The Museum of HP Calculators

HP Forum Archive 16

[ Return to Index | Top of Index ]

HP 12C - When did quality go downhill?
Message #1 Posted by Insect on 17 Oct 2006, 9:35 a.m.

Hi,

When did the 12C quality go downhill from the 80's vintage models? Would a 1995 Singapore 12C be comparable to an 80s model? e.g. would 3532Sxxxxx be OK? I know it is a 1995 model, but not sure when the cost cutting occurred.

Thanks

      
Re: HP 12C - When did quality go downhill?
Message #2 Posted by JMatias on 23 Oct 2006, 12:45 p.m.,
in response to message #1 by Insect

Same for 49G, 49G+ and 17bII+. In fact, after the 48GX there isn´t a single good calculator.

JMatias

            
Re: HP 12C - When did quality go downhill?
Message #3 Posted by Bob Smith on 26 Oct 2006, 11:06 p.m.,
in response to message #2 by JMatias

I've been considering a replacement for my 42s which rececntly wore out on keys and frame abuse (carrying around factory etc.). Is the newer 50g worth considering? It seems quite a bit larger, but a colleague's 48G with quite a few built in equations and the excellent keyboard feel looks reasonable. Is there a referance for the changes that have occourred through this set of models? Thanks

                  
Re: HP 12C - When did quality go downhill?
Message #4 Posted by Ron Ross on 27 Oct 2006, 8:40 a.m.,
in response to message #3 by Bob Smith

The Hp50G is a nice calculator.

However, if you don't use trig functions or can get by with approximations, the Hp17Bii+ is a much smaller more compact calculator. Comes with 32K RAM, all of which can be used by the unlimited variable name convention. No real organization or directory structure, but the equations just pile up in a list menu of sorts.

If your equations do require trig, but not a lot of them, the Hp33s is Hp's present pocket scientific (a bit larger than the Hp17Bii, but it is still smaller than the Hp50G). However, it only has single letter variable names ie 27 variables, total available to the user. While it is advertised to have 32K, it can probably address or use 3-5 K RAM at the very most due to the variable length issue.

You might consider letting fixthatcalc take a look at your Hp42s as well. I haven't used their services, but they do have a good reputation (also a long wait, so you would still have to buy something in the interim)

                  
42S vs. 50g (was:Re: HP 12C - When did quality go downhill?)
Message #5 Posted by Jeff O. on 27 Oct 2006, 9:11 a.m.,
in response to message #3 by Bob Smith

Bob,
I was working on a response when Ron posted his. There is some duplicate information, but I'll present it anyway.
The 50g is a powerful, versatile machine that appears to be of sufficient quality to consider on its own merits, but I do not believe that you will find the 50g to be a satisfying replacement for your 42S. The 48/49/50 series of calculators operate on the RPL system, as opposed to the pure RPN system of the 42S. The systems are of course related, so that if an RPN user picks up an RPL calculator, her or she will be able to use it (as opposed to the frustrated looks that non-RPN users get when confronted with an RPN calculator.) But they are different enough that you will likely have to work at understanding and using the 50g to its potential, and perhaps even for the uses that you formerly got out of your 42S. I am not aware of a single reference that describes the changes from the 42S through the 50g. If you are completely unfamiliar, I can offer some fairly basic information. I consider the biggest immediately obvious differences between the RPN and RPL models to be:
1) Interface is through a command line on RPL models rather than directly onto the bottom level of the stack as with RPN models
2) RPL models have an “unlimited” stack (within available memory) vs. the four level X, Y, Z, T stack of the RPN models.
The path from 42S to 50g was not directly evolutionary. The RPL models started in 1987 with the HP-28C, which was followed by the 28S, the 48S series, the 48G series, the 49 series and now the 50g. The 42S can basically trace its roots back to the HP-35 in 1972, and it is considered by many (myself included) to be the zenith of the RPN line. Unfortunately, it was discontinued in 1995. The 32Sii, a less-capable model of the same family, continued in production until about 2003 when it was likewise discontinued. In 2004, the 33s was introduced, which was basically the 32sii in a new package with a few changes. Unfortunately, its styling was a radical departure from classic HP models, and the ENTER key was moved and reduced in size. Despite its flaws (if you consider the above to be flaws), the 33s retains classic RPN functionality, and may be a good choice for you to carry around the factory. Be aware that the 33s is considerably less capable than the 42S in many ways, but it is the closest thing to a classic RPN scientific programmable that is currently easily available.
If you are familiar with the 33s and do not care for it, you may very well do just fine with a 50g. One alternative would be to buy a 50g, then buy Hrastprogrammer’s HP-42X emulator for the 49g+ which should work on the 50g. This emulator will give you the exact functionality of the 42S on the 50g with a lot of enhancements (such as mass storage, input and output of programs, etc.) This would not be an inexpensive option, however. You might be able to buy another 42S on eBay for about the same cost.
Good luck and best regards.

                        
Re: 42S vs. 50g (was:Re: HP 12C - When did quality go downhill?)
Message #6 Posted by Marcus von Cube, Germany on 27 Oct 2006, 9:42 a.m.,
in response to message #5 by Jeff O.

As an aside, the HP 28S manual contains a short section (6 pages) titled "Notes for HP RPN Calculator Users" with the following subheadings:

  1. The Dynamic Stack
  2. Stack-Lift Disable and ENTER
  3. Prefix Versus Postfix
  4. Registers Versus Variables
  5. LASTX Versus LAST

So there are significant differences in handling between these. Since the 50g is a direct descendant of the 28S this information is still valuable. You should be able to find the 28S manual online or on the museum DVD.

Marcus

                        
Re: 42S vs. 50g (was:Re: HP 12C - When did quality go downhill?)
Message #7 Posted by Thomas Okken on 27 Oct 2006, 9:51 a.m.,
in response to message #5 by Jeff O.

Quote:
One alternative would be to buy a 50g, then buy Hrastprogrammer’s HP-42X emulator for the 49g+ which should work on the 50g. This emulator will give you the exact functionality of the 42S on the 50g with a lot of enhancements (such as mass storage, input and output of programs, etc.) This would not be an inexpensive option, however.

Or, if you can live without an actual keyboard (I know, that's a big "if"), get a low-end PDA and run Free42 on it. Palm Z22: $99 at CompUSA; Free42: $0.

- Thomas

                              
Re: 42S vs. 50g (was:Re: HP 12C - When did quality go downhill?)
Message #8 Posted by Dave Boyd on 27 Oct 2006, 10:24 a.m.,
in response to message #7 by Thomas Okken

Quote:
Or, if you can live without an actual keyboard (I know, that's a big "if"), get a low-end PDA and run Free42 on it. Palm Z22: $99 at CompUSA; Free42: $0.

Agreed -- it's not quite an "emulation" since it doesn't use the actual ROM image, rather it's a re-implementation of the functionality and a good simulation of the interface -- the only places I can tell it from a real 42 (aside from the obvious physical differences) is by the fact that the package offers several versions of the program, including one which uses a very nice decimal math package to give you a much more precision and a higher range of values -- it can tell you (very quickly) that 2000! is 3.32637509245 E 5735, which would overflow a real 42S.

Personally, I like a PDA with a higher-res screen, so I use a Palm T/X -- I have a lot of calculator emulators for it, and Free42 has one version with a very nice full-screen skin, big enough so that my big fingers can easily type on it if I'm careful... I also have Power48, which is a port of Emu48, and p41CX, which is a very nice emulator for the beloved HP41 series...

                              
Re: 42S vs. 50g (was:Re: HP 12C - When did quality go downhill?)
Message #9 Posted by Jeff O. on 27 Oct 2006, 11:22 a.m.,
in response to message #7 by Thomas Okken

Thomas,
I probably should have mentioned your fine Free42 option, but for me, that "if" is just too big. I can't see using a pda with its stylus out in a factory or field environment. A keyboard is a requirement.
With the above said, an ideal option would be porting Free42 to run on the 50g using hpgcc. I'm "working" on that, estimated completion around 2015. Maybe. (Actually, I have no idea if it is even possible, it just sounds like a good idea that I'd like to pursue.)

Edited: 27 Oct 2006, 11:23 a.m.

                                    
Re: 42S vs. 50g (was:Re: HP 12C - When did quality go downhill?)
Message #10 Posted by Thomas Okken on 27 Oct 2006, 1:35 p.m.,
in response to message #9 by Jeff O.

Quote:
an ideal option would be porting Free42 to run on the 50g using hpgcc. I'm "working" on that, estimated completion around 2015. Maybe.
(Actually, I have no idea if it is even possible, it just sounds like a good idea that I'd like to pursue.)

The Free42 core was written to be portable, but it does require a C++ compiler, and I don't think there is one for the 50g -- at least, I couldn't find anything on hpcalc.org, except for gcc, but that's C only. Getting rid of the C++ dependencies is actually not very difficult; until version 1.3, Free42 was pure C (except for the Windows front-end). However, it would mean sacrificing the Decimal version, since I grafted the BCD code onto Free42 by relying on operator and function overloading, and that is a C++ feature that C does not have.

- Thomas

                                          
Re: 42S vs. 50g (was:Re: HP 12C - When did quality go downhill?)
Message #11 Posted by Jeff O. on 27 Oct 2006, 4:30 p.m.,
in response to message #10 by Thomas Okken

Well shoot, there goes my plan!
Seriously, I was under the impression that Free42 required just a C compiler, so I was hopeful that the hpgcc route might work. Maybe there will be an hpgcc++ someday.

                                                
Free42 and the whole C-vs.-C++ issue
Message #12 Posted by Thomas Okken on 27 Oct 2006, 10:18 p.m.,
in response to message #11 by Jeff O.

Well shoot, there goes my plan!

The best-laid plans of mice and men... ;-)

Seriously, I was under the impression that Free42 required just a C compiler, so I was hopeful that the hpgcc route might work.

Before version 1.3, Free42 was pure C; versions 1.3.* are C++, but only in the sense that I made whatever modifications were necessary to get everything to compile as C++ -- but the code would still compile with a plain old C compiler, as long as you changed the Makefiles appropriately, changed all the *.cc filename extensions to *.c, etc.
In version 1.4, however, I introduced the decimal code which requires C++, partly because Hugh Steers' BCD20 library is C++, and partly because of the way I integrated that library into Free42. It wouldn't be all that hard to rewrite BCD20 in pure C, but modifying Free42 would be a chore... Imagine looking for all occurrences of "a+b", where a and/or b are floating-point numbers, and having to replace that with "bcd_add(a, b)". It's possible, of course, but it would be a pain... And, on top of all that, who would maintain such a pure-C version of Free42? Not me...

Maybe there will be an hpgcc++ someday.

I hope so!
In the early days of C++, this probably wouldn't even have been an issue, because back then, compiling C++ was a two-stage process, with the first stage being a C++-to-C translator, and the second stage being a plain old C compiler. Given a C++-to-C translator, it might still be possible to convert the Free42 source code to something hpgcc can compile, but I don't know if anyone even maintains that kind of thing any more... Once the gcc suite started supporting C++ with a dedicated compiler front-end, I lost interest in the whole C++-to-C translation approach, so I really couldn't say.

The other thing to hope for is that OpenRPN will mature into an actual product; if that happens, it will no doubt be possible to compile code for that platform using state-of-the-art C++ compilers (much like we can use similar tools to develop for PDAs today).

- Thomas

Edited: 27 Oct 2006, 10:23 p.m.

                                                      
Re: Free42 and the whole C-vs.-C++ issue
Message #13 Posted by Jeff O. on 30 Oct 2006, 7:47 a.m.,
in response to message #12 by Thomas Okken

Thomas,
Thanks for the detailed explanation. Your Free42 web site does of course make repeated reference to C++. I'm not sure why I thought it was pure C.
Best regards,
Jeff

                  
One (or More) Comments 42s replacement
Message #14 Posted by Ron Ross on 27 Oct 2006, 9:46 a.m.,
in response to message #3 by Bob Smith

Actually, if you buy an Hp17Bii USED from Ebay or other source, you will get an RPN based calculator with the Exact keyboard and basic layout of the Hp42s (if you are really adventurous, you could even swap guts, to get your old 42s back but I wouldn't bank on that option).

Many engineers like the older Hp17Bii for the layout similarity to the Hp42s. Actually has a nicer (easier anyway) solver to work with. The 17Bii actually has a bit better statistics functions, but lacks trig (does have pi, e^x, N! and other basic math though). No matrix functions of course.

If you are a power user and used the higher functions of the Hp42s, nothing less than an Hp 42s will work and if you can't replace it, you will have to move up to the RPL graphics line of Hp calculators.

If you do buy a graphics, you can easily buy a new Hp50G (or the slightly less expensive Hp48Gii). As you take it to the factory floor, perhaps cost IS an issue. If you dislike the new Keyboard layout, you may consider an Hp48G as these are easily found on ebay for a bit less than the newer more capable CAS powered Hp's. However, they have an ENTER key where you expect it to be. Keyboard quality IS NOT an issue as it will be the same as your previous Hp42s (same design, perhaps different build). The two flaws of the older Hp48G are that it is slow and it is big (second flaw not corrected by the new line either).

                        
Re: One (or More) Comments 42s replacement
Message #15 Posted by Bob Smith on 27 Oct 2006, 2:17 p.m.,
in response to message #14 by Ron Ross

Thank you for all of your responses. The trig functions are basic for me, so I will try the 50g, even though it is larger than I'd like. I also noticed some keybaoard notes relative to bounce. Knowing that it is adjustable will releive any issues with entry, which can be frustrating with input such a basic need. I have considered a PDA, but had similar concern on usage in rough conditions. Again, thanks to the many informed and dedicated users.

                        
Re: One (or More) Comments 42s replacement
Message #16 Posted by Arnaud Amiel on 30 Oct 2006, 9:03 a.m.,
in response to message #14 by Ron Ross

Actually on old 48G the keyboard can be a problem as some rows of keys start not working when the connection is not made anymore by the ruber thingy (?) under the screen. I would be careful if getting a 48. They are good though

Arnaud


[ Return to Index | Top of Index ]

Go back to the main exhibit hall