Forgotten Fundamentals: the HP-33S


Forum readers --

This post could be considered tardy, but it's one that's been rattling around a while with me.

I bought a new HP-33S from a retailer for $50 after the April NCEES exam cycle, with its related profiteering, was over. I didn't want to compete with those who needed an RPN calc for the exams, and wasn't about to pay an inflated price from a gouger. In fact, I wasn't even eager to buy one, having seen the photos of its warped chevron-style keyboard. However, the 33S would be useful as a calc without RAM limits that allowed development of programs for my 32SII, so I got one.

Verdict: Quite frankly, I would have been more pleased with a 32SII improved only by 2-4 kB of additional RAM. Although the 33S includes several new and helpful features, it exhibits apalling inattention to HP's time-honored fundamentals of calculator design. It's as though KinHPo "started from scratch" without examining the many other Pioneer models which offered the features added to the 33S.

Consider what was added to the 33S that the 32SII lacked, but other Pioneer models already had:

Two-line display:           42S
AOS with precedence: 20S, 21S, 22S, 27S
RPN/AOS dual mode: 17BII

Or, what could have been added:

Ins/del equation-editing:   27S, 17B, 17BII
Complete complex functions: 42S
One-line complex display: 42S

Any of these models could have served as good examples of thoughtful design.

What's good or "right" about the 33S?

  • Unshifted x2 key (usually omitted from RPN calcs)
  • "INTG" (round to nearest integer) function
  • Scalable ENG display mode (Casios have had it for years)
  • Physical constant list

What's wrong about the 33S? Let us count the ways:

Computational bugs!

  • Rigorous testing methods were not applied.

Keyboard design:

  • The "chevron" design degrades ergonomics, with irregularly-shaped, misaligned keys that don't fit the fingertips naturally. Keys in the middle column are larger than the others, for no practical reason.
  • The four silver, raised keys that are labeled are awkward to press, being irregular and angled.


  • Tiny, virtually-indistinguishable decimal point and comma. Inexcusable!
  • Shadowing when viewed at a sharp angle.
  • Glary, reflective screen. (The 1990 Pioneer upgrades addressed that one.)
  • Tiny, hard-to-read annunciators on the upper part of the display, in the shadows
  • The display instantly changes to name of function when a key is pressed, instead of after a time delay when the key is held down, as on the 32S, 32SII, and 42S. It is disconcerting to watch the display flash each time an operation is performed.

Faceplate legends:

  • The two "shift" colors contrast nicely with the faceplate, but not with each other. This causes me (and others, I presume) to make more mistakes by using the wrong shift key.
  • Poor spacing and cluttered detail -- exacerbated by the crooked layout -- make it sometimes difficult to determine what imprint belongs to which key.

Functional grouping:

  • Remarkable illogic in placement of functions. This aspect of design was carefully planned in HP's masterful Voyagers. There was also a coherent concept in the fully menu-based 32S that was degraded in the partially menu-based 32SII, and has become haphazard in the 33S:
  1. Why isn't "HYP" adjacent to SIN/COS/TAN?
  2. Why are "RPN" and "ALG" buried in the keyboard, and not under the "MODES" menu, as on the HP-17BII?
  3. Why isn't Roll Down" adjacent to x<>y?
  4. Why aren't the probability functions (seed, rand, nCr, nPr) grouped together under a menu, as they were on the 32SII?
  5. Why aren't ABS, IP, FP, RND, and INTG consolidated under a menu named "PARTS" (32S/32SII) or "CONVERT" (42S)?
  6. There's a special "ENG/<-ENG" key. Why retain "ENG" under the "DISPLAY" menu? FIX/SCI/ALL can be used to set precision (the way Casio does it). Then, "." and "," could be moved to the "DISPLAY" menu, thus making room for "RPN" and "ALG" under the "MODES" menu.


On the 33S menus: The 2-line display made possible more selections (as "pick-a-number" or "cursor-and-enter") in the menus, but I prefer the "press-the-softkey" system of the Pioneers, in both 1-line and 2-line displays.

I've noted that I spend considerable time hunting for functions on the 33S, for the reasons listed above -- clutter, visual indistiguishability, haphazard arrangement. The hard-to-read numerical display is a disqualifying offense.

When I eventually take the PE exam, I'm bringing the 32SII. I have also lobbied NCEES to explicitly approve the 11C, 15C, and 20S -- all of which are better-designed computational tools.


-- KS

Edited: 20 Aug 2004, 2:27 a.m.


One complaint I have about the HP-33S, as well as some other HP and TI calculators, is the lack of attention given to the legibility of the labels on the keyboard. My eyes aren't as sharp as they used to be, and I often have difficulty reading the shifted functions on some calculator keyboards. Legibility is affected by the foreground color, the background color, the typeface, and the font size. The white metallic background results in glare and low contrast. I would prefer a dark matte background. The foreground colors should be bright/saturated enough to be clearly visible. The typeface should be selected for high legibility.


With respect to legibility and careful layout of keyboard, IMO HP65 is best. Reasons:

a) Big keys with 2 sides (usually with black/white and blue characters on).

b) Background colour of keys clearly separates different groups of keys.

c) Colour of characters contrasting well to background (on keys and plate). No reflections or other obstacles. Colours of prefix keys showing optimum contrast.

d) Sufficient space between the keys.

e) Superior print quality.

f) Consistent grouping of functions (f, f^-1 TMK was never seen thereafter on HP handheld calculators).

After all, this rating might be earned because it was the first HP with more than 1 prefix key. Usually, you have to present real advantages to convince your management to start something new. Later models (Woodstock and Spice) retained most of these advantages. On Voyagers, IMO print quality was fading (I will be killed for this statement, I know ;-). With first Pioneers, features a), b) and f) were dropped finally. Especially after losing the big keys every prefixed function had to be printed on the plate. Meanwhile there were so many functions making it apparently impossible to end up with a keyboard layout being as user-friendly as before. It really slows you down when you have to look through menus or you have to scan a cluttered keyboard for desired functions. And - quite naturally - all of these calculators are designed for young people with full eyesight - it didn't bother me then, I "see" it now d:^)

Well, just my 0.01 € ...

Edited: 20 Aug 2004, 9:28 a.m.


"f) Consistent grouping of functions 
(f, f^-1 TMK was never seen thereafter on HP handheld calculators)."

I think I prefer the f,g,h of the HP-67 (or the 34C)

<some rant>
The ALPHA on the 48/49 series could be considered as a third shift
When the shift&hold was introduced in a 49G Flash-ROM
there was more than enough shifts.
Wolfgang Rautenberg offers double-click and long-hold
in KeyMan
(as if the shift&hold was not enough :)


Hallo nochmals Karl!

I returned back my HP33s just for some of the problems you exposed.

In my opinion, for display, the litle comma is the most disappointing problem.

But my bigger disappointing subject, was the formula storage and SOLVE. I mean that in other HP calculators (like for instance the even non-scientific HP17BII, HP18C, and so on...) you could use two or more characters for naming variables.

On the 33s you just can use single character variables in formulas as for example:

A = B + C (excuse the simplicity of the example)

but not:


The fact 33s can use just one character variables, made ver hard for me to remember and use stored formulas. Then, if to look for, and put formulas on SOLVER... it is not easy to remember wich variable was which.

As another example, these simple formulas (used in dead-reckoning navigation) must be "converted" to HP33s as it just accepts single character variables!!!.

DLAT = D x R (difference of latitude)
DLON = A / COS( ML ) (difference of longitude and mean latitude)


X (or whatever) = D x R
Y (or whatever) = A / COS( M )

This is by far less descriptive!!!!.

Moreover, you have to add the lack of INPUT's or PRINT's with user messages. As the manual states, if you need to show an input/output message to the user, you need to create "one-string" programs (programs containing the message in a string - and just ONE string-) and call them from other programs!!!!. Wow, even my old HP41 works much better than the latter Hp33s!!!!.

Other manufacturers (don't want to mention brands), which also use single-character variables in formulas, have a solution. They permit the use of a string right after the variable in order to be more descriptive (like for instance: X"LAT"=D"DISTANCE" x COS(R"COURSE"). It is just like a workaround, but it works.

I sincerely hope that HP will "follow the right line" again, as I think their calculators are the best made ever (or should I say "were"?).

Best regards (Viele Gruesse!!!),



"Moreover, you have to add the lack of INPUT's or PRINT's with user messages.
As the manual states,
if you need to show an input/output message to the user,
you need to create "one-string" programs
(programs containing the message in a string - and just ONE string-)
and call them from other programs!!!!.
Wow, even my old HP41 works much better than the latter Hp33s!!!!."

I think you need to use an Equation as a message.
It can be on any line - no need for subprogram call.
You need to set a flag to use an Equation as a message.


Surely there's problem with 33S' design, but what makes me really disappointed is the large no. of bugs that's present in the calculator. It is not like HP calculator's long-developed reputation. I agree with Karl in that I would prefer a 32sII+ with 2-4kb of additional ram than this 33S.



My situation is similar to that faced by Karl Schneider. I am currently preparing for an NCEES PE exam, and so I bought the NCEES-approved 33S a few months ago. I already had a 32SII, but I wanted the additional program memory and equation storage capabilities offered by the 33S.

I have worked extensively with the 33S, and I basically concur with all of Karl's points. I had previously reached the conclusion that a 32SII with more memory, say 4 KB, would have been a better design than the remodeled 33S. The extra memory is the only real advantage of the 33S; I prefer the 32SII in other respects. The 2-line display of the 33S should have been a plus, but in fact the 1-line display of the 32SII is arguably better, for the reasons noted by Karl.

But I differ from Karl in one respect: the 33S, and not the 32SII, will be my primary calculator on the PE exam. While I recognize the idiosyncrasies of the 33S, I am prepared to tolerate them in exchange for the extra program and equation memory. The 32SII will go to the exam too, but only as a backup.

Incidentally, the 2004 NCEES Annual Meeting was held last week. NCEES was considering revisions to their calculator policy, including a proposal to ban all models except those on their "approved" list. However, I have no idea as to the outcome of the meeting; there don't appear to be any updates at the NCEES website.

Possibly Related Threads...
Thread Author Replies Views Last Post
  Integration Times "Old" 33s vs "New" 33s John Smitherman 21 3,775 12-14-2005, 12:04 AM
Last Post: Karl Schneider

Forum Jump: