Home Random Page


CATEGORIES:

BiologyChemistryConstructionCultureEcologyEconomyElectronicsFinanceGeographyHistoryInformaticsLawMathematicsMechanicsMedicineOtherPedagogyPhilosophyPhysicsPolicyPsychologySociologySportTourism






Periodic Flash Rebuild

To employ this mitigation method, the user must periodically place the star tracker into maintenance mode and send Rewrite NAND commands. These commands should rewrite those areas of flash that have been subject to stress. They include:

· The boot block. Instead of the Rewrite NAND command, the Make Bootable command should be used.

· The flash relocation and backup tables

· The application image

· The calibration table

· The entire star table and hash table

· The triangle table

Of these, the triangle table is probably stressed less than the others. It may be possible to rewrite it less often than the other areas.

The rewrite frequency must be carefully chosen. The program/erase endurance of the NAND flash is only 100,000 cycles. For a 10 year mission, with no margin, the maximum rewrite frequency would be once an hour. Rewriting only once every few hours will give margin.

The current health of the NAND flash can be estimated by reading the flash error counts at the beginning of the hardware telemetry. By observing these trends, and appropriate rewrite frequency can be determined.

Where possible, perform the rewrite operation in a low radiation portion of the orbit. If in LEO, avoid the South Atlantic Anomaly (SAA). Flash memory is more vulnerable to latchup while its charge pumps are active, performing program and erase functions.



Run From RAM

To employ this mitigation method, set the “Cache triangle table” bit in the functional GO bitfield in the control structure. Then, when sending GO or COMBINATION commands, ensure that bits 2 and 5 are set.

When the first GO or COMBINATION command is received, the star tracker will take slightly longer than normal to execute. In addition to the star and hash tables, the entire triangle table will be copied into RAM. The functional processor will remain turned on, and all accesses will be to RAM. There will be no repeated flash accesses, and so the flash will not accumulate read disturbance errors.



Comparison

Neither mitigation strategy is ideal. The periodic flash rebuild requires that the star tracker be taken out of operation for a number of seconds, several times a day. This may interfere with missions that demand very high ACS availability. It also consumes program/erase cycles – a finite resource. There should be sufficient cycles for any normal mission, but nevertheless it is disquieting.

Running from RAM defeats one of the prime design features of the star tracker – that the functional processor powers down and reboots between cycles. Expect to see periodic SEU events where the functional processor becomes corrupted and requires a full reset (GO code of 0). The spacecraft flight computer will need to detect these events autonomously to guarantee high ACS availability. SEL events might occur as well, and by keeping the functional processor powered continually the chance of a destructive event may increase.



Recommendation

All else being equal, Sinclair Interplanetary recommends the periodic flash rebuild mitigation method. Keep an eye on the flash error counts. Maybe your star tracker will not experience read disturbance degradation in your application, and no mitigation will be required.


Date: 2015-12-17; view: 657


<== previous page | next page ==>
Flash Memory Issues | Turner Bison System
doclecture.net - lectures - 2014-2024 year. Copyright infringement or personal data (0.008 sec.)