What was the JOHNNIAC ANelex printer text encoding?

[sebras]

The JOHNNIAC Programmers' Manual on page 42 mentions an ANelex printer:
http://www.bitsavers.org/pdf/rand/johnniac/D-3058_Johnniac_Programmers_Manual_Jul55.pdf

The printer is visible to the the left in this picture:
https://images.computerhistory.org/revonline/images/102618693-03-01.jpg?w=600

It matches none of the Anelex printer manual scans on this page exactly, but Anelex Series 4-1000 looks like the most similar:
http://www.chilton-computing.org.uk/acl/literature/chapman/p002.htm

The Instruction Manual for the Anelex Series 5 Printer System has been scanned:
http://museum.informatik.uni-kl.de/Rechner/Zuse/Allgemeine%20Peripheriegeraete/Schnelldrucker/Anelex-5/Dokumente/Instruction%20Manual%20for%20the%20ANELEX%20Series%205%20Printer%20System.pdf
As well as the scans of the schematic of the Series 5 Printer:
http://museum.informatik.uni-kl.de/Rechner/Zuse/Allgemeine%20Peripheriegeraete/Schnelldrucker/Anelex-5/Dokumente/000%20Originale/Instruction%20Manual%20for%20the%20ANELEX%20Series%205%20Printer%20System/
Oage 8 of the PDF describes the print roll character codes. Perhaps this describes the encoding sent by JOHNNIAC?

The documentation for a completely different machine also mentions an Anelex Printer and on page 11 is what looks like a character table:
https://bitsavers.informatik.uni-stuttgart.de/pdf/ict_icl/atlas/Atlas_Machine_Description_Sections_8-13_Nov64.pdf

[drboone]

I see that there were two printers, the 40 column one, and the 120 column one. I think the Series 5 is too new to be either.

The ANelex brochures and such seem to indicate that the font drums could be customized to some extent, or at least that there were a variety of them available. In this era, I'd guess that they were 5-bit devices, but there were a lot of 5-bit codes, and in this case something more or less custom may have been used.

I haven't seen much (any?) doco about how the machine was initially started. Anyone? My guess is that a short series of 10.x orders were entered from the console to initially load a card deck (or perhaps something stored permanently on the drum).

[sebras]

I see that there were two printers, the 40 column one, and the 120 column one. I think the Series 5 is too new to be either.

Another way of approaching this is to study the arguments to the 101 opcodes. I'm thinking that some of those might point to strings that may be used to infer the encoding.