RPN-65 Question.

[Matt Agajanian]

Hi all.

Given that transcendental functions and others used register nine as a scratchpad space, was there any way that the 65 could have been built so that a scratchpad register wouldn't be necessary? Likewise, would it have been possible that the 35 wouldn't need to use the T register?

Apologies for posting this in the wrong forum.

[Steve Simpkin]

I don't think so. The Arithmetic and Register (A&R) chip used on the Classic series contained 7 56-bit (14 BCD digits) registers. This was not quite enough for calculating trig functions. Three of these registers were the X, Y, and Z registers of the four-level stack. Two more were working registers and one was the user-visible memory register (STO/RCL.) The final register did double duty as the topmost register in the stack (T) and as a working register for trigonometric functions. Thus, when a trig function was used, the topmost value in the stack was lost. It appears the HP-65 used register 9 instead of stack register T. But the result was the same. 7 registers were just not enough for the Stack and working registers during trig calculations.

Among the improvements on the ACT chip used in the Woodstock series was the addition of one more internal register.

From the "20 Series Technology and Packaging" page:

"The heart of the 20 series calculators, as well as many pocket calculators that followed, was the Arithmetic, Control and Timing chip (ACT). This chip integrated the earlier generation's register, arithmetic, control, and clock driver circuit into a single package reducing size and cost.

The ACT chip had 8 56-bit data registers, an address register, two return address registers, 16 flags, and a 4-bit display format register. The chip could directly address 4096 instructions with its 12-bit address space. Because there was now one more register on the chip, trig calculations no longer overwrote the top of the stack as they did on the HP-35. "

https://www.hpmuseum.org/tech20.htm

[Matt Agajanian]

Steve, thanks for the repeat of the info about the Woodstock series. Very appreciated and illuminating.

[Steve Simpkin]

(08-20-2023 08:58 PM)Matt Agajanian Wrote:  Steve, thanks for the repeat of the info about the Woodstock series. Very appreciated and illuminating.

Matt, No problem I learned something new to me. I did not realize that the HP-65 used storage register 9 for trig functions. So did the HP-45. The HP-55 on the other hand, did not appear to need any stack or register location for trig functions.

According to Eric's HP Calculator Memory Sizes site, the HP-55 had 30 RAM registers total. Of these 20 were used for storage registers R0-R.9, 1 for Last X and 7 for the 49 program step storage. This left 2 registers left over that were not available to the user. It appears they used one of these for trig functions. I am not sure if the other leftover RAM register was used for anything (maybe the timer?).

[Matt Agajanian]

Unless I’m missing something from that page, is there a beginning homepage? Let me know, please. Thanks.

[Steve Simpkin]

(08-20-2023 10:29 PM)Matt Agajanian Wrote:  Unless I’m missing something from that page, is there a beginning homepage? Let me know, please. Thanks.

Here is Eric's HP calculator home page. There was no link to it from the HP Calculator Memory Sizes page.
http://www.brouhaha.com/~eric/hpcalc/

[Matt Agajanian]

Thanks!

[teenix]

From the emulator, RAM [20] and [21] are used for trig calculations and RAM[22] doesn't appear to be used for anything. This is strange as I thought with the cost of chips in those days, it would have some use.

RAM[20] stores the Last X value too, so if you do a trig function, the Last X is lost. I wonder why RAM[22] wasn't used for Last X then.

I suspect there just wasn't enough ROM to code in a different address for Last X using RAM[22], as the Microcode that sets up the address for RAM[20] is a common routine between trig functions and Last X.

cheers

Tony

[Steve Simpkin]

(08-21-2023 12:30 AM)teenix Wrote:  From the emulator, RAM [20] and [21] are used for trig calculations and RAM[22] doesn't appear to be used for anything. This is strange as I thought with the cost of chips in those days, it would have some use.

RAM[20] stores the Last X value too, so if you do a trig function, the Last X is lost. I wonder why RAM[22] wasn't used for Last X then.

I suspect there just wasn't enough ROM to code in a different address for Last X using RAM[22], as the Microcode that sets up the address for RAM[20] is a common routine between trig functions and Last X.

cheers

Tony

Tony, can you check that again? Using your HP Classic Emulator for PC (an Awesome program!) for the HP-55 it looks like RAM[21] is used as a scratch register during trig calculations but RAM[20] always has the correct Last X value in it even after a trig operation. Also in use, the Last X function seems to work as it should even after a trig calculation.

[teenix]

Sorry, you were right.

I didn't realize the result of a previous calculation was put in Last X when (say) [f] [SIN] was pressed.

cheers

Tony