Sun 3/50 memory layout

agw at math.canterbury.ac.nz agw at math.canterbury.ac.nz
Sun Dec 9 09:07:00 AEST 1990


We have a couple of sun 3/50's which fail to boot.  Running their
diagnostics reveals the problem is a parity error and gives the physical
address (and bit error in the associated long word) at which this occurs.
Our hardware manuals don't tell me how to translate this information into
which memory chip is faulty. The memory is arranged in 8 rows of 18 chips
with a numbering arrangement ( underside of board )

	U1420 U1400 U1320 U1300 U1220 U1200 U1120 U1100
	U1421 U1401 U1321 U1301 U1221 U1201 U1121 U1101
	U1422 U1402 U1322 U1302 U1222 U1202 U1122 U1102
	  .     .     .     .     .     .     .     .	
	  .     .     .     .     .     .     .     .	
	  .     .     .     .     .     .     .     .	
	U1428 U1408 U1328 U1308 U1228 U1208 U1128 U1108
	U1430 U1410 U1330 U1310 U1230 U1210 U1130 U1110
	  .     .     .     .     .     .     .     .	
	  .     .     .     .     .     .     .     .	
	  .     .     .     .     .     .     .     .	
	U1438 U1418 U1338 U1318 U1238 U1218 U1138 U1118

suggesting

o the first digit is bit # in the byte (which is the parity bit - 0 or 8 ?)
o the second is the byte of the (long?) word
o the third is the Mbyte of memory the (long?) word is in.

The chips are soldered in so any information as to the exact relationship
between physical memory address and chip number would be appreciated. The
local Sun agent seems unwilling/unable(?) to supply this information.

			Allen Witt,
			email:  agw at math.canterbury.ac.nz



More information about the Comp.sys.sun mailing list