Gregorian amp failures feb07
20nov06
plots:
The
position and status bits of the system during the failure (40 seconds)
(.ps) (.pdf).
The
status bits of the system during the failure (25 minutes) (.ps)
(.pdf).
Gregorian
motor torques during the failure (.ps) (.pdf)
Intro: (top)
Amplified 21 and 22 on the dome have failed with amplifier
not ready. This has occurred on :
-
13feb07 - amp 22
-
15feb07 - amp 22
-
16feb07 - amp21 and amp 22.. Twice once at 3 am. A second time at 6 am.
The failures have occurred when the dome is at high
za and starting to move downhill. The failure occurs 3 or 4 seconds after
it starts moving downhill. To recover from the error, the operator has
sent a reset command to the vertex system.
Position, status bits during failure (40 secs of data).
(top)
The
position and status bits of the system during the failure (.ps)
(.pdf).
There are 4 pages for each plot showing the four separate failures. Data
is recorded once a second. For the status plots the leftmost * on each
line is the zero (false) value for each bit. Each plot contains 40
seconds of data.
-
Plots 1-4. Az, Za position. The top plot is the azimuth position.
The bottom plot is the dome za. The dashed vertical line is when the failure
occurred. Each * is spaced by 1 second. Each failure occurred 3 to 4 seconds
after the dome started down hill. Some days it had been moving uphill and
then started down (15feb, 16feb_6am). Other days it had been sitting at
the high za and then started down (13feb, 16feb_3am).
-
Plots 5-8. General system status: This includes the power status.
There were no power problems during the time period.
-
Plots 9-12. Dome axis status: There is a servo fail with the drive
going from enable to disable at each failure. At the same time the brakes
are go from released to brakes on. The brake fail comes on 1 or two seconds
after the original event (i'm not sure whether this is meaningful or not).
-
Plots 13-15. Tracking status and Dome power status: The top plot
shows the system going from tracking to stopped. The bottom plot shows
that:
-
DCBusFlt: There are no failures on the dc bus. This is the output
of the power supplies that drive the amplifiers.
-
MnContOn: The master contactors remain closed providing power to the power
supplies.
-
PwrSFlt: there are no power supply faults.
-
Plots 16-20. Dome amplifier status: These 4 pages show each
amplifier read and fault status. The first two faults show amplifier 22
going not ready. The last two days show amplifier 21 and 22 going not ready.
There are no amplifier faults occurring.
Status bits during failure (25 minutes of data).
(top)
The 2nd set of plots shows the status bits for 25 minutes of time
starting before the failure.
The
status bits of the system during the failure (25 minutes) (.ps)
(.pdf).
-
page 1-4 Dome axis status.
-
page 5-8 tracking status, Dome power supply status
-
Plots 9-15. Dome amplifier status.
These plots show the same data as the first set but
extended to about 25 minutes. The only interesting behavior is that some
of the other amplifiers seem to be going not ready a while after the failures.
I looked at the normal
behaviour of the status bits when a reset/stop sequence was issued.
It did not change the amplifier ready status. The software must be in a
different state when the first error occurs.
Motor torques during failure. (top)
Gregorian
motor torques during the failure (.ps) (.pdf).
These plots show the motor torques for each motor during the failures.
The torques are measured once a second.
-
Page 1. Torques during failure. The 4 plots show 40 seconds of data
around each failure. The dashed lines shows when the failure occurred.
Each * is spaced by 1 second.
-
Page 2. Torques when no failure occurred. On 16feb07 i tried to
duplicate the failure motion to see if i could generate an error. The top
plot shows the dome za motion. The bottom plot shows the measured torques.
The dashed vertical line shows 3 seconds after the motion where the previous
errors tended to occur. This motion did not generate an error. I repeated
this about a dozen time without an error occurring.
The motor torques when the error occurred do not differ
appreciably from when the error did occur. The torques are complicated
since:
-
When sitting at the high za and not moving, the motors are holding the
dome. This takes lots of torque.
-
The hydraulic brake begins to resist the dome motion when the velocity
increases. This causes the motor torques to decrease as the velocity
increases.
The torques for motor 31 (which never failed) seem to
differ from the others (when there was a failure and when there was no
failure). This may be a calibration problem or some other problem.
Summary (top)
-
There have been 4 amplifier not ready failures on the dome.
-
Each failure has occurred 3 to 4 seconds after starting downhill at high
za.
-
The first 2 failures had amplifier 22 going not ready by itself.
-
The last 2 failures and amplifier 21 and 22 both going not ready together.
-
There have been no power or power supply faults during these failures.
-
There have been no amplifier faults during these failures. This is a bit
strange. If there was an amplifier error (over speed, fold back, etc.)
you would expect the amplifier fault bit to come on.
-
An amplifier not ready will occur if there is an amplifier fault. This
bit is not coming on. The amplifier not ready can also be commanded to
the not ready state by the PLC by bringing the commanded enable request
bit to 0.
-
The torques do not look very different when there is and is not a failure.
-
The torque for motor 31 looks different than the other 7.
Processing: x101/070213/agcampfail.pro
<-
page up
home_~phil