Here we go again...
3 views
Skip to first unread message
Marc-X
Nov 9, 2010, 3:34:42 PM11/9/10
to UAVHeliBoard
Guys,
As earlier reported, I have often experienced "strange" behavior when
cold starting the UDB. I have seen a clear coincidence between this
behavior and the connection of a XBee to the debug UART. The troubles
reported has been servo jerkiness, servo drift, extreme "slow motion"
of the gyro reaction, etc.
Now, after that I have used MP in HILSIM for a while, I have severe
problems with my UDB. After first successful compilation and flashing,
I got no visible reaction on the servos in STABILIZED mode. After
resetting, rebooting, power down the whole system for a few minutes
and a couple of other magic tricks, I now have a UDB which seems to
react normally in the ROLL axis. The pitch axis is going in "slow
motion". The behavior is visible both as servo motion, and in the
debug data stream which scrolls on a laptop beside me.
I have camera stabilization enabled, but the camera doesn't move.
(This means nothing since I haven't verified that my camera config
works before).
I feel this smells hardware again, but I will do a few more "magic
tricks and see if I can get it to "loosen up" and behave as it should.
Just thought you should know that the problem reported earlier is in
no way gone. I have not used the IMU in a while (since HILSIM feeds
simulated IMU data to the UDB through the GPS UART. This makes me
think that my IMU unit is faulty, or possibly there is a bug in the
gyro calibration / feedback routine).
Cheers!
Marc
As earlier reported, I have often experienced "strange" behavior when
cold starting the UDB. I have seen a clear coincidence between this
behavior and the connection of a XBee to the debug UART. The troubles
reported has been servo jerkiness, servo drift, extreme "slow motion"
of the gyro reaction, etc.
Now, after that I have used MP in HILSIM for a while, I have severe
problems with my UDB. After first successful compilation and flashing,
I got no visible reaction on the servos in STABILIZED mode. After
resetting, rebooting, power down the whole system for a few minutes
and a couple of other magic tricks, I now have a UDB which seems to
react normally in the ROLL axis. The pitch axis is going in "slow
motion". The behavior is visible both as servo motion, and in the
debug data stream which scrolls on a laptop beside me.
I have camera stabilization enabled, but the camera doesn't move.
(This means nothing since I haven't verified that my camera config
works before).
I feel this smells hardware again, but I will do a few more "magic
tricks and see if I can get it to "loosen up" and behave as it should.
Just thought you should know that the problem reported earlier is in
no way gone. I have not used the IMU in a while (since HILSIM feeds
simulated IMU data to the UDB through the GPS UART. This makes me
think that my IMU unit is faulty, or possibly there is a bug in the
gyro calibration / feedback routine).
Cheers!
Marc
Marc-X
Nov 10, 2010, 11:57:18 AM11/10/10
to UAVHeliBoard
No matter what I did, I couldn't get the UDB to behave again.
Last time I had this issue "fixed" (then I had re-heated the
solderings on the daughter card), I had the board almost completely
covered in "hot-glue". This was to protect it, and cover some bare
spots that came when I used a scalpel to scrape of the solder mask
for cable attachments. This hot-glue is a real H*LL to get rid of if
you have to do repairs later. It really protects the board well, but
hey... Anyway; I went over the board once again without anything
making the UDB happy.
Finally I used my stereo "microscope", which really is a 20X stereo
loupe, but this does really show you how the world look like. When
viewing the gyros through the loupe it became obvious that there were
at least one soldering that was NOT ok. You could see that there were
space between the chips "pin" and the board. I tried to solder this
point several times, which is under a 10:th of a mm in diameter, with
varying results....
Now the board "works" again. I had all the axis respond at a normal
rate again, but I don't trust this board anymore. It will probably
never fly anything else than a "drone" in a simulator. I really had
high hopes for this and my helicopter, but after several real life
"issues" and one expensive crash with this board, I can't afford to
take any chances. At least not as the board is now. If you know of any
way of getting a new daughter card, or I get hold of some EXTREMELY
fine soldering tips it might be a future for me and this board,
otherwise it's doomed to the eternal SIM-WORLD.
A tired and pretty fed up Marc...
Last time I had this issue "fixed" (then I had re-heated the
solderings on the daughter card), I had the board almost completely
covered in "hot-glue". This was to protect it, and cover some bare
spots that came when I used a scalpel to scrape of the solder mask
for cable attachments. This hot-glue is a real H*LL to get rid of if
you have to do repairs later. It really protects the board well, but
hey... Anyway; I went over the board once again without anything
making the UDB happy.
Finally I used my stereo "microscope", which really is a 20X stereo
loupe, but this does really show you how the world look like. When
viewing the gyros through the loupe it became obvious that there were
at least one soldering that was NOT ok. You could see that there were
space between the chips "pin" and the board. I tried to solder this
point several times, which is under a 10:th of a mm in diameter, with
varying results....
Now the board "works" again. I had all the axis respond at a normal
rate again, but I don't trust this board anymore. It will probably
never fly anything else than a "drone" in a simulator. I really had
high hopes for this and my helicopter, but after several real life
"issues" and one expensive crash with this board, I can't afford to
take any chances. At least not as the board is now. If you know of any
way of getting a new daughter card, or I get hold of some EXTREMELY
fine soldering tips it might be a future for me and this board,
otherwise it's doomed to the eternal SIM-WORLD.
A tired and pretty fed up Marc...
Riccardo Kuebler
Nov 10, 2010, 4:38:37 PM11/10/10
to uavhel...@googlegroups.com
Marc,
is this a new UDB3 board ?
If yes, please tell that to SparkFun.
Best regards,
Reply all
Reply to author
Forward
0 new messages