self-test Message #5 Posted by Eric Smith on 21 Apr 2008, 2:25 a.m., in response to message #4 by Randy Sloyer
Randy wrote:
Quote:
The 20 series units have no self-tests of any kind.
And in fact it was not even possible for a calculator based on the ACT processor to test the ROM(s), as the ACT has no instruction to read an arbitrary word of the ROM as data. None of HP's calculator processors had an instruction like that until CXISA instruction of the the Nut processor used in the HP-41C.
"But wait!" you're probably thinking now. "The 30 series had a self-test!"
Yes, they did, but the way they did it was by adding an instruction that computes the CRC-10 of a 1K block of ROM. The instruction takes 1025 full cycles (57400 clocks) to execute. That was useful for self-test, but had no other benefit.
The 41C doesn't have a built-in self-test, but there were test ROMs for it, which were not offered for sale. The Voyager series (10C/11C/12C/15C/16C) do have a self-test, and use the CXISA instruction in a loop to compute a ROM checksum of the 6Kwords of ROM. (The 15C does that twice for the two ROMs.) The use of the software loop for the ROM test is the main reason that the Voyager self-test takes longer than the Spice self-test.
|