During the May 24, 2010 flight we lost data from the OIP and CDP. Before the start of this flight, the data system locked up and had to be rebooted. This resulted in serial box and instrument running for 5-10 minutes without the M300 running. This resulted in large > 65000 bytes in the OIP (Type 39) data buffer.
On June 8, 2010 the following trouble shooting was conducted.
Started M300 and record data to file 20100608143939.sea OIP started off 14:40:50 OIP On 14:54:50 M300 Off. This resulted in all good OIP buffers.
Started M300 and OIP normally, with power on both M300 and OIP at the same time. 14:45:25 - M300 started, record data to 20100608144523.sea. 14:51:23 - M300 stop. This resulted in all good data, first buffer size was 1219.
The OIP was not turn off, reset M300. 14:58:20 M300 started, recording to 20100608145755.sea 15:00:42 M300 Off This resulted in some bad (invalid) data buffer sizes (< 1219)
Left OIP turned on, changed acq.300 table so all offset were set to zero. Did not reboot the m300. 15:35:15 M300 on and recording data to 20100608153507.sea 15:38:25 M300 Off This resulted in bad buffer size that were greater than 1219, buffer sizes like on the May 24, 2010 flight. This reproduced the problem.
Restarted the M300, recorded some data to 20100608154011.sea.
Updated the NT 960 Drivers per email from Augie. Had qnx4.F file but not ntpci003.f file. The qnx4.F had different (smaller) file size. Saved old version of the files. The driver version was different since the readme file was different. Installed the updated drivers.
Boot M300 and record data to 20100608162940.sea file. 16:30:17 M300 On 16:30:40 OIP On 16:32:77 M300 off This resulted in all valid data from the OIP.
Reboot the m300, OIP was left on 16:47:48 M300 started, recording data to 201006081644.sea 16:49:48 M300 Off This resulted in all good data, first buffer size 1219, second 1086, and third 1118.
After Lunch, left OIP run. 18:48:23 M300 started, recorded to 2010608104820.sea file. 18:57:10 15 um Bead test for CDP. 19:10:00 - 30 um bead test for CDP. This resulted in all good data buffer sizes.
After Aircraft tour, left OIP run for 5-10 minutes. 20:25:40 Started M300, recorded data to 20100608202456.sea. This resulted in all good data buffer size and reproduced again the situation that had resulted in the problem seen on May 24, 2010 flight. Since all data buffer sizes were correct, the upgrade of the driver fixed the problem.
Problem with the NT 960 QNX driver.
Upgraded the NT 960 QNX driver with those supplied by Augie.