The Museum of HP Calculators

HP Forum Archive 11

[ Return to Index | Top of Index ]

Calculators I would market
Message #1 Posted by Ron Ross on 17 Mar 2003, 11:26 a.m.

I feel that Hp has missed the boat (BIG TIME).

I think Hp should market 3 types of calculators (scientific).

The first would be a low end calculator like the recently discontinued Hp20s with RPN/Algebraic option. This was a quality calculator that could have just as easily been released in RPN and the price should be about $30. Give it plenty of RAM for a low end unit ie 1-8K of RAM. More is a waste since you want to keep it cheap so no I/O or real programming support.

Next is the calculator I would Have to Have. A full blown POCKET calculator. All the functions and features of a 48G with 128K RAM, but no graphics, to keep it the size to fit in a standard shirt pocket, ie a 42sX type unit. If you could squeeze in an Hp28s type display GREAT. If not, Don't. But give it everything else. I/O, And RAM. 128K or even 1 Meg with it being able to run 48G type software with little or no modifications (of course not the graphics). Keep everything compatable with the 48G if possible aside from graphics. Make it definitely RPN, but with this being such an all powerful pocket calc, maybe algebraic as well. Price $100-200, I would probably even pay more.

Third calculator, would be to re-release the 49G with new Hardware, ie new keyboard (more like the 48 with a thinner form factor) and a faster CPU. Double or triple speed so that it can compete with Ti's Motorola CPU. Price has to be less than $200 to compete with any school graphics.

None of these suggestions require any large expendurture of R&D from HP. These would be quick turnarounds. And I think would grab quick market share. Would Hp ever gain market control?? No, I do not believe that is possible while selling quality as todays calculator market is dominated by a disposable mentality of calculators. Many people want a $10 calculator so that if it is lost or broken, it is not a real loss. HP calculators were not ever supposed to be in that market. (At least not MY HP's)

      
I agree, except on one point...
Message #2 Posted by Jeremy on 17 Mar 2003, 12:43 p.m.,
in response to message #1 by Ron Ross

The middle model should be less than $100. This is fully possible with HPs muscle. If it cost more than $100, people would compare it to the TI86 and HP48G and it would never sell. (except to us... ;)) -Jeremy

      
Re: Calculators I would market
Message #3 Posted by Tom (UK) on 17 Mar 2003, 1:08 p.m.,
in response to message #1 by Ron Ross

All the suggestions use code that has to run on a Saturn processor. This is (or will be very soon) obsolete.

So HP will have to either:

1) Redesign the Saturn using a modern semiconductor process

2) Re-write the code to run on another processor

3) Have an emulator running on a faster processor

I wish they would do one of the above (my preference would be '3') but all of them will take a fair amount of time and money - something HP have not been famous for lately.

            
Re: Calculators I would market
Message #4 Posted by Chan Tran on 17 Mar 2003, 2:24 p.m.,
in response to message #3 by Tom (UK)

I disagree with Jeremy on price. I don't think HP can ever steal customers away from TI by selling calculators at lower price. If the calc is of very high quality I think $300 isn't too much. I think HP users are used to that kind of prices like the 65,67,41,48 etc....

                  
A comment for Chan...
Message #5 Posted by Jeremy on 17 Mar 2003, 6:40 p.m.,
in response to message #4 by Chan Tran

I respect what you're saying, but many of the best future consumers of HP calculators simply cannot afford that much. In a market where a high school student can get a perfectly functional Casio for $12, $100 + is too much for a scientific calculator. (especially one that doesn't graph) Maybe it's not too much for you, but for many a struggling high school or college student, $300 is way too much. I suppose you could argue in favor of the cheaper RPN/algebraic model in this case... But then it might not have all the requisite features. It's tough to say, I guess. -Jeremy

                        
Re: A comment for Chan...
Message #6 Posted by Steve Hanson on 17 Mar 2003, 10:51 p.m.,
in response to message #5 by Jeremy

You might be surprised.

I recently bought my first luxury car after 35 years of driving and then I found a forum dedicated to it. Within the first week, someone posted a poll that just said "what's your age" and the options were:

Under 18, 18, 19-20, 21-22, 23-25, or over 25.

I thought that was ridiculous - where would 25 year-olds come up with that kind of money? Well the top answer was 18 and I was the ONLY vote in the over 25 category.

There are a lot of kids (or parents) with a LOT of money. Of course not EVERY kid maybe not even most kids but I don't think HP needs to sell to everyone. Right now I'd settle for SOMEONE.

                  
Re: Calculators I would market
Message #7 Posted by Christof on 17 Mar 2003, 11:32 p.m.,
in response to message #4 by Chan Tran

I think that something like a 42X (maybe running full rpl, maybe running an updated keystroke 42 like model) with 256-1024k, clock, and the display capabilities of the 28S cna be done. probably at 4mhz. probably sold in a pioneer form factor for under $120.

what do you do about the more than $30, but less than $120 range? an 8k 32sii variant. you ahve to be able to do it for $65, though.

on the low end, I don't think you can do a $30 calc with 8K RAM. I really don't- not with any sort of decet quality. the 20S was fairly priced, and with technological advances can probably be bumped up to rpn/alg switchable, some solver ability, and more ram- but maybe closer to 2k.

But let's look at it- 99steps of fully merged keystroke programming? give it a stack (make it RPN) and for $30 the 20SII would be the equivalent of any pre 67. it's not that bad, even with 99 steps. Since you have to keep the simpler LCD to do that price range (probably), anything over about 300 steps is just too bothersome to program. give me alphanumerics and it's a different story.

            
Re: Calculators I would market
Message #8 Posted by Monte Dalrymple on 17 Mar 2003, 2:51 p.m.,
in response to message #3 by Tom (UK)

Option 1 could be done in 6 months and for about $200K in the absense of a bloated beaurocracy (sp?). I would stay away from the bleeding edge of technology and use a .35u gate array, giving a packaged chip cost of less than $1.50. The advantage of the Saturn architecture is it's power consumption characteristics, but I expect that it would run in excess of 60MHz in this kind of technology if you really wanted speed. If there were more documentation available and I didn't have three kids in college I might even start this as a back-burner project (but put it in an FPGA to start with).

                  
Re: Calculators I would market
Message #9 Posted by Masao Kinoshita on 18 Mar 2003, 9:16 a.m.,
in response to message #8 by Monte Dalrymple

I don't know about this Forum designing its own calculator. I remember back when Richard Nelson left PPC and started CHHU. All he (we that followed him to CHHU) wanted was a users' group, not a business to compete with HP. I don't think HP will ever ask Dave to do anything again like sending RPN manuals to Kinpo, if they felt this Forum could become their competitor.

Deja vu

PPC #8585, CHHU #123, HPCC #1031

Dean Schwartz denied me an HP 71B back then. Then my parents gave me a Casio FX-802P. I was grateful and all, but ...

                        
Re: Calculators I would market
Message #10 Posted by Masao Kinoshita on 18 Mar 2003, 9:59 a.m.,
in response to message #9 by Masao Kinoshita

And Paul Ceruzzi gave me his Sharp PC-1250. This was one reason I decided against Seiko's wrist terminal with the Z8 BASIC programmable base station. I didn't want to have to carry around a mini-keyboard, but still wanted to be able to input info on the fly. Casio's Data Bank watches, though more limited, let you do this.

                        
Re: Calculators I would market
Message #11 Posted by Monte Dalrymple on 18 Mar 2003, 12:45 p.m.,
in response to message #9 by Masao Kinoshita

I wasn't proposing to start a group project. If HP is reading this I'd be happy to talk to them. Designing microprocessors is what I do for a living.

            
Re: Calculators I would market - comments about emulation
Message #12 Posted by Bill Wiese on 17 Mar 2003, 4:46 p.m.,
in response to message #3 by Tom (UK)

Tom & folks...

> Tom(UK) wrote: > ...(snip) So HP will have to either: > 1) Redesign the Saturn using a modern > semiconductor process > 2) Re-write the code to run on another processor > 3) Have an emulator running on a faster processor > I wish they would do one of the above (my preference > would be '3') but all of them will take a fair amount > of time and money - something HP have not been famous

Option 3 is not so bad. Saturn is probably quite similar; a 68HC05 8051+4K ROM could do an emulation of 'Nut architecture and still have room for some application code and a small hardware-abstraction layer.

If function call overhead minimized (steals speed) this could even be done in C on such a microcontroller. Prob some of the code structure/flow of the HP45 emulator on this website could be used!!

Doing a professional job of this (i.e, planning ahead to not back oneself into a corner, plan for future changes, doing test suites, documentation, etc.) might be a 2 month job - quite a bit faster to just do a one-time hack to get something out.

I have a nice Rockwell 350 desktop sci calc (algebraic) from mid-70s. VFD and nice keyboard and fullsize keys - but it's inaccurate as hell. I think I can find a double-size keytop that says "Enter" and that fits this (standard keyswitch) keyboard - and redo this into an RPN monster with an 8051. (I don't have to worry about power consumption here.) One of these days when I have time - reall soon now :)

As HP calcs' innards start to fail more & more - even though physical case, KB, displays may be OK - putting an emulator inside a microcontroller may be the way to go.

Bill Wiese San Jose, CA

                  
What should the price points be?
Message #13 Posted by Scuba Diver on 17 Mar 2003, 11:12 p.m.,
in response to message #12 by Bill Wiese

Someone mentioned a 3-calc scientific line. If the first one is cheap, and has RPN/Algebraic, then it would attract a large population, namely high school students. High school students don't have much disposable income, and this calc would meet all of their needs (let's be honest, high school math and physics do not require powerful calculators!) Another nice feature would be a built-in RPN tutorial; it would allow people to learn RPN at their leisure if they don't already understand it.

The other two units could easily be priced at $100 and $300 accordingly; the market there would be university and professionals.

You can argue that university students don't have much disposable income (When I went to school, Kraft Dinner was a staple of my diet!) however, I know many people who bought 28Ss and 48Ss at the campus bookstore.

People would have an expectation that the higher end units would have the old HP feel to them; they would have to be rugged, and well designed. And don't forget the documentation! Concise, yet comprehensive. Everything you need to know, plus loads of stuff you may look up someday.

And, finally, revert back to the old serial numbering system (eg. 3424AXXXXX)! The new protocol (ID1XXXXXXX) mandates a maximum shelf life of 10 years for a product (otherwise the serial numbers would start repeating!)

B.

            
Re: Calculators I would market
Message #14 Posted by Christof on 17 Mar 2003, 11:34 p.m.,
in response to message #3 by Tom (UK)

the z80 still lives....

yeah, redoing the saturn with modern chipmaking techniques would be nice.

but the 48GX isn't functionally much slower than the ti89 in anything except graphics. it's not that bad :)

                  
Re: Calculators I would market
Message #15 Posted by Eddie Mabillard on 18 Mar 2003, 9:45 a.m.,
in response to message #14 by Christof

hello HP fans, very glad to read this serie "calculator I woukd market", as I already wrote before, the entry one should be cheap and RPN only, once those kids having tasted the power and ease of calculation of RPN will be ripe to put big bucks for the old HP quality whith the new capabilities of modern hardware, and then HP could gain a big share of the market of calculators. Eddie

                  
Re: (Voice from the Wilderness) The Z-80 is AVAILABLE today . . .
Message #16 Posted by Paul Brogger on 20 Mar 2003, 12:04 p.m.,
in response to message #14 by Christof

. . . if you haven't yet, consider reading this. (Especially, the paragraph beginning, "Also, ..." and thereafter.)

I keep promising to shut up about this, but it seems everyone is circling about and yet not seeing a very nearly complete solution, and with an underlying air of futile resignation -- perhaps just because of the label on the hardware. That, for me, is the most frustrating aspect of all! I mean, you're allowed to do so, but if what you want is a FULLY configurable/programmable calculator platform with a computer link and (relatively) open, currently-supported technology, it's not HP, but it's just ~$100.00 away . . .

If anyone cares to rant at me in private (using non-Forum-compatible language, for example), take the "NoHormel" (i.e., No Spam) out of my email address above, and fire away. Meanwhile, I'll go to my Nagger's Anonymous meeting and promise (again) to get back "on the (be QUIET!) wagon". (I'm sorry -- I can't help myself . . .)

                        
Re: (Voice from the Wilderness) The Z-80 is AVAILABLE today . . .
Message #17 Posted by Bill Wiese on 20 Mar 2003, 4:09 p.m.,
in response to message #16 by Paul Brogger

Hi...

Yes, newer TI calcs (TI-8X) can indeed be re-flashed and reconfigured. Their Z80 architecture is reasonably speedy, and it could in fact be purged & replaced with a Coconut emulator that'd run stuff roughly as fast as a 41C, maybe 42S.

Unfortunately the real issue is not software. We can pretty much, if there's a way to introduce new program firmware into a handheld device, make "anything" be an HP. Whether that's done in BASIC, C, Tiny-C, or assembly is moot, as long as it's fast enough.

The problem is the physical embodiment and 'making a nice product' (well, device) of it. You have to deal with keys and keytops, legends above keys, etc. You'd really need to have a double-sized [ENTER] key too.

RPN on a TI platform without appropriate key layout isn't a real solution. HP spent some nice time figuring key layouts, spacings, etc for each line of calcs and did a great job - look how they've stood the test of time.

Frankly, if we found someone that could injection-mold keys and silk-screen a metallic KB overlay/mask we could buy HP12Cs, put in a new CPU (based on a low-power 8051 w/HP emulator firmware). Then we'd have a nice robust HP. Heck, we could even add a 10 or 12 char ASCII matrix display as well.

It's all about the packaging & form-factor now: we can handle the software/firmware readily in a variety of ways.

Bill Wiese San Jose, CA

                              
Re: (Don't encourage him!)
Message #18 Posted by Paul Brogger on 20 Mar 2003, 7:49 p.m.,
in response to message #17 by Bill Wiese

Soooooo . . .

Rather than a full-blown hardware and software development effort, a gonzo RPN calculator creation approach might consist of:

Software -- Implement RPN OS with following features:

  • four-level stack, Roll up/down, swap, Last x , etc. with appropriate display behavior and keyboard mapping
  • connect operation of all existing math functions to RPN/stack context. (e.g., y^x function is now called as a strictly postfix, rather than an infix operation)
  • implement storage register (A-Z) access with STO/RCL “storage register arithmetic” (RCL + x, etc.)
  • create keystroke programming environment (entering, editing, and executing) with alphanumeric routine and variable naming, subroutine call & return, SST, etc., etc.
  • implement ability to load and save programs & data to PC (TI has already implemented something like this, but seems to maintain an obsessive degree of control over aspects of application “signing” and permissions)
  • (there is certainly more . . . )
Hardware -- Given (for example) a working TI-83+:
  • disassemble case front and keyboard
  • cut path between two chosen keys
  • epoxy a new “enter” key over two existing keys (or replace with a newly-molded unit)
  • paint the front (& back?) case and keys
  • silkscreen new keyboard and keytop legends
  • (is there anything else???)
  • reassemble
  • install latest version of new RPN OS & test

The above is certainly non-trivial, but isn't it a much shorter list of doable details than it would be if one were to start from scratch? By working with a system that has most of the key elements already packaged together, couldn't the whole job be made substantially simpler and more realistic?

I suppose your comment about retrofitting an HP-12C approaches the same idea from a different starting point -- but the 12C doesn't have PC connectivity, and (probably because I'm intimidated by electronics design) that seems to be a rather more significant drawback than the lack of a big "Enter" key . . .

Just some thoughts . . .

                                    
Re: HP coated keys are getting better
Message #19 Posted by Frank Travis on 21 Mar 2003, 8:26 a.m.,
in response to message #18 by Paul Brogger

A few years ago I called HP's calculator tech support number in Oregon. The person who answered said that although HP no longer injection molds the symbol through the keys, the coatings on the hard plastic keys of HP calculators such as the HP 48GX have gotten much better than before. I have not had any key symbols wear out on any of my several HP 48GX calculators(newer or older). If this is the case, HP could still save on cost, yet produce a high quality keyboard.

                                    
Re: (Don't encourage him!)
Message #20 Posted by Bill Wiese on 21 Mar 2003, 3:09 p.m.,
in response to message #18 by Paul Brogger

Paul..

> I suppose your comment about retrofitting an HP-12C > approaches the same idea from a different starting > point -- but the 12C doesn't have PC connectivity, > and (probably because I'm intimidated by electronics > design) that seems to be a rather more significant > drawback than the lack of a big "Enter" key . . .

Nice looking cases & KBs, etc are harder to do and can cost more than electronics!!! Right now, the cost of assembly, packaging (cardboard box + manual), device packaging (case + keys + PCB) exceeeds the cost of basic calc CPU + LCD.

For mfgrs using off-the-shelf chips there's no NRE for electronics. But molds, etc. for quality-appearing product can run into lots of dollars just to get Product One out the door.

Some PROM microcontrollers can be had for $1-$2.50 in very small quantities (TI MSP430 for example). In even small mfg qtys much much less.

Bill Wiese San Jose, CA

      
I need new HP 200LX !!!
Message #21 Posted by mapet on 18 Mar 2003, 4:23 a.m.,
in response to message #1 by Ron Ross

I agree on HP20S and HP42S but the third one should be a HP 200LX like calc with VGA color display and Intel 486 CPU with 16 MB RAM.

            
Re: I need new HP 200LX !!!
Message #22 Posted by Mike (Stgt) on 18 Mar 2003, 5:32 a.m.,
in response to message #21 by mapet

If you are good in soldering you may build your own. See http://member.nifty.ne.jp/toyozou/palmpc/release_English.html

Ciao.....Mike

            
Re: I need new HP 200LX !!!
Message #23 Posted by Erik Ehrling (Sweden) on 18 Mar 2003, 12:09 p.m.,
in response to message #21 by mapet

...and running Linux...

            
Re: I need new HP 200LX !!!
Message #24 Posted by Christof on 18 Mar 2003, 11:59 p.m.,
in response to message #21 by mapet

Why bother with a 486?

the jornada 720 or 728 have everything you need to replace a 200LX- except one thing. a real OS. the ARM is supported by many real OS flavors that cna provide you with the software you need. In rough order of suitability:

QNX

NetBSD

Linux. (yeah, I'm a heretic, I put linux last. it's really not all the hot for this application. too much cruft and bloat and ....stuff!)

It would be harder at this point to redo a 486 than to use a mk arch for a dragonball. (though it'd be nice to have the MMU back, you know).

16MB RAM is... not competitive. if it isn't at least 32MB, no one outside of a 200 fanatic will buy it- and a 200 fanatic will just spend the money to upgrade his RAM and put a backlight on his 200LX (whihc I'm considering, I can't stand the windowsCE on my 720, as much as I like the wireless pcmcia connectivity)

Sharp has a new ARM based linux core 4 inch VGA keyboard palmtop. it's only produced for japan, but some importers are brining it into the US priced at about $600.

(this is sounding cool enough to almost convince me to sell off the 680, 720, both 320LX machines and my 200LX to afford a pair for me and the budding mommy next to me)

the 200LX is, alas, too far behind the modern world to really work in an upgrade type of system. I love my baby, but it's just not going to be succeeded by another DOS based device.

                  
There is no need for a 200LX
Message #25 Posted by Nenad (Croatia) on 20 Mar 2003, 5:12 a.m.,
in response to message #24 by Christof

Just to add a little bit to Christof's excellent point:

The Jornada 720 and 728 would have a funcionality of a 200LX. There is a program, called PocketDOS, which brings up the complete funcionality of your good, old, well tested, reliable DOS programs to your Jornada.

After trying PocketDOS on my 720, I have decided that I would not look for a 200LX anymore.

And, do not forget Jornada 710. It often sells for a much less price than both 720 and 728, but, AFAIK, the only important difference is 710's lack of modem.

      
Re: Calculators I would market
Message #26 Posted by James M. Prange on 19 Mar 2003, 7:23 p.m.,
in response to message #1 by Ron Ross

At least three (all RPN entry, of course) that I think *some company* should market. I'm not convinced that HP will ever give its calculator group (Does it have a name? Is it a "division"?) a fair chance, so maybe a different company will pick up what HP seems to have mostly discarded. On the other hand, HP has dropped calculator development and good (at least they seemed good to me) products before, and then started up development and introduced new products again, so perhaps it'll do it this time, although it's lost a lot of credibility with it's recent models.

1: A very simple "4-banger" but with RPN entry (not the 4-level stack; two levels should suffice for this). Just the basic arithmetic +, -, X, and / keys, and, of course, ENTER, Clear Entry, and perhap Clear or Clear All. Maybe keys for one memory register, change sign, square root, 1/x, and SWAP. Maybe even a % key, although I personally find that about as useful as a screen door on a submarine. Something that's cheap enough but also robust enough to give to elementary school students as their first calculators, before they get indoctrinated into using an algebraic entry calculator. Or to carry in a pocket, or toss in a glove box, toolbox, briefcase, backpack, book-bag, or whatever (oh yeah, a lady's purse! I know a few geeky ladies, and this wouldn't be a geeky calculator anyway) without worrying that it may be too expensive if it gets lost or broken. For that matter, for the company to use as promotional handouts instead of putting their logo on the same calculator that other companies are handing out. Or to have the company's logo molded in somewhere on calculators that other companies are handing out (imagine a TI "promotional item" with HP's logo molded in). If someone "catches on" to RPN entry (and it isn't the least bit difficult, despite its reputation), would they ever go back to algebraic entry? This could have a low profit margin; the main purpose would be to introduce them to RPN entry calculators and doing things the right way; profit would come mostly from future sales of more advanced RPN entry models that many users would eventually want. For an established company with deep pockets (like, I suppose, HP), this could even be a no-profit model.

2: A more advanced "scientific" RPN calculator small enough to fit in a typical shirt pocket and robust enough that it could stand some rough handling, with a reasonable profit margin but still at a reasonable price. I expect that some of HP's discontinued models would be just fine, except perhaps for HP's typical premium price.

3: An "everything but the kitchen sink" RPL model redesigned from the 48 series, which could have a high price and profit margin. Of course, with the high price, one should expect it to be very reliable and robust, and have a long warranty. Put the operating system in flash memory like the 49G so that it can be upgraded, but thoroughly debug it before release; don't take the attitude that it can be released with known bugs and fixed up after the customers complain too much. Add a USB port both for speed and because many new PCs don't have a serial port built-in. Maybe add a port for more or less "industry standard" (relatively) low cost flash (no battery) removable memory, like what's used in some cameras, or perhaps use one of those miniature hard-disk drives. Maybe keep a port for additional system (high speed) memory, if there's any realistic possibility that the built-in system memory might not be enough for some users. Add the useful features from the 49G, but not its flaky keyboard or its habit of not restoring modes and flags at the end of a command or various other bugs. If there's really a need to cover up the display with a chunk of plastic, at least make it user-replaceable. Keep the keyboard overlays like the 48 series for user or third-party customization. Maybe the ability to switch to algebraic entry mode if it can be done without introducing new bugs in RPN mode; but I have my doubts about that one; I'd never use it and I'm afraid that the algebraic entry option encourages those who don't know any better to continue their bad habits.

Of course there's plenty of room for various intermediate and special-purpose calculators. Some of HP's out-of-production models might fit in the gaps quite nicely.

It seems to me that a lot of HP calculator's problems may be from each HP group being unwilling to share technology with other groups. What might HP calculators be like if they shared more technology developed for "computers"? After all, a 41 series (perhaps with HPIL peripherals) or an RPL model seems to me to be very much like a full-fledged "computer", but designed for hand-held use and with a keyboard designed for crunching numbers.

Given that calculators seem to have a very low priority at HP, I wonder whether they could be persuaded to sell off their calculator product line (to someone with a *lot* more money than I have). Or maybe they could spin it off as a separate company that could go it's own way, sink or swim, without interference from HP corporate headquarters; I'm convinced that a good profit could be made on RPN calculators. On the other hand, if HP gets out of the calculator business or continues as a weak competitor, then it may be a golden opportunity for someone else to start up an RPN calculator company, or for a company already established in calculator products to market RPN models as a premium line of calculators.

Regards,
James


[ Return to Index | Top of Index ]

Go back to the main exhibit hall