RPN83P: RPN calculator for TI-83+ TI-84+ inspired by HP-42S

[bxparks]

I just released version 0.7.0 of RPN83P, an RPN calculator app for the TI-83 Plus and TI-84 Plus calculators, inspired mostly by the HP-42S, with significant features from the HP-12C, and HP-16C. I am probably heading towards a v1.0 milestone release in a couple of months. I would appreciate any feedback from users who may notice things that I may have overlooked.

The program is written in Z80 assembly language. It consumes 2 pages (32 kiB) of flash memory. The project home page has screenshots.

Project Home: https://github.com/bxparks/rpn83p

Installation:

• Download the rpn83p.8xk file from the RPN83P Releases page.
• Upload the file to the TI-83 Plus or TI-84 Plus calculator.
• Run the program using the APPS button and selecting "RPN83P".
  

Exiting:

• Quit app: 2ND QUIT
• Turn off device: 2ND OFF
  

Features [Updated on 2024-03-31 for v0.10.0]:

• traditional 4-level RPN stack (X, Y, Z, T) with LastX register
• 8-line display showing all stack registers
• hierarchical menu system similar to HP-42S
• quick reference HELP menu
• storage registers
  
  • store and recall:STO nn, RCL nn
  • storage arithmetics: STO+ nn, STO- nn, STO* nn, STO/ nn, RCL+nn, RCL- nn, RCL* nn, RCL/ nn
  • 25 storage registers: nn = 00..24
  • 27 single-letter variables (nn = A..Z,Theta)
    
• support for all math functions with dedicated buttons on the TI-83 Plus and TI-84 Plus
  
  • arithmetic: /, *, -, +
  • algebraic: 1/X, X^2, SQRT
  • transcendental: ^ (i.e. Y^X), LOG, 10^X, LN, e^X
  • trigonometric: SIN, COS, TAN, ASIN, ACOS, ATAN
  • constants: pi and e
    
• additional menu functions:
  
  • arithmetic: %, %CH, GCD, LCM, PRIM (prime factor), IP (integer part), FP (fractional part), FLR (floor), CEIL (ceiling), NEAR (nearest integer), ABS, SIGN, MOD, MIN, MAX
  • algebraic: X^3, 3RootX
  • transcendental: XRootY, 2^X, LOG2, LOGB, E^X- (e^x-1), LN1+ (log(1+x))
  • trigonometric: ATN2
  • probability: PERM, COMB, N!, RAND, SEED
  • hyperbolic: SINH, COSH, TANH, ASNH, ACSH, ATNH
  • angle conversions: >DEG, >RAD, >HR, >HMS, >REC, >POL
  • unit conversions: >C, >F, >hPa, >inHg, >km, >mi, >m, >ft, >cm, >in, >um, >mil, >kg, >lbs, >g, >oz, >L, >gal, >floz, >kJ, >cal, >kW, >hp
    
• statistics and curve fitting, inspired by HP-42S
  
  • statistics: Sigma+, Sigma-, SUM, MEAN, WMN (weighted mean), SDEV, SCOV, PDEV, PCOV
  • curve fitting: Y>X, X>Y, SLOP (slope), YINT (y intercept), CORR (correlation coefficent)
  • curve fit models: LINF (linear), LOGF (logarithmic), EXPF (exponential), PWRF (power)
    
• base conversion and bitwise operations, inspired by HP-16C and HP-42S
  
  • base conversions: DEC, HEX, OCT, BIN
  • bitwise operations: AND, OR, XOR, NOT, NEG, REVB (reverse bits), CNTB (count bits)
  • integer arithmetics: B+, B-, B*, B/, BDIV (divide with remainder)
  • shift and rotate: SL, SR, ASR, RL, RR, RLC, RRC, SLn, SRn, RLn, RRn, RLCn, RRCn
  • carry flag and bit masks: CCF, SCF, CF?, CB, SB, B?
  • word sizes: 8, 16, 24, 32 bits
    
• time value of money (TVM), inspired by HP-12C HP-17B, and HP-30b
  
  • N, I%YR, PV, PMT, FV, P/YR, BEG, END, CLTV (clear TVM)
    
• complex numbers, inspired by HP-42S and HP-35s
  
  • stored in RPN stack registers (X, Y, Z, T, LastX) and storage registers R00-R24
  • computation modes: RRES (real results), CRES (complex results)
  • display modes: RECT, PRAD (polar radians), PDEG (polar degrees)
  • linking/unlinking: 2ND LINK (convert 2 reals to 1 complex, same as COMPLEX on HP-42S)
  • number entry: 2ND i (rectangular), 2ND ANGLE (polar degrees), 2ND ANGLE 2ND ANGLE (polar radians)
  • extended regular functions: +, -, *, /, 1/x, x^2, SQRT, Y^X, X^3, 3RootY, XRootY, LOG, LN, 10^x, e^x, 2^x, LOG2, LOGB
  • complex specific functions: REAL, IMAG, CONJ, CABS, CANG
  • unsupported: trigonometric and hyperbolic functions (not supported by TI-OS)
    
• date functions
  
  • date, time, datetime, timezone, and hardware clock
  • add or subtract dates, times, datetimes
  • convert datetime to different timezones
  • convert between datetime and epochseconds
  • support alternative Epoch dates (Unix, NTP, GPS, TIOS, Y2K, custom)
  • set and retrieve datetime from the hardware clock (84+/84+SE only)
  • display time and date objects in RFC 3339 (ISO 8601) format
    
• various modes (MODE)
  
  • floating display: FIX, SCI, ENG
  • trigonometric computation: RAD, DEG
  • complex computation: RRES, CRES
  • complex display: RECT, PRAD, PDEG
  • SHOW (2ND ENTRY): display all 14 internal digits
    

Missing features (partial list):

• vectors and matrices
• keystroke programming
  

[dm319]

That's very cool!

You've made me want a TI calculator. Could be a reasonable option for those who want RPN but not the price of HPs.

How does roll and x-y work?

[bxparks]

(11-20-2023 11:31 PM)dm319 Wrote:  You've made me want a TI calculator. Could be a reasonable option for those who want RPN but not the price of HPs.

How does roll and x-y work?

Yeah, these TI-83+ and TI-84+ are dirty cheap on eBay. I recommend the 84+, it has the better LCD screen. I have now implemented enough of the HP-16C (which I don't have), that I find myself preferring to use my app to do hexadecimal, bitwise, and logical calculations.

I follow the HP-30b conventions with regards to roll and x<->y:

• Left paren ( is roll
• Right paren ) is x<->y
  

I think this is the convention used by other HP calculators with dual ALG and RPN modes, like the HP-17BII.

[Thomas Klemm]

It's nice to be reminded of an old thread: Looking for TVM formulas

[Gerald H]

Very nice work, bxparks, bravo!

PRIM actually finds a factor, otherwise returning 1 for prime input.

PRIM works very fast - What algorithm have you used?

[floppy]

Interesting. Widely spread hardware.
Has the TIs possibilities for external memory storage? (PC-link or SD or..)

[bxparks]

(11-21-2023 09:26 AM)Gerald H Wrote:  Very nice work, bxparks, bravo!

PRIM actually finds a factor, otherwise returning 1 for prime input.

PRIM works very fast - What algorithm have you used?

Thanks! Early versions of PRIM just returned a boolean (0 or 1) to indicate prime, but it was more useful to return the smallest prime factor. I have updated the README.md and USER_GUIDE.md to make that more clear. I don't know if you saw it, but I explained how to get all the prime factors of a number N (< 2^32) iteratively (https://github.com/bxparks/rpn83p/blob/d...me-factors), because PRIM lifts the original N to Y, then pushes the prime factor into X. So you can just hit / to get the next number to factor, and hit PRIM again.

The largest prime number <= 2^16 is 65521, so the longest execution time of PRIM is the factorization of 65521*65221=4 293 001 441. That takes 33 seconds on a TI-84+SE. I spent far too much time optimizing the PRIM function. :-) I just added notes about how PRIM works in the DEVELOPER.md notes here: https://github.com/bxparks/rpn83p/blob/d...VELOPER.md . I am copying the info below to save a click:

• The basics of the algorithm is to test all the candidate prime factors from 2 to sqrt(N).
• We could simply start at 3 and increment by 2 to test every odd number to sqrt(N). But we can do slightly better. All prime numbers >=5 are of the form 6k-1 and 6k+1. So each iteration can increment by 6, but perform 2 checks. This effectively means that we step by 3 through the candidate prime factors, instead of just by 2 (for all odd numbers), which makes the loop 50% faster.
• We use integer operations instead of TI-OS floating point ops. If I recall, this makes it about 2-3X faster (I thought it would be more, but floating point ops in TI-OS are surprisingly fast).
• Z80 does not support integer division operations in hardware, so we have to write our own in software. The integer size of N is limited to 32 bits, so we need to write a div(u32, u32) routine.
• But the loop only needs to go up to sqrt(N), so we actually only need a div(u32, u16) routine, which if I recall is about 2X faster. This is because the bit-wise loop is reduced by 2X, but also because the dividend can be stored in a 16-bit Z80 register, instead of stored in 4 bytes of RAM.
• Finally, we actually don't need a full div() operation for the PRIM function. We don't need the quotient, we need only the remainder. So we implement a custom mod(u32, u16) function which is about 25% faster than the full div(u32, u16) function.
  

I think there are additional micro-optimizations left on the table that could make the PRIM function maybe 1.5X to 2X faster, without resorting to a completely different algorithm. But I suspect that the resulting code would be difficult to understand and maintain. So I decided to stop here.

[bxparks]

(11-21-2023 10:46 AM)floppy Wrote:  Interesting. Widely spread hardware.
Has the TIs possibilities for external memory storage? (PC-link or SD or..)

I am not aware of an SD card extension (although it sounds like something someone, somewhere would have hacked together). But the USB-serial link to upload the RPN83P program from the host computer to the calculator is bi-directional. That means the external storage limit of the calculator is limited only by your host computer. (Unless I misinterpreted your question...)

[bxparks]

(11-21-2023 03:34 AM)Thomas Klemm Wrote:  It's nice to be reminded of an old thread: Looking for TVM formulas

Yes, definitely, that is one of the links that I referenced in my USER_GUIDE.md. During the implementation of the TVM functionality in the RPN83P app, I researched and read almost every post in every thread about TVM formulas and algorithms on the hpmuseum.org site. They were extremely helpful. I was planning to summarize my findings to the recent TVM formula error in programming manual? thread shortly.

[spiff72]

I don't have an 83 or an 84, but I do have an 89 Titanium laying in a drawer. Is there any possibility that this could be ported to that platform? I think there was an RPN app of some kind for the non-Titanium version of the 89, but it doesn't work on the Titanium.

[bxparks]

(11-21-2023 07:38 PM)spiff72 Wrote:  I don't have an 83 or an 84, but I do have an 89 Titanium laying in a drawer. Is there any possibility that this could be ported to that platform? I think there was an RPN app of some kind for the non-Titanium version of the 89, but it doesn't work on the Titanium.

I guess the important questions are:

* What is the state of the development tools for those calculators? How well are they supported and documented?
* How much time and energy do I have?

The TI-89, TI-89 Titanium, TI-92, TI-92+, and the TI Voyage 200 calculators use the Motorola 68000 processor, not the Z80. Basically nothing of the RPN83P code base would be reusable. The development tools of the 89/92/200 are completely different as far as I know, and I have not done any research on those tools. I don't know if TI ever published an SDK for those calculators, similar to their 83 Plus SDK. I know that those calculators can be programmed in C instead of assembly language, which would increase productivity.

On the question of time and energy, I have so many projects in my backlog, I'm not sure what would cause the RPN89T/RPN92P project to rise above the others. By the time I get to v1.0 on this RPN83P project (in a couple of month hopefully), I will have spent 2-3X longer on it than I had expected. The 83+ and 84+ calculators were attractive targets, because there are millions of them out there, and they are dirt cheap on eBay. You can pick one up for $15-$35 on eBay including shipping. I get the impression that the 89, 92, and 200 calculators are far less popular. It seems like a LOT of work, for not many potential users.

I tell you what. If you want to increase the chances of seeing a port to the 89T, 92+, and 200, I request that you scout ahead on the development tools for the TI-89T. Figure out how to download and setup the C compiler on Ubuntu 22.04 (my dev environment). Figure out if there is an SDK, figure out where the docs are located, figure out how to create a "hello world" program for the 89T, how to transfer the executable to the calculator, how to read the keyboard, how to print to the screen, how to read and store variables. There is so much more, but I guess that would be a start. That's probably a month or two of effort. If you come back and tell me that the dev environment for the 89/92/200 is easy to set up, and the OS routines are well-documented, then I would be more likely to be persuaded to do a port. Or, you will have learned enough to do the port yourself. :-)

[floppy]

(11-21-2023 05:46 PM)bxparks Wrote:  

(11-21-2023 10:46 AM)floppy Wrote:  Interesting. Widely spread hardware.
Has the TIs possibilities for external memory storage? (PC-link or SD or..)

I am not aware of an SD card extension (although it sounds like something someone, somewhere would have hacked together). But the USB-serial link to upload the RPN83P program from the host computer to the calculator is bi-directional. That means the external storage limit of the calculator is limited only by your host computer. (Unless I misinterpreted your question...)

Rewording: any program on the TI83/84+ can be
a) uploaded into a PC storage place (or SD or whatever standard storage)
b) deleted from the TI
c) uploaded again into the TI if any need
?

[bxparks]

(11-21-2023 09:07 PM)floppy Wrote:  Rewording: any program on the TI83/84+ can be
a) uploaded into a PC storage place (or SD or whatever standard storage)
b) deleted from the TI
c) uploaded again into the TI if any need
?

Yes. There is a host app that runs on Windows (TI Connect), Mac (TI Connect) or Linux (TiLP2). It gives you a view of the contents of the calculator's memory that looks like a file system. You can drag-and-drop any app or variable from the calculator to the host. You can drag-and-drop those files from the host back to the calculator.

You can also connect 2 TI calculators together through a 2.5 mm cable, and transfer files between them. I've actually never done it myself, but apparently it's possible.

[floppy]

(11-21-2023 09:26 PM)bxparks Wrote:  Yes. There is a host app that runs on Windows (TI Connect), Mac (TI Connect) or Linux (TiLP2). It gives you a view of the contents of the calculator's memory that looks like a file system. You can drag-and-drop any app or variable from the calculator to the host. You can drag-and-drop those files from the host back to the calculator.

You can also connect 2 TI calculators together through a 2.5 mm cable, and transfer files between them. I've actually never done it myself, but apparently it's possible.

Ouch. Perhaps you just signed the death of my HP35S (no exchange with PC possible with that weird one). Now the TI84+ is on my amazon list. In few days will probably make its way to me. Then after install/test probably bye bye HP35S via eBay.

The TI-84 Plus C Silver Edition would be ok? lt seems to be the highest spec version.

[spiff72]

(11-21-2023 08:50 PM)bxparks Wrote:  

(11-21-2023 07:38 PM)spiff72 Wrote:  I don't have an 83 or an 84, but I do have an 89 Titanium laying in a drawer. Is there any possibility that this could be ported to that platform? I think there was an RPN app of some kind for the non-Titanium version of the 89, but it doesn't work on the Titanium.

I guess the important questions are:

* What is the state of the development tools for those calculators? How well are they supported and documented?
* How much time and energy do I have?

The TI-89, TI-89 Titanium, TI-92, TI-92+, and the TI Voyage 200 calculators use the Motorola 68000 processor, not the Z80. Basically nothing of the RPN83P code base would be reusable. The development tools of the 89/92/200 are completely different as far as I know, and I have not done any research on those tools. I don't know if TI ever published an SDK for those calculators, similar to their 83 Plus SDK. I know that those calculators can be programmed in C instead of assembly language, which would increase productivity.

On the question of time and energy, I have so many projects in my backlog, I'm not sure what would cause the RPN89T/RPN92P project to rise above the others. By the time I get to v1.0 on this RPN83P project (in a couple of month hopefully), I will have spent 2-3X longer on it than I had expected. The 83+ and 84+ calculators were attractive targets, because there are millions of them out there, and they are dirt cheap on eBay. You can pick one up for $15-$35 on eBay including shipping. I get the impression that the 89, 92, and 200 calculators are far less popular. It seems like a LOT of work, for not many potential users.

I tell you what. If you want to increase the chances of seeing a port to the 89T, 92+, and 200, I request that you scout ahead on the development tools for the TI-89T. Figure out how to download and setup the C compiler on Ubuntu 22.04 (my dev environment). Figure out if there is an SDK, figure out where the docs are located, figure out how to create a "hello world" program for the 89T, how to transfer the executable to the calculator, how to read the keyboard, how to print to the screen, how to read and store variables. There is so much more, but I guess that would be a start. That's probably a month or two of effort. If you come back and tell me that the dev environment for the 89/92/200 is easy to set up, and the OS routines are well-documented, then I would be more likely to be persuaded to do a port. Or, you will have learned enough to do the port yourself. :-)

Sorry - I didn't realize they ran different processors and "porting" wasn't that simple. (I'm more of a hardware/sticker guy (see my signature)...

Maybe I will look into picking up an 83+ or 84+ to add to the collection. Is there a preferred one for this project? I'm guessing the 84+ silver is the better option?

[dm319]

(11-21-2023 06:00 PM)bxparks Wrote:  

(11-21-2023 03:34 AM)Thomas Klemm Wrote:  It's nice to be reminded of an old thread: Looking for TVM formulas

Yes, definitely, that is one of the links that I referenced in my USER_GUIDE.md. During the implementation of the TVM functionality in the RPN83P app, I researched and read almost every post in every thread about TVM formulas and algorithms on the hpmuseum.org site. They were extremely helpful. I was planning to summarize my findings to the recent TVM formula error in programming manual? thread shortly.

Please do! I wrote a TVM solver for the DM42 (available here) specifically for NSTK mode. I suspect it may need a bit more work/optimising though as just managed to get a divide by 0 error putting in a really big N.

[bxparks]

(11-21-2023 10:02 PM)floppy Wrote:  Ouch. Perhaps you just signed the death of my HP35S (no exchange with PC possible with that weird one). Now the TI84+ is on my amazon list. In few days will probably make its way to me. Then after install/test probably bye bye HP35S via eBay.

Texas Instruments seems to have thoroughly abused the "TI-84 Plus" brand. There are at least 4 calculators by that name:

* TI-84 Plus
* TI-84 Plus Silver Edition
* TI-84 Plus C Silver Edition
* TI-84 Plus CE

Plus the "Python" versions of some of those. Some of them are quite incompatible with each other. Make sure you are getting what you think you are getting.

The RPN83P works with the TI-84 Plus, and TI-84 Plus SE (and the TI-83 Plus), but *not* the TI-84 Plus C SE, and most *definitely* not the TI-84 Plus CE.

With regards to the HP-35S, it all depends on what you want the calculator for. Most reviewers seem to love the retro-quality of the 35S, with echoes of the old Voyager (10C to 15C) and Pioneer series (20S, 32S, 42S). All of its functions seem to be available directly through the buttons instead of menus. Some people really like that. I'd probably enjoy having this calculator, but not at the prices that I see on eBay.

(11-21-2023 10:02 PM)floppy Wrote:  The TI-84 Plus C Silver Edition would be ok? lt seems to be the highest spec version.

I don't have this calculator, but everything I read on the internet says, *never* get this calculator. It uses the same 15 MHz Z80 processor as the 84+ and 84+SE, but has a 320×240 color screen that requires a much higher CPU and memory resources. The Z80 processor apparently cannot render the fonts fast enough and does not have enough bandwidth to push the bits quickly enough to the LCD screen.

[Gerald H]

My own preference is the 83+SE - awfully elegant. The 84+SE is merely greyish & looks much bulkier although it in fact weighs less.

[floppy]

TI 84 PLUS
This cable is for the connectivity to the PC? https://www.amazon.de/TI-Graph-Texas-Ins...6BXBS?th=1
Or this one ? https://www.amazon.de/Datenkabel-Ladekab...10&sr=8-17

Looks like 2 different types are listed. Not sure which one is the best.

[Thomas Klemm]

(11-21-2023 06:00 PM)bxparks Wrote:  Yes, definitely, that is one of the links that I referenced in my USER_GUIDE.md.

I noticed the link while reading the source code of lnOnePlus in float.asm.

---

A few lines above we can find the following comment:

Quote: ; This uses ln(1+x) = x * log(1+x) / ((1+x)-1). I think this algorithm is
; faster than above, but apparently it doesn't work on certain computers
; (though I don't know why).
; See https://math.stackexchange.com/questions/175891

What do you mean exactly with "certain computers"?
Does it work with these TI calculators?
Otherwise, can you debug the code and maybe provide a trace of OP1, OP2 and FPS?