HP-67 and TI SR-52 compared
#1

A recent reply from Valentin to Stefan inspired me to start a new thread about the comparison of two vintage calculators, the SR-52 and the HP-67. Here is Valentin's statement:

Quote:
(Re the TI SR-52, the underlining is mine) "This calculator had 224 steps of program memory, 20 registers, and a magnetic card reader. It was only slightly less powerful than the HP-67. "

    Absolutely nope. It was much less powerful than the HP-67, not "only slightly". I did own an HP-67 the instant they were released, wrote many, many highly complex programs for it (many of them I did submit to the HP User's Library), and also had the opportunity of being lent an SR-52 for a month or so. The difference in capabilities was tremendous, there was no way most of my best programs could fit there. I would estimate the SR-52 to be no more than 30%-40% as powerful as the HP-67, if at all.

I agree with Valentin. Even if the specs sound similar, they aren't. The main difference between the two machines is that the TI had no merged steps: a simple absolute GTO takes 4 steps in memory, a labeled GTO takes two steps and the label another two steps. Likewise with STO and RCL.

AOS isn't bad, but from a keystroke programmer's point of view, RPN is far more elegant.

The SR-52 has only a limited number of comparison operators. The way they work is different, too: You can only provide a step number or label, so it is always a jump that follows. Unlike newer TI incarnations, the SR-52 has no PSE command so you can only display intermediate results with HLT or print them on the PC-100.

So let's start a more detailed comparison here.

Marcus

#2

I think Valentin is correct in his assessment, and I will reword my web site when I get a chance. It's been a long time since I programmed the SR-52, and I had forgotten how much overhead the non-merged keystrokes took.

I also agree with you that AOS is a pain for keystroke programming. One requirement was that any numerical subroutine should begin with "(" and end with ")", otherwise it might not do what you want when called from within another expression. For example, a subroutine to add the contents of registers 01 and 02 (forgive the syntax, I don't remember all the SR-52 instructions):

LBL
0
RCL
01
+
RCL
02
INV
SBR

If you now type (in run mode, not program mode),

2 x SBR 0 =

you'll get (2 x R01) + R02, not 2 x (R01 + R02) like you probably intended.

Hence the subroutine should be:

LBL
0
(
RCL
01
+
RCL
02
)
INV
SBR

So basically every subroutine has 6 keystroke steps of overhead!:

LBL
xxx
(
...
)
INV
SBR

Contrast this with an HP, which has only three keystrokes (and only two steps) of overhead:

LBL xxx
...
RTN

So I think Valentin is correct.

I guess I remember the SR-52 as being rather powerful compared to my previous programmable, which was a Commodore PR-100 with 72 non-merged steps, no labels, and a single test.

Stefan

#3

My SR-52 is just sitting besides me, so I can correct your program:

2nd LBL
+
(
RCL
0
1
+
RCL
0
2
)
2nd rtn

2nd is always merged. Labels are normal keys, not numbers.

In the TI-58/59 the steps were at least partially merged so that a GTO line number took only 3 steps instead of 4 and a STO only 2.

Edited to add the following information:

The conditional jumps in the SR-52 are:

if flg

if err

if pos

if zro

and their inverses. Why no x/y comparisons? Because the Y register isn't accessible at all! There is no T register as in later models and no x<>y exchange instruction.

Edited: 7 May 2008, 11:29 a.m.

#4

Quote:
Labels are normal keys, not numbers.

True, but if I remember correctly, the calculator had number keys too. Or was it not possible to use these as labels?

Stefan

#5

In 1976, I had the ship on which I served procur an SR-52 and PC-100 for use in the engineering (radiochemistry) lab. It was very useful. When the HP-67 came out in 1977, I bought one for myself and found it to be a tremedously more capable machine, except for lacking the printer capability. I then had my ship procur an HP-97, which was the best thing available for our applications.

I never considered the SR-52, which came out well before the HP-67, to be fairly compared to the HP-67. It should more likely be compared to the HP-65. The leading TI competitor to the HP-67 was the TI-59. I bought one of those, and found it to generally be more capable, faster, and innovative than my HP-67. But it was much harder to program. It was also far less reliable. IIRC, I went through three TI-59s in the first year. They wouldn't last, and the mechanical quality was stinko. So I reverted back to my trustworthy HP-67 until it began seriously misbehaving four years after I bought it. By that time, the HP-41C was available.

#6

Numerical labels are only possible if shifted. This has to do with the way jumps are interpreted by the calculator. If the command (SBR, GTO, if ...) is followed by a digit, two more digits are read and interpreted as an absolute address. If the code after the jump is not a digit, it is a label and searched for as jump destination. This is true for the TI-59 as well. The latter packs the last two digits of the jump address in one step, while the 52 needs a step for each digit.

One more note: SBR in calculator mode does not switch to run mode automatically, you have to hit RUN to start the program. Only the function keys A..F and their shifted counterparts start execution immediately.

#7

Good points!

Printing with the 67 meant buying a different calculator: the 97.

A fair comparison should include the 65.

The mechanical properties of the 52 where better than those of its successors.

#8

I've just checked the HP-65 manual ad must admit that the SR-52 has some more power than the 65.

More steps (233 vs. 100) - The steps in the 65 are only partially merged.

More Registers (20 vs. 9) - The 65 has something like 8.5 registers because R9 is often used internally. R8 serves the same purpose as R00 in the 52 (DSZ).

Subroutines - The 52 can use any label for a subroutine and has a 2 level return stack; the 65 is limited to 1 level and only 5 subroutines.

Flags (4 vs. 2.)

Program editing - The 65 does not show the program location, just the key code.

Advantages of the 65 over the 52:

Branching - The 65 is more versatile than the 52. It has more comparisons and can execute any instruction, not only branch on a conditional.

RPN. :-)

I looks like TI tried to overcome the 65 limitations with their offering but were outperformed by its successor, the 67/97 combo.

Marcus

Edited: 8 May 2008, 7:39 a.m.

#9

Another advantage is the indirect addressing capability of the SR-52 compared to HP-65.

I'm not sure, but I guess the SR-52 is also a bit faster than the 65/67.

#10

Hi, all:

    Visit the Images section of my calc web page to have a look at the six pages of promotional leaflets featuring TI's SR-52 I've included there.

    I've scanned all six pages myself so that they're fully readable, just click on the thumbnail to get the quite large, readable versions.

    Enjoy. Comments welcome.

Best regards from V.

#11

Now it would be really interesting to compare the marketing speech of both companies. Valentin, do you plan to upload readable copies of the HP material any time soon?

#12

Hi, Marcus:

    (by the way, being very mathematically-oriented, I really, really do like your surname, "von Cube" :-) I don't know if it means anything in German or not, but it's really wonderful to my eyes. It would be a real pun for me to have such a surname, can you imagine a "Valentin von Gamma" ? But I digress ...)

Marcus posted:

    "Valentin, do you plan to upload readable copies of the HP material any time soon? "

      It's been an idea of mine since old to scan and upload to some web site some of the many, many calc leaflets I do own, be they HP, TI, Casio, Sinclair, or whatever.

      However, there's the problem of size: to be reasonably readable, the scanned images must be fairly large, and they take a sizable amount of web space. My free web account doesn't have the necessary space and further, images that big are considered 'heavy' downloads and would involve a bandwidth penalty for my calc site which could result in it getting shut offline.

      This being so, if someone can offer a place where I could upload scanned images to be freely available to everyone, I would certainly take the opportunity to do so. The requisites are simple:

      • Sufficient space: around 1 Gb would be adequate

      • full ftp access for me to upload the images, where "full" means I can upload them whenever I want, list them, eliminate them, rename them, etc. at my leisure with no access restrictions to me.

      and that's it. If this can be arranged, I would consider allocating time to scan and upload the images.

    By the way, the six SR-52-related images I uploaded to my site are there only temporarily, I'll delete them within a week or so to make room.

Best regards from V.

#13

Quote:
It's been an idea of mine since old to scan and upload to some web site some of the many, many calc leaflets I do own, be they HP, TI, Casio, Sinclair, or whatever.

However, there's the problem of size: to be reasonably readable, the scanned images must be fairly large, and they take a sizable amount of web space. My free web account doesn't have the necessary space and further, images that big are considered 'heavy' downloads and would involve a bandwidth penalty for my calc site which could result in it getting shut offline.

This being so, if someone can offer a place where I could upload scanned images to be freely available to everyone, I would certainly take the opportunity to do so. The requisites are simple:

  • Sufficient space: around 1 Gb would be adequate

  • full ftp access for me to upload the images, where "full" means I can upload them whenever I want, list them, eliminate them, rename them, etc. at my leisure with no access restrictions to me.

and that's it. If this can be arranged, I would consider allocating time to scan and upload the images.

Best regards from V.


This isn't a problem, since I run my own server and there's plenty of space available. Outbound speed is limited to about 50K/sec, though, since it's on a residential DSL line. Contact me at dramsey at neko.com if you're interested.

-- David

#14

Quote:
More Registers (20 vs. 9) - The 65 has something like 8.5 registers because R9 is often used internally. R8 serves the same purpose as R00 in the 52 (DSZ).

The SR-52 has 22 data registers, but using your reasoning, that is more like 21.5 since as page 51 of Owner's Manual states "... You should exercise discretion in using register R00 because the SR-52 occasionally uses this register for its own purposes, namely during polar/rectangular conversions and DSZ execution. ..."

#15

Touche' !

Whoa, RAM must have been really expensive in the 1970's (of course, we all knew that...)

Still, that's bad design. Imagine storing a number in one of those "shared" registers, then getting an incorrect result after retrieval because the calculator altered the value.

I assume that such pitfalls were not possible on the HP-67 and later models.

-- KS

Edited: 11 May 2008, 12:03 a.m.

#16

Quote:
(by the way, being very mathematically-oriented, I really, really do like your surname, "von Cube" :-) I don't know if it means anything in German or not, but it's really wonderful to my eyes.

No, its is pronounced "Koobe(r)" and is of ancient origin. The "von" has become part of the name in 18 century and is a noble title (no powers or money involved.) I was working as a teacher, lately, teaching about SAP business warehouse. One key concept is that of a multidimensional "InfoCube". I kept telling the people that it had nothing to do with my name.

Quote:
It would be a real pun for me to have such a surname, can you imagine a "Valentin von Gamma" ? But I digress ...)

Wouldn't that be more suitable vor Viktor Toth? ;)

Marcus

#17

Quote:
I assume that such pitfalls were not possible on the HP-67 and later models.

I suppose that the manufacturers take the position that if the user has been fairly warned in the documentation then it is the user's responsibility; for example, page 109 of the HP-67 Owner's Handbook and Programming Guide states "..Unlike storage register arithmetic, the Sum+ function allows overflows (i.e., numbers whose magnitudes are greater than .999999999x10^99) in storage registers RS4 through RS9 without registering Error in the display. ..."

Later machines such as the TI-59, the HP-41, and others use parts of the user memory for the statistical accumulation. Again, it's up to the user to understand. Then a machine such as the hp33s comes along.



Possibly Related Threads…
Thread Author Replies Views Last Post
  HP Prime vs TI : Factoring ? HP Pioneer 7 2,336 10-29-2013, 02:00 PM
Last Post: CompSystems
  [alpha]SR wp34s Andrew Nikitin 9 2,966 06-17-2013, 11:48 PM
Last Post: Andrew Nikitin
  HP-27S speed compared to HP-42S Tommi 3 1,731 06-05-2013, 05:55 PM
Last Post: Christoph Giesselink
  SR-52 flaky display Colin Verrilli 11 3,415 07-24-2012, 01:22 PM
Last Post: Colin Verrilli
  SR-52 vs 15c: more questions of NIMB nina scholz 6 2,123 05-27-2012, 05:59 PM
Last Post: Anders Persson
  Help wanted SR-52 nina scholz 29 7,535 05-16-2012, 07:48 AM
Last Post: Anders Persson
  HP & TI in the 70's revisited Palmer O. Hanson, Jr. 1 1,203 05-12-2012, 02:17 PM
Last Post: John W Kercheval
  OT--HP & TI in the 70s Matt Agajanian 12 3,167 05-09-2012, 01:20 AM
Last Post: Matt Agajanian
  OT--TI's SR-60 Matt Agajanian 8 2,624 04-25-2012, 11:34 AM
Last Post: Joerg Woerner
  Emu28 v1.25 and Emu48 v1.52 update Christoph Giesselink 2 1,176 04-05-2012, 02:01 AM
Last Post: Gilles Carpentier

Forum Jump: