ARM-based 12c compared to the old one



#2

Here you can find a short comparison of the venerable HP12c (made in Brazil) compared with a new, ARM based HP 12c

HP12c new vs. old comparison

There is hope of a new HP15c being released - as Eric would say, "those who know do not speak, those who speak..."


#3

>Only HP and Eric Smith know how much Nonpareil lies within the new calculator.

None at all. Cyrille was very thankful to have a resource like Eric available to discuss with as needed, and to have a reliable way to check his emulator was behaving properly, but it is all original HP code. Granted that was before my time at HP, but I do have it direct from at least 1 of the parties involved. :-)

Cyrille's emulator is different since it was designed from the start for a hardware product and so some things were optimized away in order to maximize battery life (sometimes causing problems since special cases are handled individually - the SST capability comes to mind).

Great page!

Regarding the 12cp, have you taken a look at the current models? They have no known bugs I am aware of. I think everyone agrees the first two revisions were rather bad. Once the 25th anniversary one came out though I've not heard any complaints (regarding software at least).

TW

Edited: 12 Mar 2011, 5:15 p.m.


#4

I believe Tim is correct; I don't think there is any of my Nonpareil code in any HP product. The closest thing is that I contributed a short snippet of C code that does parallel BCD addition, vs. digit-at-a-time. That C code has been posted here, and Cyrille also posted a highly-optimized ARM assembly language version here. (HP has filed a patent application on how Cyrille implemented the carry out.) I don't know which calculators HP is using this in, but the 12C, 20b, and 30b are obvious candidates.

As Tim says, I have from time to time tried to help Cyrille with ensuring that some details of his simulation code are accurate.


#5

With due respect to people more knowledgeable than me (Eric and Cyrille)... may there be enough originality in a carry-out method to deserve a patent? Would it be generic or processor-architecture dependent? Would it be enforceable?

I understand and respect copyright claims; but I suppose a patent needs more than this kind of substance. Just my humble opinion.

Disclaimer: I'm no lawyer, no patent expert or otherwise intellectual-property expert... I just learned a little about these matters reading IEEE articles about these issues (Richard Stern's MicroLaw column in IEEE Micro magazine, among others) many years ago.


#6

I believe that there is prior art for the way the carry out is done. It remains to be seen whether a patent is actually issued.


#7

hello,

I beleive that the patent has actually been issued... one more useless SW patent... but companies love to have large "portfolio" of patent.. my understanding is that it's the size of the patent portfolio (mesured in inches of printed paper) that is important...


To be honnest, I personally do not like SW patents, but hey, the company likes them... so...


all Atmel based calcualtors use this code.

Eric has been of great help, providing me with a modified version of NoPareil that would generate log file of executed instructions... this has been extremly usefull to see when my code path execution was varying from his and find where my emulator was buggy... however, I have never looked at his code to avoid any possible questions about HP using free code without respecting lisences...

cyrille


#8

Cyrille, thank you for your comments and openess. Very enlightening! Best regards, Andrés.

#9

Quote:
There is hope of a new HP15c being released [...]
Why do you think so? While development costs might be rather low, that thing has to be produced and sold ;-). Imagine a 15C between all those calculators with graphical displays and lots of RAM. Only few will buy, since the strengths of the 15C cannot be marketed easily.

You know, 'ease of use' today means 'simple' or 'entry level'. And 'bugfreeness'? One even don't knows if it is. And people are over-optimistic in the light of a high-specs gadget anyway, I fear.

In the end, some more important people might ask why HP needs to get back to a model that ancient. What's their concept of 'innovate'? ;-)

But ... HP could camouflage it as a retro-toy sold for their next anniversary :-D.


#10

Quote:
While development costs might be rather low, that thing has to be produced and sold ;-). Imagine a 15C between all those calculators with graphical displays and lots of RAM. Only few will buy, since the strengths of the 15C cannot be marketed easily.

I can't imagine a new 15c being sold at retail. Very few HP calcs are sold at retail and as you point out, the 15C would not compare well.

Internet sales should be very inexpensive.

As for production costs, has anyone tried the 15c firmware replacing the 12c firmware component in the current emulated 12c?

If the display is sufficient, and the emulator is sufficient or could be fixed, then incremental production costs for the 15c over the 12c calculator are limited to a different paint job on the keys and a different firmware load. Conceivably even the firmware load could be done post-manufacture if this is a small scale effort.

If no packaging and manual are needed it seems that financing the cost of a small production run would be well within the means of many of the HP calculator hobbyists. (With legal permission from HP, of course.) HP would have no means to deal with a small production run like that, but for an individual or small group selling via the internet then the missing "commercial" package and documentation seem reasonable.


#11

Quote:
If the display is sufficient, and the emulator is sufficient or could be fixed, then incremental production costs for the 15c over the 12c calculator are limited to a different paint job on the keys and a different firmware load.
I have no clue what that costs, I admit. It is an interesting question for the 20b/30b repurposing projects anyway.

Quote:
If no packaging and manual are needed it seems that financing the cost of a small production run would be well within the means of many of the HP calculator hobbyists. (With legal permission from HP, of course.) HP would have no means to deal with a small production run like that, but for an individual or small group selling via the internet then the missing "commercial" package and documentation seem reasonable.
Right. Would HP deal with a bunch of hobbyists willing to do the job?

Thanks for your interesting thoughts about this matter!

#12

would someone like to humour me and tell me where i might find any repository of information on the ARM based 12C? has anyone had a go at developing a GCC toolchain to do programming of it? is there documentation on port mappings, etc?

am toying with the idea of getting hold of one and having a go at doing something useful with it...


cheers,
rob :-)


#13

There's an SDK from HP for developing on the 20b. The new 12c has basically the same hardware, with different keyboard and LCD mapping, so the 20b SDK is mostly applicable.

As supplied, the SDK works with the IAR toolchain. The "Kickstart" version of the IAR toolchain is available for download from IAR at no charge. It has almost all of the features of the full version, except that it won't link an executable larger than 32KB of object code.

The GNU toolchain from CodeSourcery should be perfectly suitable for developing for the 12c/20b/30b, but using the 20b SDK would then require some modification.

I'm currently using IAR Kickstart on a different platform (ARM Cortex-M3), but plan to switch to CodeSourcery at some point.

#14

Quote:
would someone like to humour me and tell me where i might find any repository of information on the ARM based 12C? has anyone had a go at developing a GCC toolchain to do programming of it? is there documentation on port mappings, etc?

Yes, GCC works. I've mapped out the keypad and LCD.


#15

sounds like what i'm after - is your work posted anywhere on the net?

cheers,
rob :-)

#16

I'm working on a 20b/30b repurposing project (wp34s). I've plans to use the Yagarto toolchain. It's freely downloadable and came with my Olimex JTAG adapter. I'll keep you posted here how it works out.


Possibly Related Threads...
Thread Author Replies Views Last Post
  HP-27S speed compared to HP-42S Tommi 3 591 06-05-2013, 05:55 PM
Last Post: Christoph Giesselink
  HP-9100 used in computer-based education (video) Juergen Keller 5 786 12-07-2012, 11:04 AM
Last Post: hpnut
  Homemade ARM-based Calculator Programming Cable Katie Wasserman 18 2,023 08-27-2012, 11:25 PM
Last Post: Katie Wasserman
  Saturn based emulators for android Olivier De Smet 16 1,260 07-29-2012, 01:32 PM
Last Post: Hans-Erik Lehndal
  NCEES and Computer Based Testing Richard Garner 2 411 04-10-2012, 07:14 PM
Last Post: Steve Fennell
  Different keyboards on ARM 12c Jose Gonzalez Divasson 6 665 10-17-2011, 01:52 PM
Last Post: Paulo MO
  12C ARM Firmware? Lincoln R. 0 243 10-01-2011, 06:04 PM
Last Post: Lincoln R.
  Is there anything the 15c can't do compared to the 35s? snaggs 156 9,350 09-02-2011, 05:36 PM
Last Post: Egan Ford
  34s - stack lift with PI compared to 15c - difference - anyone else able to repeat this? Gene Wright 39 2,704 07-16-2011, 11:30 PM
Last Post: Jake Schwartz
  HP 12C, 12C Platinum & 15C iOS App Walter Lam 2 450 06-02-2011, 01:25 PM
Last Post: Andrés C. Rodríguez (Argentina)

Forum Jump: