Update README.md

Added documentation for statements about SUNDIAL assembly versioning.

SundialE/README.md

@@ -16,4 +16,7 @@ On the available information -- and admitting that there may be AGC units not co
 
 The discrepancy is that the fixed-memory modules listed are known to be from SUNDIAL assembly 2021104-051.  However, both the 2021104-041 and -51 modules contain SUNDIAL E software, so if the modules were exchanged at some point it wouldn't have affected functionality in any way.
 
+By the way, the documentation on which the assertions about fixed-memory modules 2021104-041 vs 2021104-051 are based is this:  All of these dash numbers are past the dash-number range for which we have obtained explicit [engineering drawings (namely, 2021104-011 and -021)](https://archive.org/details/apertureCardBox467Part2NARASW_images/page/n16/mode/1up?view=theater), although the drawings we have _do_ identify 2021104-x as a SUNDIAL assembly. However, [document ND-1021043](https://www.ibiblio.org/apollo/Documents/HSI-208435-002.pdf#page=19) has "compatibility" tables showing what module part numbers are in which positions for 2021104-011 through 2021104-061.  From the tables, it is clear that the combination 2003053-121/2003053-151/2003972-211 is indeed one of the three possible combinations acceptable for 2021104-051.  The compatibility table also states that SUNDIAL assemblies 2021104-041, 2021104-051, and 2021104-061 are compatible replacements for each other under all circumstances (and that -031 is compatible as well, for testing purposes, but is "not as good").
+
+But why SUNDIAL E, as opposed to SUNDIAL A through D?  I haven't found any explicit documentation, but the indirect evidence is that in the afore-mentioned compatibility table, it is noted that 2021104-021 "contains gyro compassing error", whereas -031 through -061 presumably do not.  If we look at [document E-1142 Rev 50, "System Status Report"](https://www.ibiblio.org/apollo/Documents/E1142-50.pdf#page=27), it is noted that "Verification of fix to Sundial C and D gyrocompass program incorporation into Sundial E was completed." From this combination of statements, it seems that we would have to infer that SUNDIAL assemblies 2021104-031 through -061 each contained SUNDIAL E, whereas 2021104-021 likely contained SUNDIAL C or D.