Re: [amsat-bb] Re: [gamma rays and chips]

["John P. Toscano" <tosca005@xxxxxxxxxx>]

Bob Bruninga (WB4APR) <bruninga@usna.edu> wrote:
>
> Which brings me back to PCsat's TNC's in orbit.  All the Pros said
> you cannot shield things in space with lead and that in some cases
> it can make things worse (by increasing the probabiliy of the above
> efffect).  But then when I investigated replacing our $3 ROM with a
> SPACE HARDENED $3,000 part (with one year lead-time), I asked what
> made it so different.  And the answer was "testing" and it had a
> layer of Tantalum (externally applied) over the chip die.

to which, Daniel Schultz (N8FGV) wrote:
> 
> The entire subject of whether shielding is a good thing or makes
> things worse for electronics protection is open to much research,
> the answer probably varies depending on the particle species and
> energy spectrum of the incident radiation.

and in a different thread, Bob also wrote:
>
>  so many launch opportunities... so little time...
 
Sounds like an interesting opportunity for some research that could
be put together on a low budget without a whole lot of complex
engineering.  Fabricate a PC board with lots of positions into which
one could install a large number of identical memory chips (SRAM,
DRAM, EPROM, EEPROM, Flash, etc.),  Add differing amounts of radiation
shielding.  Fill all the chips with identical data.  Include a simple
microprocessor (with a large memory address space, or else with some
sort of bank-switching to access so many different memory addresses),
and then repetitively scan all the memory components looking for
either transient or permanent bit changes.  Send the results down to
earth via telemetry.

Variations on the theme with little or no change to the design of
the previously-mentioned circuit board:

1) Compare memory cell density instead of thickness of shielding for
   bit flip rate (e.g., does a 64 Kbit EPROM show MORE or LESS
   radiation sensitivity than a 256 KBit EPROM, vs. a 1 MBit EPROM,
   etc.?)  Dan said smaller transistors are better than bigger, but
   it seemed to me that the reverse would have been true -- the same
   size "hit" spread over a larger chip area would lessen the impact
   on the component.  But that's just intuition, which can be way
   wrong.
2) Compare brands or lots of components of otherwise "interchangeable"
   parts (e.g., do 1 MBit EPROMS manufactured by company XXX show more
   or less radiation sensitivity than 1 MBit EPROMS manufactured by
   company YYY?)
3) Compare different shielding materials (Lead, Copper, Tantalum,
   Aluminum, Quartz, Glass, Plexiglas, etc.).  Or stack 2, 3, 4, or
   more of the PC boards and check the error rate on the "outermost"
   board vs. the boards located more inward (do the outer boards act
   as shields for the inner boards, or do they act as a source of
   more secondary particle emissions, in PRACTICE?)
4) If the spacecraft had a means to keep one side facing the earth
   at all times, place these test boards on various faces of the
   satellite to see if the position matters, e.g. does the Earth
   act as a shield or not?
5) etc.

Variations on the theme that would require some additional PC board
engineering:
1) Design a PC board with spaces for numerous identical
   microprocessors instead of numerous different memory
   components.  Have each processor run an identical CPU-intensive
   task, and compare the failure rates as you vary the shielding,
   or the position in the spacecraft, or the clock rate, or the
   manufacturer, or whatever else you want to try varying.
2) Design a PC board with spaces for numerous DIFFERENT micro-
   processors.  Does a Z80 show more or less radiation resistance
   than an 8052 or a 68000 or whatever non-Radiation-hardened
   processor you might want to consider flying on a future space-
   craft.
3) etc.

Just food for thought.
73 de W0JT
----
Via the amsat-bb mailing list at AMSAT.ORG courtesy of AMSAT-NA.
To unsubscribe, send "unsubscribe amsat-bb" to Majordomo@amsat.org