would do much better.

sanjiv1994

for better results start by checking the motors and escs.Calibrate each component step by step rather than putting everything together and then figuring out what's wrong.then check the receivers signals received by apm and finally gps.You will be able to fly properly

SideWinder

Dont you log? Can you put up a log of your crashes? Logs are kind of a 'blackbox' for these aircraft's and APM has it.
It really makes things easier if you can get one. I would suggest anyone beginning with APM to enable logging to all data.

rohan.khairnar

Thankyou for all the suggestions :)

Quote from: miginstruments on March 07, 2015, 11:08:13 AM
one more thing that I would suggest you is go for lower kv motors. for your size of quad a 980kv sunnysky motor would do much better.
yeah i got that later. Initially while buying i had very less idea about kv thing, i am assuming total weight of my quad to be approx 1.8kg, maximum 2kgs.

Quote from: sanjiv1994 on March 07, 2015, 01:21:24 PM
for better results start by checking the motors and escs.Calibrate each component step by step rather than putting everything together and then figuring out what's wrong.then check the receivers signals received by apm and finally gps.You will be able to fly properly
I did all the calibration step by step, with one esc at a time. i have been through logs and i think gps needs to be placed a little away from apm.
also, i am uploading the logs now.

Quote from: SideWinder on March 07, 2015, 04:21:48 PM
Dont you log? Can you put up a log of your crashes? Logs are kind of a 'blackbox' for these aircraft's and APM has it.
It really makes things easier if you can get one. I would suggest anyone beginning with APM to enable logging to all data.

sir yes sir i'm uploading my last two logs(same day flight tests) right away. I went through the files but couldnt understand the parameters much properly. I should have done that earlier, thanks for the idea. Also i have downloaded some ApmLog analyzer applications, hoping that should help.

rohan.khairnar

can anyone tell me how do i upload logs here? sorry for being such a noob.
it says maximum attachment size should be less than 200kb. I tried to zip the file but still it counts 259kb. any further compression techniques?

rohan.khairnar

Quote from: rohan.khairnar on March 08, 2015, 04:23:45 AM
can anyone tell me how do i upload logs here? sorry for being such a noob.
it says maximum attachment size should be less than 200kb. I tried to zip the file but still it counts 259kb. any further compression techniques?

LOG FILES ;D

just found a public upload file sharing service.
Here is a link to download log files: http://www.filedropper.com/2logsapmdriftcrash
bin & kmz files: http://www.filedropper.com/2logsbinkmzfiles

i understand that gps and loss in rx/tx connection to be the problem. still most of parameters are confusing. Also i plan to order reptile/deadcat frame on which i can properly place all the things onboard, keeping everything well organized.

SideWinder

#14
Hi

First of all please post a pic of your setup.Now exhaustive logs, specifically the second one with so many fail safes and errors.

1)Your logs show too much vibrations, x/y axis are between -5/5 while z axis seems to be even poor and very bad magnetic interference. So you definitely need to mount your controller properly and place the mag module away from power cables or motors on a mast etc as it is 200% approx in your case while recommended value is below 30%. Also autotune/tune your copter as bad PID's can also be the reason.

2)CG is not centered as two motors(rcout ch1&2) are working harder than the rest of the two which clearly answers your drift issue. I also see some unrequested yawing, which is usually caused by windy conditions or improper yaw  PID's or magnetic interference which seems to be your issue. So place mag module away from possible emi sources and orient it correctly.

3)Your VCC too is going down to 4.7v whenever your are throttling up, so use a ubec etc. Powering via esc's/PM is a bad idea. APM 2.6 needs 5v supply and you need to supply at least 5.3v as a diode drops the voltage by 0.3v I think. It is recommended to use a dedicated ubec etc.

4) Your throttle in & out track really well, and as per log you are getting hover at 45/47% throttle which is decent in my opinion, but if you want to add more payload, you would need to upgrade battery/motors/props etc. Go for a H frame instead of deadcat, as it would provide you some more real estate.

5) Probably need to calibrate your power module as it shows your pack voltage at 11.5 before flight or was it the case? And disable battery failsafe in APM & use a lipo alarm instead for the moment.


Your first logs show you were suddenly loosing gps link for reasons I cannot understand since you were flying in a field. Just check if your GPS/Mag module is properly connected or mounted and also place it away from FPV Tx if any. Wait for atleast 5 mins after powering copter to get a 3d gps lock. Check lights on APM to ascertain it has a gps lock. Altho you were flying in stabilize mode hence GPS failsafe dint trigger alt hold or land as in your case. GPS plotting shows you had crossed St.Louis Convents road but i dont think that would be the case.

Your second log shows a crash due to battery going down, but I see that you had requested very high throttle in the end after you switched to stabilize from RTL which probably lead to brownout of esc's or FC itself. Battery voltage is being shown at 10.3v in the end. In the end I see that you increased the altitude of the copter to 40m and then to decrease it you bring your throttle to ZERO(which should never be the case) along with a radio failsafe which triggers RTL and it clears . This leads to a freefall for your copter which tries to maintain it attitude. Then you see it coming down really fast & give it full throttle which leads to a brownout and crash. I think probably the FC gave up because the FC hasnt recorded the crash . This log also has ekf/dcm failsafe which is triggered "when the GPS implied heading and DCM’s estimated heading disagree by at least 60 degrees (configurable with the DCM_CHECK_THRESH parameter) for a full second." which can be caused by a disturbed mag module.

To summarize:

-Mount FC properly on a foam pad etc. Google to find options. I prefer anti vibration mount which has provided me the best results(between -1/1 for x,y axis).
-Calibrate acc and mag,tune PID's, place mag module in correct orientation.
-Place the gps/mag module away from power cables and any radio Tx such as FPV/Telemetry.
-Check CoG and adjust it properly.
-Get a telemetry as you could have seen the errors and landed instead of continuing to fly. Ideally when a failsafe is triggered, you should bring the copter down and check and mitigate the reason for failsafe, instead of continuing to fly.Thats the primary reason that caused your crash.  :(

Learn to read logs, infact you should be really good at it if you want to do SLAM with APM+companion image processor. Some links to start you off.

http://copter.ardupilot.com/wiki/common-downloading-and-analyzing-data-logs-in-mission-planner/
http://copter.ardupilot.com/wiki/common-diagnosing-problems-using-logs/

There are a couple more things which I have overlooked. We can get to them once you sort the primary issues.

EDIT:
One more thing is if you were using stock module with your radio tx, then it shouldnt trigger radio failsafe because, the rx repeats the last received values till the time it regains link, in case of lost link. And this behaviour cannot be changed as far as I know. However powering off the receiver would trigger radio failsafe. How are you powering everything up? Please post pics.

rohan.khairnar

Quote from: SideWinder on March 08, 2015, 01:47:45 PM
Hi

First of all please post a pic of your setup.Now exhaustive logs, specifically the second one with so many fail safes and errors.

1)Your logs show too much vibrations, x/y axis are between -5/5 while z axis seems to be even poor and very bad magnetic interference. So you definitely need to mount your controller properly and place the mag module away from power cables or motors on a mast etc as it is 200% approx in your case while recommended value is below 30%. Also autotune/tune your copter as bad PID's can also be the reason.

2)CG is not centered as two motors(rcout ch1&2) are working harder than the rest of the two which clearly answers your drift issue. I also see some unrequested yawing, which is usually caused by windy conditions or improper yaw  PID's or magnetic interference which seems to be your issue. So place mag module away from possible emi sources and orient it correctly.

3)Your VCC too is going down to 4.7v whenever your are throttling up, so use a ubec etc. Powering via esc's/PM is a bad idea. APM 2.6 needs 5v supply and you need to supply at least 5.3v as a diode drops the voltage by 0.3v I think. It is recommended to use a dedicated ubec etc.

4) Your throttle in & out track really well, and as per log you are getting hover at 45/47% throttle which is decent in my opinion, but if you want to add more payload, you would need to upgrade battery/motors/props etc. Go for a H frame instead of deadcat, as it would provide you some more real estate.

5) Probably need to calibrate your power module as it shows your pack voltage at 11.5 before flight or was it the case? And disable battery failsafe in APM & use a lipo alarm instead for the moment.


Your first logs show you were suddenly loosing gps link for reasons I cannot understand since you were flying in a field. Just check if your GPS/Mag module is properly connected or mounted and also place it away from FPV Tx if any. Wait for atleast 5 mins after powering copter to get a 3d gps lock. Check lights on APM to ascertain it has a gps lock. Altho you were flying in stabilize mode hence GPS failsafe dint trigger alt hold or land as in your case. GPS plotting shows you had crossed St.Louis Convents road but i dont think that would be the case.

Your second log shows a crash due to battery going down, but I see that you had requested very high throttle in the end after you switched to stabilize from RTL which probably lead to brownout of esc's or FC itself. Battery voltage is being shown at 10.3v in the end. In the end I see that you increased the altitude of the copter to 40m and then to decrease it you bring your throttle to ZERO(which should never be the case) along with a radio failsafe which triggers RTL and it clears . This leads to a freefall for your copter which tries to maintain it attitude. Then you see it coming down really fast & give it full throttle which leads to a brownout and crash. I think probably the FC gave up because the FC hasnt recorded the crash . This log also has ekf/dcm failsafe which is triggered "when the GPS implied heading and DCM's estimated heading disagree by at least 60 degrees (configurable with the DCM_CHECK_THRESH parameter) for a full second." which can be caused by a disturbed mag module.

To summarize:

-Mount FC properly on a foam pad etc. Google to find options. I prefer anti vibration mount which has provided me the best results(between -1/1 for x,y axis).
-Calibrate acc and mag,tune PID's, place mag module in correct orientation.
-Place the gps/mag module away from power cables and any radio Tx such as FPV/Telemetry.
-Check CoG and adjust it properly.
-Get a telemetry as you could have seen the errors and landed instead of continuing to fly. Ideally when a failsafe is triggered, you should bring the copter down and check and mitigate the reason for failsafe, instead of continuing to fly.Thats the primary reason that caused your crash.  :(

Learn to read logs, infact you should be really good at it if you want to do SLAM with APM+companion image processor. Some links to start you off.

http://copter.ardupilot.com/wiki/common-downloading-and-analyzing-data-logs-in-mission-planner/
http://copter.ardupilot.com/wiki/common-diagnosing-problems-using-logs/

There are a couple more things which I have overlooked. We can get to them once you sort the primary issues.

EDIT:
One more thing is if you were using stock module with your radio tx, then it shouldnt trigger radio failsafe because, the rx repeats the last received values till the time it regains link, in case of lost link. And this behaviour cannot be changed as far as I know. However powering off the receiver would trigger radio failsafe. How are you powering everything up? Please post pics.

What a theory brother, totally impressed and really really appreciate for all your valuable inputs. I am sure this is not only going to help me but everyone else no matter the problem be same or different. really thanks a lot!
Everything that you have stated makes perfect sense and while reading it i could actually rewind back the time of last flight test. I am sure this will fix all my problems. For the unrequested yawning, im sure its PIDs since it wasnt that windy then. PIDs is something i really need to focus on.

This is how how i am starting, firstly since i have already ordered reptile 500 i will go with it. Will place all modules as mentioned, gps away from any emi, for which i have also bought a gps stand. Also, yes PID tuning. I will be doing autotune first. Those links helped me a lot going throough logs and yes the gps data is false. I never crossed St.Louis Convent road, infact, nowhere close to that.
Will be using anti-vibrations rubber balls(or something) i forgot its name but i have ordered those as well which reduce the vibrations. Also i will be upgrading to FrSky DJT combo but as of now working with stock one.
I am yet to receive the frame, mostly i will get by today evening and yes i will keep posted with pictures.
APM, centered.
All ESCs calibrated, one by one.
Recalibration of accel, compas.
GPS - solid blue signal states gps fix.
Using UBEC to power apm.(yet to order)
Battery replaced to 3S 20C 4300mah.
All the arrangements to be done rel. to CG.
Learning to read and understand logs, downloaded few software tools that make it easier.
Understanding the concepts of different flight mode, still sticking to stabilize mode for a while.
PIDs, which i dont know how but aim to figure out the proper settings.
Neat and clean setup.

Two queries:
1. How to calibrate power module. Did you mean simply resetting it(on/off)?
heres a way i have come accross from a post on rcgroups:
Lipo's are very efficient in that if you pull out 1,234 mah's, your charger ends up putting in almost the same amount of current 1,234 mah. (It probably puts back one or 2 % more but these current monitors aren't that precise anyway)
So... to calibrate within a few %...

1) Plug in a fresh battery and immediately do a flight for 5 minutes or so
2) Immediately unplug the battery
3) Charge the battery and note how many mah's it took to charge. Let's say that was 1,500 mah
4) Download your log and look at maximum CurrTot. Let's say it was 1,300 mah. This implies that your APM Power monitor is calibrated too low.
5) Grab your BATT_AMP_PERVOLT setting from Mission Planner. It defaults to 17
6) Do some math...
NEW_AMPS_PER_VOLT = OLD_AMPS_PER_VOLT * CHARGED_MAH / LOGGED_MAH
With my example numbers, you'd get
NEW_AMPS_PER_VOLT = 17 * 1,500 / 1,300 = 19.61
7) Set the new value in the BATT_AMP_PERVOLT parameter and save to your copter
8) Fly to check


2. I have soldered the receiver antenna as it had once come off after a crash, will that create any problem with the range or break in signal?

With all this done and checking every point that you mentioned i am with highly positive hopes for an improved and successful flight test.
Thanks a ton again bro! really appreciate your effort :) :) :bow: . Will keep posted.

SideWinder

1) Method for power module calibration can be found in detail at
http://copter.ardupilot.com/wiki/common-measuring-battery-voltage-and-current-consumption-with-apm/

2) Yes, a broken antenna can be cause of trouble if it has stripped really bad with missing solder pads, or a bad soldering job. It depends on the extent of damage. Do a range check for your Tx. Method can be found via google for your respective Tx. I suspect something was wrong either with the receiver or the power supply to receiver. Keep an eye on your Tx battery too.

For vibration you need something like this
http://www.quadkopters.com/product/flight-controllers/glass-fiber-flight-controller-anti-vibration-mount/