Viewing the full mantissa on a 41C

[Marcus von Cube, Germany]

Hello to all!

On most HP RPN calculators there is a function "clear prefix" which duplicates as a means to view the full mantissa of the value in X while held down.

I couldn't find anything similar on my HP 41CV. My fault or an omission by HP?

[Vassilis Prevelakis]

Marcus von Cube wrote:

> On most HP RPN calculators there is a function "clear prefix"

> which duplicates as a means to view the full mantissa of the

> value in X while held down.

>

> I couldn't find anything similar on my HP 41CV. 

I think the 41 was caught in the midst of a transition from the era where a separate command (DSP) was used to specify the precision to the new way of integrating it with the mode command (FIX/SCI/ENG).

E.g. on the HP-67/97 FIX/SCI/ENG do not take an argument, while on the HP-41 and later models they take the number of digits to display.

On the Spice series (introduced in the same timeframe as the 41) the small display (in the sense that they could display fewer digits than previous models because of the larger digits) the use of the MANT command was imperative.

**vp

[Thomas Okken]

...or to look at it differently, the MANT command became necessary with the Woodstocks, since unlike the Classics before them, they could not display 10 digits in SCI mode.

But, it wasn't until the Spice series that the MANT command actually appeared -- and that's after the introduction of the 41C.

I don't have a 41 any more, but I seem to remember that, in SCI 9 mode, ARCL would recall the full 10-digit mantissa to the alpha register. If that's true, you could use that to create a pseudo-MANT command.

[Andrés C. Rodríguez (Argentina)]

The decision to implement separate DSP and FIX/SCI/ENG instructions has to do with op-code space preservation. The HP 67 (as the HP 25 and others) used just one byte to encode each instruction, hence there are only 256 op-codes available. With a rather rich set of instructions, you need to be careful in order to avoid running out of possible op-codes. The separate option uses just 13 op-codes for all display formats (FIX, SCI, ENG, DSP 0, DSP 1, ... DSP 9), while the unified style requires 30 op-codes (FIX 0, FIX 1, ... FIX 9; SCI 0, ... SCI 9; ENG 0, ... ENG 9), too costly a sacrifice for the HP 67.

As the HP 25 had a not-so-rich instruction set, it could spend more op-code space without too much concern.

The HP 41 introduced multi-byte instructions, so the restriction was avoided.

Of course, op-codes are the manner in which instructions are stored as program steps.

[Etienne Victoria]

Hallo Marcus,

The PPC developed a very efficient code for the Mantissa function.

I keep it permanently assigned to my SHIFT ENTER key.

This function, named MT, was built in the PPC ROM Module :

01 LBL 'MT'

02 CLA

03 STO [

04 ASTO [

05 INT

06 X<>0?

07 APPEND DIAMOND

08 X=0?

09 APPEND

10 CLX

11 ST+ [

12 X<>[

13 RTN

APPEND & DIAMOND: Hex characters in string (refer to combined Hex table).

SOURCE: The PPC ROM USERS MANUAL P. 349

Friendly regards from France.

Etienne

Edited: 5 Nov 2005, 11:05 a.m.

[Etienne Victoria]

Sorry,

It came to my mind that many people still have an IBM PC compatible and can't get all characters.

Therefore, please find hereunder the "graphical listing" for MT:

[ is register M and DIAMOND is hex 00.

Oh yes, and you can LastX the full number after looking at the mantissa!

Thanks.

Etienne

Edited: 5 Nov 2005, 12:59 p.m.

[Ángel Martin]

Hi,

in case this is useful, both the SandBox and the SandMath-III include a VMANT function. M-Code, based on that published in the Ken Emery's book, it shows the mantissa without disturbing X.

Besides, the SandMath-III also has MANT function to replace x with its mantissa.

Best,
ÁM