The Museum of HP Calculators

HP Forum Archive 13

[ Return to Index | Top of Index ]

HP-42S With 32K
Message #1 Posted by Tom Cox on 14 Aug 2003, 8:34 p.m.

Have had a modified unit for a couple of months now and it works just great. Fortunately, the 42S is rock-stable and there is not much danger in losing memory. It took me many hours to key in the nearly 8k of programs I had in machine before the unit was modified.

I would be really upset if I had machine nearly full of 32k and crashed it as it would take a looooooooong time to key in that much stuff.

Will reserve judgment on 33S until more details and reports are available -- but, the features appear to make it a suitable substitute for 32SII. For someone that uses matrices, it would not seem to be a reasonable substitute for 42S.

Unless the reviews are outstandingly bad, will probably purchase one just to see what it can do. As for the keyboard, I can probably get used to it. Perhaps HP was copying the TI-34 when they made Integer Divide a top level key.

Does anyone know when the 33S will actually be available for sale?

All things considered, for $50-$60 the HP-33S appears to be a reasonably priced machine.

Re: HP-33S With 32K
Message #2 Posted by Paul Brogger on 15 Aug 2003, 10:55 a.m.,
in response to message #1 by Tom Cox

I think the reasoning behind 32K in the 33s is "no practical limit" -- a vast improvement over the mere several hundred bytes offered by the 32s/32sii. From that perspective, the lack of I/O, while unfortunate, is not illogical.

Also (and without really knowing anything about the subject) I suspect that, given the reliability of current fabrication processes and chip yields, 32K is not significantly more expensive than would have been a 2K or 8K addition to the CPU circuit. (That assumes they're on the same chip. If no one gets there before me, mine will be dismantled soon after I get my hands on it.)

Note that the 32s/sii programming model for line numbers (three digits unlabeled, plus 26 labels with two-digit suffixes for each) has a potential range of at most ~3,600 lines. The 33s' 14-character display lines (up from 12) must be dictated by the availability of 32K bytes of storage, if that is to be made available to programmers.

I note the addition of a "space" function, implying a "space" character -- that wasn't available in the 32s/sii, was it? That could imply variable and program naming, or maybe it's just a syntactic requirement elsewhere?

Space in 32sii
Message #3 Posted by bill platt on 15 Aug 2003, 11:13 a.m.,
in response to message #2 by Paul Brogger

Hi Paul,

There is a space in the 32sii---it is on the R/S key and it is only active when in the equation writer. Remember that in the 32sii, the equation writer can be used a number of ways:

1. You can write equations in the "equation list" and solve them for any varible. The "equation can be either an assignment, an expression, or an equality. Then, you can solve it, using "solve"

2. You can write equations in PROGRAM mode. This way, you can sneak algebraics into your programs. The equation takes up one "line" in hte program, and it can be solved in the program etc.

3. By setting a flag, you can use the equation writer in PROGRAM mode, but instead of treating the equation as a mathematical expression, it will instead just dump the contents to the display via a "pause" so that you can have text messages!

And also, you can integrate, too. So, for lots of stuff, this is just soo much more than the 11c ever was (and 19 times faster, too). But I still like the 11c. It is a beautiful thing.

So that's the long of the short of it.



Edited: 15 Aug 2003, 11:16 a.m.

Re: HP-33S With 32K
Message #4 Posted by John Moore on 20 Aug 2003, 6:16 a.m.,
in response to message #2 by Paul Brogger

I do not think that the 33s has 32K-bytes. It is more likely to be 32K-bits. The CPU is a SPLB31A. The specification can be found here:

and here:

The CPU has 4 K Bytes. This is the same as 32K-bits. I think someone has made a mistake.

It could use an extenal RAM but this would add a lot to the cost. The 12C Platinum uses a CPU from this range.

Re: HP-33S With 32K (bits?)
Message #5 Posted by Paul Brogger on 20 Aug 2003, 9:39 a.m.,
in response to message #4 by John Moore

In the 33s' description (as provided) there are enough grammatical mistakes and clumsy turns of phrase to make your observation very plausible.

4K would be more in line with the lack of electronic I/O and the apparent limitation of 27 registers, while still being a vast improvement over the severely hamstrung 32s/sii memory. Another benefit would be reduced need to alter the 32s/sii user programming address scheme (limited to something like 3573 lines, if I've got it right).

Quite apart from the obviously ugly keyboard, all this speculation about actual capabilities has me checking regularly . . .

Re: HP-33S With 32K (bits?)
Message #6 Posted by bill platt on 20 Aug 2003, 11:05 a.m.,
in response to message #5 by Paul Brogger

Hi Paul,

But isn't most of the memory in these calculators used up in the built-in functions and routines? In other words, this 4k may not be any more than the existing memory in the 32Sii?



Re: Not 4K (but I'm no expert)
Message #7 Posted by Paul Brogger on 20 Aug 2003, 11:52 a.m.,
in response to message #6 by bill platt

I understand that "secret" registers are used for STAT & complex number operations, and probably for SOLVE, but I would think only a handful.

Remember (for example) in the Woodstocks, the statistics features took up 6 registers that were otherwise made available for general use. I think as memory got cheaper, these registers were simply allocated in the background, and dedicated to statistics uses to simplify the user interface.

I assume similar things are done with SOLVE, etc. But again, that shouldn't take too much memory.

But then, what do I know? Maybe they've implemented a huge jump table in RAM to allow dynamic patches to the OS Kernal and support viruses & worms . . .

Re: Not 4K (but I'm no expert)
Message #8 Posted by bill platt on 21 Aug 2003, 12:23 p.m.,
in response to message #7 by Paul Brogger

Hi Paul,

Somewhere here recently someone posted about how much ROM is used up in implementing the machine's functions (like sine cosine etc) I think it was Gene who posted.

Anyway, what I am guessing is that much of that 4 kB would be used up in ROM as opposed to user RAM. Am I confused?

BTW, just read your contribution regarding taking apart Pioneers----so, how is your black stealth 14-BS doing?!



Re: Stealth 42s
Message #9 Posted by Paul Brogger on 22 Aug 2003, 4:24 p.m.,
in response to message #8 by bill platt

I was assuming the 4K referred to RAM, and that the ROM was a different number. If they're combined as 4K either/or, I think we're not going to see much user memory in a 33s! Someone probably has some idea how much ROM it takes to implement something like a 32s/33s -- I 'll bet it's a LOT more than 4K!

I have two Dr. Frahnkenshteen HP-42s'. One is all black, upgraded to 32K, with some sort of keyboard or power-short problem. IIRC, it randomly resets itself, quickly uses batteries and dies. The other is a 32K model too, in a 17BII with most keys and the faceplate black, but the numerals, operations, and a few other shared keytops exposed. It's doing great.

In addition, I bought 42s with a sticky keyboard a while ago. A few attempts at un-sticking it don't seem to have worked. Way down on my projects list is performing "keyboard surgery" to replace its internals, and documenting the whole process with pictures. But that's just not going to be done anytime soon.

Re: Stealth 42s
Message #10 Posted by John Moore on 22 Aug 2003, 7:16 p.m.,
in response to message #9 by Paul Brogger

THe SPLB31A CPU has 256K ROM and 4288 bytes of RAM. It can address up to 4 megabytes externally. It has all the hardware needed for a display and keyboard. Using just the CPU would make the calculator cheap to build, adding external chips puts up the cost.

The calculator must have a least 4K-bytes or 32K-bits of RAM. It could have RAM external to the CPU , though I doubt this. It has 256K of ROM, more than enough for all the functionality

Re: Stealth 42s - A guess...
Message #11 Posted by Vieira, Luiz C. (Brazil) on 23 Aug 2003, 3:43 a.m.,
in response to message #9 by Paul Brogger

Hi, Paul;

I read your post and it made me remember I want to catalog the available 32KRAM chips for the HP42S/HP42S+ upgrade. BTW, have you tried to remove/replace the 32KRAM chip in the all-black, self resetting Frahnkeshteen? The high-current drain made me think of RAM chip as the cause. Just a guess...

In time: do you know where did I find related 32KRAM with ID code close to the ones I used in some HP42? Ancient hard-disk controler board (two so far, one with two 32K IC's and another with only one), an old sound board (with three IC's, two already used) and a controler ISA-type board for a hand-pulled, colored Genius scanner. The only board I could remove the IC's is the sound board; the others are still in use. As I mentioned, I'll collect data and post later.

Best regards.

Luiz (Brazil)

[ Return to Index | Top of Index ]

Go back to the main exhibit hall