Building a calculator

[Freeman]

I was discussing with my dad and brother about graphing calculatros when my bro suggested that parts are so cheap that I might as well just build a calculator. Well, I have been thinking about this but does anyone have any suggestions or experience in this field?

[mmalluck]

You could probably make a simple 4 function calculator, but anything more complicated than that will be a real pain, if not impossible.

 

Graphing calculators are one part hardware and another part software. Both can be daunting to figure out and do on your own. You'd have better luck coding a calculator program for the computer. This will give you some idea of the complexity involved.

[CPL.Luke]

the complexity and price of the parts for a graphing calculator will make it not worth your while, as attempting to construct any form of modern calculator would be like trying to build a computer from scratch

 

however if you really wanted to make some form of calculator, you could make a simple four function analog calculator using a series of transistors, capacitors, and resistors. but you would have to read the answer to any problem in terms of an electrical signal

[Freeman]

Well, wouldn't a USB hardrive work? And I wouldn't need a fast processor, or much RAM. Wouldn't it be more along the lines of a subpar handheld computer?

 

But what about graphics?!

[Thomas Kirby]

You get into buying so much hardware that if your goal was to own a graphing calculator, you might as well have bought one of those, or a cheap old laptop that could run some graphing calculator programs.

[CPL.Luke]

Well, wouldn't a USB hardrive work? And I wouldn't need a fast processor, or much RAM. Wouldn't it be more along the lines of a subpar handheld computer?

 

still need software and to design and build a motherboard

[Freeman]

Well, I can get the motherboard used, I can get mathematica from my friend, and I can use...uh...Linux for the os.

 

Wait...why don't I just load mathematica onto a palm pilot do-higgy?

[CPL.Luke]

I don't think mathmatica would run on a palm pilot

 

and now what your talking about is building a computer, not building a calculator

[herme3]

the complexity and price of the parts for a graphing calculator will make it not worth your while' date=' as attempting to construct any form of modern calculator would be like trying to build a computer from scratch

 

however if you really wanted to make some form of calculator, you could make a simple four function analog calculator using a series of transistors, capacitors, and resistors. but you would have to read the answer to any problem in terms of an electrical signal[/quote']

 

I could probably build a desktop computer from scratch. However, I don't think I could build a laptop because everything needs to fit perfectly. I imagine you would have the same issue if you tried to build a calculator.

 

You'd have better luck coding a calculator program for the computer. This will give you some idea of the complexity involved.

 

I once did that using C++. It was a very simple calculator. All I really did was declare APStrings to the C++ functions. For example, if somebody typed "Multiply" the computer would use the "*" function. After that, they would type two integers. Then I used a cout command to display the answer.

[CPL.Luke]

I don't think anybody could build a comp from scratch, (by scratch I mean no mobo)

[YT2095]

I don't think anybody could build a comp from scratch, (by scratch I mean no mobo)

 

Oh I wouldn`t be so sure of that

check this out: http://www.scienceforums.net/forums/showthread.php?t=2893

[Thomas Kirby]

I don't have time to be very specific right now, but there are hardware and software readily available that can be adapted to this job. People did a lot of work with the 8031 types of microcontroller chips. There is a lot out on the net for it and its descendents. Some of them are pretty fast, low power, can support a lot of memory. In uses like a graphing calculator they would be way faster than a 10 MHz AT. They don't require hugely complex circuit boards, either, and the displays are available.

[mmalluck]

I know some of the older TI calculators (82, 83, 85, 86) used the Z80 processor running around 2 Mhz.

 

If you really wanted to build a calc, I suggest using a pic microcontroller. You just need to interface a keyboard and a display screen. That will be the easy part. The hard part is writting the software for the pic-processor.

[CPL.Luke]

thats amazing YT, but are an of those parts still readily available?

[Klaynos]

http://members.lycos.co.uk/racegame/phone/Picture%20(2).jpg some ram I made a little while ago... a whole 16 bytes worth

[YT2095]

thats amazing YT, but are an of those parts still readily available?

 

absoloutley! I have tins full of them each and then some!

you can also pick them up over the net cheap as anything too (you`ll have to program your own OS onto the EPROM though, like I did) but sure, and you`ll not pay the sort of prices I paid for mine either!

 

if you get stuck at all, let me know and I`ll gladly post you any parts that you have trouble getting if I have them, I`m a kinda retro-tech Museum curator

 

as for the OP, I have no experience with LCD dot matrix displays other than the single line (32 chr 7x8 displays) certainly nothing that will do gfx that you`re after.

[5614]

[url=http://members.lycos.co.uk/racegame/phone/Picture%20(2).jpg]http://members.lycos.co.uk/racegame/phone/Picture%20(2).jpg[/url'] some ram I made a little while ago... a whole 16 bytes worth

Did you have a circuit diagram to work from? (Can I have it if you do)

Did it use PICs or standard ICCs?

How did you know how to make it?

 

Same questions applies to YT's computer!

[YT2095]

(1)Did you have a circuit diagram to work from? (Can I have it if you do)

(2)Did it use PICs or standard ICCs?

(3)How did you know how to make it?

 

Same questions applies to YT's computer!

the () numbers are mine.

 

(1) yes, but I had to design and draw it 1`st.

(2) standard "off the shelf" IC`s.

(3) I studied electronics and took exams and read books and spent time experimenting, basicly alot of time and hard work on my part (but fun) )

 

also some questions of importance that you didn`t ask, you need to know the CPU you wish to use and then write your own OS for it (according to your hardware and mem stack), else it`ll have no boot sequence!

oh yeah, and I had to build my own EPROM programer to blast the EPROM, as I couldn`t afford a shop one at the time.

[Thomas Kirby]

The problem is that there are not too few choices for programmable microcontrollers. There are too many. A 2 mHz Z80 is not only primitive these days, it's a power hog. 10 or so milliamps can power a pretty fast device, in the 25 mHz or better region, that even has a fair amount of memory on chip, really fancy communications capabilities, and costs 20 dollars or less.

 

The thing to do is to look around for prior art. People write all sorts of micro OSes and put them on the net. I've even written a multi-tasker for a PIC 16F84A chip that made it possible to run a few tasks concurrently. There is no big difference between it and the vectored interrupt routines for the old IBM PC.

 

Here is a good place to start.

 

There are some fairly cheap displays out there at places like Electronics Goldmine and All Electronics. The worst is coming up with a keyboard.

[YT2095]

the kbd shouldn`t be a problem, all you need to do is wire them Row and Column and use a PIO, from there you can assign what you like to them in your routines (that`s all I did).

[Thomas Kirby]

"A" keyboard isn't that hard to find. I don't know where I would get one that is the size and shape of the one on the TI 86 graphing calculator.

[Klaynos]

Did you have a circuit diagram to work from? (Can I have it if you do)

Did it use PICs or standard ICCs?

How did you know how to make it?

 

Same questions applies to YT's computer!

 

Last semester I had a module on digital electronics

 

So yes I had a circuite diagram, unfortunatly I had to hand it in to be assessed so I no longer have it (got a first for the module though)

 

Standard ICC's I belive.

 

Just followed the circuite diagram then spend about 3 hours trying to work out why it wasn't working :@ then it worked

 

The most annoying thing was adding data to it as it wasn't connected to anything to do it automatically :s

[mmalluck]

Why reinvent the keyboard? A standard computer keyboard is easy to interface with a PIC processor. You could probably get away uing a non-standard (I'm thinking a numberical pad add-on for labtops) just as easily.

 

All you need is the list of scan codes to see what key is being pressed.

[YT2095]

Why reinvent the keyboard?

 

well by the same philosophy, why re-invent the calculator then?

 

simple: Because it`s FUN!

[mmalluck]

True, it is fun.

 

I'd just get annoyed with all the repetitive soldering that making your own keyboard would mean. Of course, this is being said by the man who's spent the last month soldering together his 1024 LED matrix board, which is only a little over half way done. I'm just a little jaded