| Comment |
This is a pretty neat unit! At a first glance
the empty keyboard suggest that there's not much functionality. But
there's a surprising number of features hidden in various menus. And
best of all is of course the support for both RPN and algebraic entry
mode. (The only other model with this capability is the HP-49G.) In RPN mode the arrow
up/down buttons scroll the stack - very neat.
The only drawback I can see is the permanent "waste" of one display
line for menu choices.
The main difference between the two version is that the HP-17BII
supports both algebraic and RPN entry mode. To support RPN it features
stack roll down and X-Y register swap keys. Furthermore, there is a RND
key (which rounds to the selected number of display digits) and under
the MODES menu an option for double payments. Other than that they seem
to be identical.
Unfortunately, I don't have a manual for the HP-17B and it is not
included on the MoHPC CD-ROMs.
From Mike at CSUN about the HP-17B:
Has anyone else played with machine code
programming on the HP-17B, 27S, or 42S? I seem to have acquired a 17BII
accidentally, and have been exploring its hardware. This machine has a
memory editor just like the 48 does (so do the 17,27, and 42.) You get
in the same way: ON-4th menu key, then <- to enter editor. Press .
immediately after entering to show your version number. Keys are: ^v X
divide + - move around by varying amounts, 0-9 and menu keys enter hex,
PRT sends line to IR (potentially to 48 running INPRT, then to PC for
disassembly), and . executes at the current address. Note that execute
is a call, doing a RTN goes back to the editor from your ML code. Basic
architecture is 64KB (128K nybbles!) of ROM at
00000, display at 40000, and 8K of RAM at 50000. The first part of RAM
changes
on the fly, so around 51000 or 52000 is good for experim- ental code.
One
can do small machine language programs simply by writing the code on
the
PC, using SASM on it, and typing in the hex digits in the .L file. If
one
creates a Solver equation with a recognizable pattern, then scans from
54000
(end of RAM) back- wards, one can easily find it. This is a way to
enter
ML without wrecking normal memory. Has anyone played with this? Where
are
the hardware registers? How well could one integrate user-written code
with
the operating systems on these machines? They are RPL machines
internally;
could you write system RPL as well as machine code?
Bug reports on the HP-17BII:
- From Wlodek Mier-Jedrzejowicz: There is an oddity in
the RPN stack of this (and the 19bii) in that the stack ranges in size
from 1 to 4 entries, depending upon how much data has been entered. For
example, if you do:
Switch to RPN mode, 1, Enter, 2, Rv, Rv
You will see "2" in the display, not zero.
- From John H. Meyers:
Early models of the HP17BII have a bug which is activated when the
HP17BII is
in RPN mode and the [+/-] and [x<>y] keys are pressed one after
the
other. A harmless way to see the bug is as follows:
- Put a positive number on the stack, for example
press 9 and [INPUT].
- Change the sign by pressing the [+/-] key next to
[INPUT].
- Exchange x and y by pressing the key with
x<>y above it.
- Now type a digit, for example 8; you will see a
minus in front of it. (This is the bug; after +/- to change a positive
number
to negative, immediately followed by x<>y and a number, the
number
has a spurious minus sign placed in front of it.)
- Press [INPUT] and you will see the number is
really positive.
- Type another number; it will again have a minus
sign.
- To deactivate the bug safely press the [CLR] key.
The bug is dangerous, if you get to step 4 and
see a number which is negative though it should not be, you might be
tempted
to press the backarrow key to remove the number and the minus sign.
Since
the minus sign is not really there, if you try to delete it, you
confuse
the calculator - it stops for a moment, then clears the stack and
displays
the MACHINE RESET message. The bug can also lock up your keyboard - in
that
case you should reset the calculator by pressing the [CLR] key and the
third
key from the left in the top row both at the same time. If you ever see
the
bug, press [CLR] at once to kill it!
There are some variations on the bug. If you type a number, immediately
change its sign, then press x<>y, the calculator will beep and
refuse to swap x and y. If you press the rolldown key next to
x<>y, that will beep and refuse to work too. If you press the
backarrow key until only the minus sign is in the display, then press
[INPUT] or try to use this minus sign, nothing will work except that
the [-] key will put a second minus in the display. If you do steps 1
through 4, then press [INPUT], [+/-], x<>y, and a digit key
again, you will see two minus signs, and you can repeat
this process adding one more minus sign each time. If you press [CLR]
now,
then type a number, press [+/-], [INPUT], x<>y, [+/-], x<>y
and then a digit key, you will see a zero (if the display mode is ALL)
or
a fraction mark (if the display mode is FIX) in front of the digit
instead
of a minus.
It can also appear whenever there is a negative number already on the
bottom of the stack when CALC is pressed in the Solver, any typing of a
new value also generates strange stuff in the display, and the
backspace key not only causes a warmstart, but has given me a "memory
lost" as well.
The same bug exists in early HP19BII
models, though it behaves a little differently. It was removed early in
1991.
|
