I thought I'd better try the learned code on the projector, but nothing. It
is all now making a bit of sense though. With the projector I was not
covering the IR LED with the other hand. With OFA 23 button push on its own
, isolated, produced no OSD. It now explains why I could only preset with
code 1004 and advance 23 steps as somewhere in between was a pre-code . By
elimination , by selective covering of the LED , it is step 6.
So programmed in 2 codes in the learner, "6" and "23". It now works with the
projector bringing up the Keystone + on OSD.
Codes are not sequential in the OFA , preset to 1004 and +6 pushes reads out
as code number 631 and 1004 + 23 reads out as 1928

So I thought set with learned "6" and run through all the OFA codes. If
there is a time-out reset between 6 and 23 it is greater than 2 minutes, as
I timed it . In TV and SAT codes something resets the 6 , in less than 2
minutes so something active in the wrong sense , but did not ellicit any
other OSD responses.

If the "23" preceeded the "6" in the OFA listing then I would not have seen
the keystone response , although (not specifically stated in the manual) you
can multi-step down as well as up.
Will set up scope and DSO tomorrow. I will also try learned 631 code and OFA
going down all its code banks to see if it hits something active before
those lurking reset codes.
Lools like a more general hack routine is to use 2 URC , preferably the same
model, for these 2 stage coded systems and automate the button pushes
electronically