IC's for simple calculator applications?

[Chaos Master]

hi.

Are there any cheap IC's around, to do the simple calculator operations (add,
subtract, multiply, divide, square root)? Or is a microcontroller better to do
this?

thanks!
--
Chaos Master® | "I'm going under,
Posting from Brazil! | drowning in you
Win 98 + BB4Win | I'm falling forever,
Slackware Linux | I've got to break through"
---------------------. -- Evanescence, "Going Under"

 

[Tim Shoppa]

Are there any cheap IC's around, to do the simple calculator operations (add,
subtract, multiply, divide, square root)? Or is a microcontroller better to do
this?

If this was a quarter-century ago, then the AMD 9511 would fit the bill.
(Well, they were never cheap, I remember them listing for $100+.)

Do you need a keyboard/display driver interface? If so, then you really
just want a caclulator chip, and you get the keyboard and display driver
and display and case for free for as little as 49 cents.

If not, you get a general-purpose CPU driven
by your computational and interface needs (precision, range, floating-point
vs fixed point vs BCD, ops/sec, etc.)

Tim.

 

[Tim Wescott]

hi.

Are there any cheap IC's around, to do the simple calculator operations (add,
subtract, multiply, divide, square root)? Or is a microcontroller better to do
this?

thanks!

What are you trying to do? If you want a calculator, I suspect that
those chips are custom-made by the same folks who make calculators.

If you want a machine to do the calculations then a microprocessor (even
a little bitty 8-bitter) is going to be _much_ faster than a calculator.

 

[Spehro Pefhany]

hi.

Are there any cheap IC's around, to do the simple calculator operations (add,
subtract, multiply, divide, square root)? Or is a microcontroller better to do
this?

If you're making a calculator, with display and keyboard, then a
calculator chip is extremely cheap in high volume (and otherwise
pretty much unattainable) and will do that trick (and nothing else).

A suitable microcontroller (pretty much any of them with a few K bytes
of code memory or more) can be programmed to do those operations, and
if you use a compiler (and can live with the generally lower number of
digits and other issues associated with the calculations being done in
binary rather than BCD) you may be able to avoid much of the
programming.


Best regards,
Spehro Pefhany

 

[Tim Wescott]

If you're making a calculator, with display and keyboard, then a
calculator chip is extremely cheap in high volume (and otherwise
pretty much unattainable) and will do that trick (and nothing else).

A suitable microcontroller (pretty much any of them with a few K bytes
of code memory or more) can be programmed to do those operations, and
if you use a compiler (and can live with the generally lower number of
digits and other issues associated with the calculations being done in
binary rather than BCD) you may be able to avoid much of the
programming.


Best regards,
Spehro Pefhany

Or you can use a compiler and do the calculations in BCD anyway -- but
why not just buy a calculator?

 

[Chaos Master]

Tim Wescott([email protected]) showed us the following results:

Or you can use a compiler and do the calculations in BCD anyway -- but
why not just buy a calculator?

Just curious about what IC's are/were used for calculators.

I have plenty of calculators, and a lot of calculation software (Excel, GNUPLOT,
'bc' - Linux command-line calculator - good when I am doing stuff on a Linux
console, HP-48GX/49G calculator emulators, and others) on my PC.

[]s


--
Chaos Master® | "I'm going under,
Posting from Brazil! | drowning in you
Win 98 + BB4Win | I'm falling forever,
Slackware Linux | I've got to break through"
---------------------. -- Evanescence, "Going Under"

 

[Tim Shoppa]

A suitable microcontroller (pretty much any of them with a few K bytes
of code memory or more) can be programmed to do those operations

I would think that a few kbytes would be enough to implement a full-fledged
scientific calculator. A 4-banger could probably be made to fit in less
than 512 bytes and the bulk of the code would be keyboard decoding and
display driving.

Tim.

 

[Spehro Pefhany]

I would think that a few kbytes would be enough to implement a full-fledged
scientific calculator. A 4-banger could probably be made to fit in less
than 512 bytes and the bulk of the code would be keyboard decoding and
display driving.

Tim.

Sure. Reasonably tightly coded in assembler (which isn't that big a
deal for four functions plus square root). OTOH, linking in a compiler
floating-point math package and/or using printf() "can" make short
work of 4K bytes before you even touch math.h or equivalent. 8-(

In the PIC line, using assembler, the PIC16F72 (28 pins, 2K x 14)
would probably be a safe (but perhaps not optimal) choice for a
one-chip LED calculator (which could use RPN). The cheapest with 28
pins looks like the 16C55A, which would save about 38 cents in 100's
(18%) but has the 12-bit core, only 512x12 code memory, is OTP not
flash and has only 25 bytes of RAM rather than 128, and might need an
extra external part or two for the clock.

This might be a good project for a student (make an 8-digit LED
calculator with 4 functions, square root and memory, with a nice RPN
stack).

There are far fewer choices if you want to drive a typical multiplexed
LCD display and want flash. TI's MSP430 line would probably be
attractive.

An even nicer project would be to make an RPN calculator using a 20x4
VFD or organic LED display.

Best regards,
Spehro Pefhany

 

[Boris Gjenero]

Tim Wescott([email protected]) showed us the following results:

Or you can use a compiler and do the calculations in BCD anyway -- but
why not just buy a calculator?

Just curious about what IC's are/were used for calculators.

I have plenty of calculators, and a lot of calculation software (Excel, GNUPLOT,
'bc' - Linux command-line calculator - good when I am doing stuff on a Linux
console, HP-48GX/49G calculator emulators, and others) on my PC.

[]s

The few cheap non-programmable ones I've examined seem to use ASICs.
TI programmable calculators use CPUs (not microcontrollers). The
TI-85 uses a Z-80 and the TI-89 and 92 use a Motorola 68000. I
suppose they're always CMOS versions of the CPUs due to power
consumption. You can program the TI graphing calculators in assembler
if you want and there are a lot of programs out there. See
http://www.ticalc.org .

I've never used an HP programmable calculator but I'm curious... so
Google. Interesting, it seems HP developed a CPU for their
calculators. It's called Saturn.
http://www.rhoads.nu/bjorn/hp48/starinst.html

 

[Tim Shoppa]

An even nicer project would be to make an RPN calculator using a 20x4
VFD or organic LED display.

That's sounding a lot like my HP 9100A, circa 1968. Of course it didn't
use any IC's, and used a CRT:

http://www.hpmuseum.org/hp9100.htm
http://www.hpmuseum.org/tech9100.htm

Tim.

 

[Eric Smith]

Are there any cheap IC's around, to do the simple calculator
operations (add, subtract, multiply, divide, square root)? Or is a
microcontroller better to do this?

A suitable microcontroller (pretty much any of them with a few K bytes
of code memory or more) can be programmed to do those operations

I would think that a few kbytes would be enough to implement a
full-fledged scientific calculator.

I should hope so! The world's first handheld scientific calculator,
the HP-35, had only 768 words of ROM. The ROM was 10 bits wide, and
one word was unused, so that was equivalent to 958 3/4 bytes.

The HP-35 did not have some features people would expect now, such as
selectable angle mode (degrees vs. radians) and engineering notation,
nor did it have hyperbolic trig functions, factorial, combinations and
permutations, and statistical functions. But it did have all the basic
arithmetic, logarithmic, exponential, trigonometric, and inverse
trigonometric functions.

Nonpareil includes a microcode-level simulator, a microassembler, and
reverse-engineered source code of the HP-35 ROM:

http://nonpareil.brouhaha.com/

Nonpareil also simulates various other HP calculators, including the
HP-45, HP-55, HP-32E, HP-33C, and HP-41C.

Currently Nonpareil runs on Linux. I've partially completed a Windows
port, which I will offer for sale once it is debugged.

 

[Spehro Pefhany]

That's sounding a lot like my HP 9100A, circa 1968. Of course it didn't
use any IC's, and used a CRT:

http://www.hpmuseum.org/hp9100.htm
http://www.hpmuseum.org/tech9100.htm

Tim.

I HAD ONE OF THOSE!! Bought it used ex-JPL in a stationery store in
Downey CA (home of the oldest McDonalds in the world). Just exuded
quality from the gold-plated PCBs to the core memory. It "disappeared"
during a move, unfortunately. It still worked perfectly, especially
the keyboard, although the CRT was suffering from a bit of burn-in.

Is there any way to duplicate the quality of that keyboard in a
one-off?

Best regards,
Spehro Pefhany

 

[William P.N. Smith]

[Since the topic is drifting that way:]

http://power48.mobilevoodoo.com/ is a port of Emu48 for the Palm that
emulates three different HP48 calculators

 

[Wing Fong Wong]

Tim Wescott([email protected]) showed us the following results:

Or you can use a compiler and do the calculations in BCD anyway -- but
why not just buy a calculator?

Just curious about what IC's are/were used for calculators.

I have plenty of calculators, and a lot of calculation software (Excel, GNUPLOT,
'bc' - Linux command-line calculator - good when I am doing stuff on a Linux
console, HP-48GX/49G calculator emulators, and others) on my PC.

[]s

The few cheap non-programmable ones I've examined seem to use ASICs.
TI programmable calculators use CPUs (not microcontrollers). The
TI-85 uses a Z-80 and the TI-89 and 92 use a Motorola 68000. I
suppose they're always CMOS versions of the CPUs due to power
consumption. You can program the TI graphing calculators in assembler
if you want and there are a lot of programs out there. See
http://www.ticalc.org .

I've never used an HP programmable calculator but I'm curious... so
Google. Interesting, it seems HP developed a CPU for their
calculators. It's called Saturn.
http://www.rhoads.nu/bjorn/hp48/starinst.html

The newest hp's uses an ARM processor.
--


Wing Wong.
Webpage: http://wing.ucc.asn.au

FAQs about me:
Is this related to homework?
Definately not
Are you and undergrad?
Yes
Are you looking for work and do you want to work for us?
Yes, but only if it pays.
Are you insane?
No, not at the moment.

 

[Sergiusz Roszczyk]

I have plenty of calculators, and a lot of calculation software (Excel, GNUPLOT,
'bc' - Linux command-line calculator - good when I am doing stuff on a Linux
console, HP-48GX/49G calculator emulators, and others) on my PC.

[]s

The few cheap non-programmable ones I've examined seem to use ASICs.
TI programmable calculators use CPUs (not microcontrollers). The
TI-85 uses a Z-80 and the TI-89 and 92 use a Motorola 68000. I
suppose they're always CMOS versions of the CPUs due to power
consumption. You can program the TI graphing calculators in assembler
if you want and there are a lot of programs out there. See
http://www.ticalc.org .

I've never used an HP programmable calculator but I'm curious... so
Google. Interesting, it seems HP developed a CPU for their
calculators. It's called Saturn.
http://www.rhoads.nu/bjorn/hp48/starinst.html

The newest hp's uses an ARM processor.

My HP49G uses TI cpu. You can find photos of cpu by google.

Sergio

 

[Tim Shoppa]

It still worked perfectly, especially
the keyboard, although the CRT was suffering from a bit of burn-in.

One could imagine a homebrew calculator with a vector-driven scope tube
for a display... the A/D and slope generation is trivial with modern
electronics. The Chinese make many classic round scope tubes, but
getting the funky rectangular front shape may be harder.

Is there any way to duplicate the quality of that keyboard in a
one-off?

Good keyswitches aren't all that hard to come by, but the molded-in
markings that HP used are probably impossible to duplicate on a small
scale.

There may be something similar that's possible through CAD/CAM engraving
followed by a filling plastic. There are a few vendors that produce
small-quantity instrument panels using this method and what I've seen at
trade shows looks (and feels) really nice. It probably doesn't stand up
to the decades like the HP molding though.

I liked the *scale* of the HP9100's buttons. No finger can feel fat on
that keyboard. Today's GUI designers lose out by making point-and-drool
interfaces when a good button panel is vastly more efficient. (Where I
work there's a project to replace a mimic/switch panel that's roughly 8 feet by
4 feet with LCD screens... I feel it is completely misguided although it
had all the necessary buzzwords to get funding.)

Tim.

 

[Chaos Master]

Boris Gjenero([email protected]) showed us the following
results:

consumption. You can program the TI graphing calculators in assembler
if you want and there are a lot of programs out there. See
http://www.ticalc.org .

I know TI's, but I love RPN and prefer HP's.
I consider TI's being more of small computers.

I've never used an HP programmable calculator but I'm curious... so
Google. Interesting, it seems HP developed a CPU for their
calculators. It's called Saturn.
http://www.rhoads.nu/bjorn/hp48/starinst.html

HP's use CPU's, and the most recent HP-49G+ use a 75MHz ARM CPU. I've readed
that they allow you to use SD cards but I've not yet found one in any shop on my
city, so I have not looked at it. :/

--
Chaos Master® | "I'm going under,
Posting from Brazil! | drowning in you
Win 98 + BB4Win | I'm falling forever,
Slackware Linux | I've got to break through"
---------------------. -- Evanescence, "Going Under"

 

[Eric Smith]

My HP49G uses TI cpu. You can find photos of cpu by google.

No, it's an HP CPU. It was fabricated by NEC under contract to HP.