In this teardown it would seem that the eprom be two am29F010 flash.
https://www.eevblog.com/forum/reviews/hp-33120a-teardown/
In your 33120A is instead only a 27C020?
it is possible that welding the memory back in the working instrument you've shorted a few pin or pin/track?
Good luck.
I think that the chip and functionality is mistakenly identified in those teardown photos. there is a sticker over the firmware chip so I bet it also has 29F010 for user setttings(arb memory) and 27C020 OTP ROM for firmware.
Things get more weird by day: now the borrowed unit (dead after firmware reading attempt) started working...Fix? let it sit on the table over night and it started working
Was the memory too hot when we tried it first time? No. it had plenty of time to cool and we tried countless times with no success...
I don't think I made any shorts or dry joints, usually PLCC-32 is reasonably easy piece of cake for me.