Dear INMS Science Team,
Janet polled us for information earlier today....or was it yesterday....my days and nights are flowing together so as to become indistinguishable. Anyway, Janet asked for info on Saturday. I'm sorry that we haven't put more information on the web site, but we have been in overdrive here trying to analyze the earlier anomaly and prepare for other activities on the spacecraft (including the Quiet Test [an instrument interference test], our contingency test [a second attempt at executing an INMS baseline], and the RADAR/INMS Mode Test). I am sending this email instead since my web maintainers are taking the weekend off.
You should have received a report earlier in the week about our anomalies on Day Of Year 005 (DOY 005), during the execution of the INMS baseline test. If you are curious for details, here is an example of the translation problem. If not curious for details, continue reading at ********* below. The Spacecraft Activity Sequence File (SASF) format (the format required and documented by JPL) of an INMS command that was incorrectly translated is:
74GO_MEAS_TABLE(0,"CYCLE",8,0,[0x0004,0x0001,0x0102,0x0f02,0x07fc]
In the ATOL language, the language used in the lab at GSFC to communicate to the EM, the command that corresponded to the above command was:
tc 74go_meas_table 7,0,2,8,0,4,#1,#0102,#0f02,#07fc
The 7 and the 4 in the above line are not valid parameters to the instrument, rather they are only parameters that communicate something to the ATOL language interpretor. As you see the 7 was correctly ignored when creating the SASF version, but the 4 was not. The confusion came about because there is no documentation on the ATOL format of the INMS commands, and also because there are other versions of the ATOL command formats that also map to the unique SASF 74GO_MEAS_TABLE command format. For example, another ATOL version of the same command listed above is:
tc 74go_table_4 7,0,2,8,0,#1,#0102,#0f02,#07fc
(notice the absence of the 4 in the parameter list). These two ATOL format types are interspersed throughout the INMS baseline test.
Unfortunately an engineer from GSFC, an analyst from SwRI, and I all
reviewed the SASF built sequence and did not catch the translation error.
**********
Now what effect did this translation error cause? First of all it corrupted some of our onboard operating tables. That, in turn, caused us to execute with settings on the instrument that we did not intend. One of the most disturbing of these erroneous settings caused us to cycle Filaments 1 and 2 (open source filaments), especially Filament 1, on and off many times.
How have we and do we proceed in the remainder of the Instrument Checkout
(ICO) period? First of all, we have developed real-time plotting software
since DOY 005 so that there is much more immediate visibility into
unexpected values of key monitoring values from INMS. (One of the
mistakes of the team on DOY 005 was to leave the instrument on longer than
necessary because we did not have good visibility into all anomalies that
were occurring due to the corrupted tables.) Secondly, prior to the Quiet
Test (instrument interference test) on DOY 010, we reloaded our operating
tables and confirmed that they were all as expected, in order to fix any
corrupted tables. We are currently participating in the Quiet Test.
The instrument is performing well, although we only use Filament 3 (closed
source filament) during this test. At first glance, I don't seeany
significant interference caused by other instruments, but this will have
to be further investigated with our post processing tools.
On Monday, we will put the final touches on the corrected SASF baseline test to be run during our contingency period on DOY 013. To double-check this sequence, a GSFC engineer and I will go through the commands step-by-step together. Also, one of our SwRI programmers has developed a basic tool to automatically check the SASF baseline sequence for the type of error that was introduced in the sequence run on DOY 005. We will use that tool to check the SASF contingency sequence, as well.
We are continuing to analyze the effects that the table corruption error had on the instrument, and we will keep you posted as more info becomes available.
Regards,
Dana Burket
INMS and CAPS OTL
Return to INMS homepage
[ INMS ] [ Related Links ]
e-mail
the webkeeper