could it be that the lower display clamp spring (a profiled strip made from thin steel sheet with black finish, running all the way across the edge of the display P.C. board) is contacting circuit traces on the top edge of the keyboard P.C.? I realized that this spring is very (VERY!) close to the gold plated traces on the underside of the P.C. There it eventually shorts out signal traces or even the supply traces; the result can be a temporary short of the battery, leaving the battery drained down, and producing strange effects & memory losses.
My HP-41CV once went completely crazy with hieroglyphic characters and producing wheezing sounds out of a sudden, and it turned out to suffer exactly from such short-circuit. A piece of insulating tape on the display clamp perfectly recovered my companion.
Inserting the tape where it's needed is not quite straightforward because the culprit lies underneath the top edge of the keyboard P.C., access is barred by the display's leadwires.
Method 1 (best but labor intensive): remove the keyboard P.C. together with the display, allows to clean the keyboard at the same time. Keyboard P.C. can be remounted with solvent based contact glue once the plastic rivets are clipped off; pressure from the silicone strips helps keeping the board in place. Make sure the glue is only on the rivet positions and does not contaminate the contact strip areas.
Method 2 (easier but beware of static electricity): desolder the leadwires to the display P.C. board, take out the display and insulate the spring clamp.
Method 3 (quick & dirty): insert some plastic pieces (thin wire insulation, e.g.) to force the display P.C. board away from the circuit traces on the keyboard P.C.
For restarting the fixed calculator, short first the supply contacts to let the unit start from scratch in MEMORY LOST condition. I used method 1 and my HP-41CV is now like new. Hope this helps.
P.S. everything in technology has its definite cause, even in HP41s