03feb05 td12 trouble
03feb05
On 03feb05 around 8am,
td 12 began making a loud noise on the speakers in the control room. It
sounded like hard rain (but it was not raining). After a few moments the
motor went into foldback. A little later we stopped the tiedowns. Something
similar occurred to td12 on 31jan05 at 23:20 pm. At that time it also went
into foldback.
-
fold back is when the amplifier is requested to provide its peak current
for more than a few seconds. If this occurs then the amplifier switches
to fold back. In the fold back mode the amplifier will only supply a fraction
of its peak current (although it keeps running). This mode is to protect
the amplifier from overheating..
Plots were made of some of the measured values
of the 3 tiedowns covering 8:27 to 8:33 on 03feb05. The tiedowns are color
coded so that td12 is black, td4 is red, and td8 is green. The tiedowns
were turned off at 8:30 and turned back on at 8:31:38.
The plots
during the problem show:
-
Top plot: This is the commanded velocity that is computed by the
little star computer and then sent to the d/a converter. From there
it goes to the amplifier. It is in d/a counts. The bump in the blue
line is where td12 went into fold back. Around 8.475 td 12 velocity started
to change (while the other two remained 0).
-
2nd plot: The commanded velocity is output by the d/a converter.
On it's way to the amplifier, this voltage is read back by the a/d converter.
It is plotted here in the second plot (the scale is different than the
first plot). The velocity read back is not the same as the commanded velocity.
This can be because the d/a, signal conditioning board, or a/d board is
misbehaving. Rfi can also cause this readout to jump. The values
read back continue to jump even after the tiedowns have been turned off
(between the dashed red lines).
-
3rd plot: This is the motor torque as measured by the current provided
by the amplifier. The torque is also jumping around. When the td's are
turned off the torque goes to 0 as it should.
-
4th plot: This plot is the motor speed as measured by the
amplifier commutation read back (the units are linear a/d counts). This
also shows that the motor motion is erratic.
5th plot: The encoder position vs time (the mean value has been
removed). The encoder also sees td12 moving erratically during this period.
Td 12 is definitely having problems. If the commanded
velocity read back is valid, then there must be something wrong with the
d/a, signal conditioning board , or the little star. Unfortunately this
signal is often corrupted by rfi. The little start is probably ok,
since the other values seem reasonable (for the requested read back velocity).
If the commanded velocity to the amp is really what we see in the commanded
read back, then the PI loop will cause the other parameters to change as
they do.
When the tiedowns are turned off, all of the signals
except the command velocity read back go to zero (as they should). So my
first guess is to replace the dac and signal conditioning board.
This problem has occurred twice (that we know of
) in two days so it is intermittent. We may have to way a while to see
if the changes really have fixed things.
processing: x101/050201/tdprob03feb.pro
home_~phil