OT: uCalc proto



#2

Here is the latest uCalc prototype design.

Almost laid out, just a few more things to take care of and then it's prototype time :-D

Final specs have turned out to be:

- PIC24FJ256GA106 processor, 256KB/16KB

- 3 Axis Tilt Sensor

- Piezo transducer

- 24 key touch panel with on/off switch on back

- 8KB EEPROM

- Micro SD Card

- Dual CR1620 battery

- IR LED

- 128 x 32 dot matrix reflective display

- 3.3mm thick for the first proto run to save cost, production should be 3mm.

Was hoping to have it in time for xmas, but I've been slack.

The front panel will look exactly as shown (but black), unless I change something at the last minute.

I'm using a more risky touch panel approach to lower cost and complexity, don't know if it's going to work reliably. Guess the prototype will answer that. In fact the whole thing has now been optimised for the lowest cost. With this new approach I'm actually not limited to 24 keys, but 24 works nicer for technical reasons.

Comments before I push the GO button?

Dave.


Edited: 15 Dec 2009, 2:46 a.m.


#3

How would the keyboard alpha be only the letters of A up to Z?
Change Roll up e Roll Down to STO and RCL.


Edited: 15 Dec 2009, 3:23 a.m.

#4

Awesome!

Is there a matching IR photo transistor, or is the IR LED for use as a remote control only? Why is there tilt sensor, do you have any specific ideas in mind -- or is it just for potential use in gaming?

-Katie


#5

Quote:
Is there a matching IR photo transistor, or is the IR LED for use as a remote control only? Why is there tilt sensor, do you have any specific ideas in mind -- or is it just for potential use in gaming?

Originally I had a full IRDA transceiver, but that got trimmed back for cost reasons.

No real idea in mind for the tilt sensor, although might be fun for games or menu navigation.

I contemplated being brutal and leaving both the features out, but I caved in :)

Dave.

#6

Hi, Dave;

looks great to me!

Please, let us know as soon as you are accepting orders!

Congrats!

Luiz (Brazil)

#7

Great looking!

Assume shifted x-y is ARC or ^-1 ?

I'd also vote for STO and RCL replacing the ROLLs as unshifted operations, like Saile did.

Will be interested as soon as you take orders, though your uWatch lies still in my shelf to be assembled ;)

Edited: 15 Dec 2009, 4:21 a.m.

#8

Oh what the heck, 4 extra function buttons actually fits.

Expanded alpha layout, and STO/RCL as primary (doubles as menu arrows).

W/X/Y/Z on the top because it just "feels right".

Dave.

Edited: 15 Dec 2009, 7:11 a.m.


#9

Please press GO!


#10

@Jeff,

Why do you want to divide first and subtract then? ;)


#11

Operator precedence.

#12

Looks good to me. Freeze the baseline & go.

#13

Dave,

Function buttons are great as is expending the alphabetical keys, but having the buttons above the display makes little sense since your fingers will block the display when you press them. Is it hard to change the PCB layout in order to put the keys below the display?

-Katie


#14

Quote:
Function buttons are great as is expending the alphabetical keys, but having the buttons above the display makes little sense since your fingers will block the display when you press them. Is it hard to change the PCB layout in order to put the keys below the display?

Not just hard, impossible I'm afraid.

It has to do with the physical size of the LCD behind the panel.

Dave.

#15

Would it make sense to have the W, X, Y, Z keys access diretly to the corresponding stack registers? (Well W standing in for T...)


#16

Quote:
Would it make sense to have the W, X, Y, Z keys access diretly to the corresponding stack registers? (Well W standing in for T...)

That is a nice possible benefit if you have a big font. But the screen is capable of displaying 4 lines of at least 16 characters anyway.

Dave.

#17

The only thing that I would suggest is to swap plus and divide over since addition and subtraction are likely to be used more often and having them on the edge of the keypad means less chance of getting something else by mistake.


#18

I second this suggestion. What are the dimensions of the front panel?


#19

Quote:
I second this suggestion. What are the dimensions of the front panel?

Exactly the same as a credit card. That's the entire goal of the project. Why?, just because I can :)

Sorry, the operator keys stay where they are, that's the Casio format I like.

Dave.

#20

Having forgotten most of the previous discussions, I'm assuming this will have complex number capabilities. Does it have an imaginary "i" key, or will that be obtained from the MENU? It'd be nice to have it next to pi instead of the // (but I'm not a programmer.)

CHUCK


#21

i suggest we use the "double point" (pressing "." twice) solution used in the uWatch. this saves a key and keeps complex out of your face if you dont want them.


#22

Quote:
i suggest we use the "double point" (pressing "." twice) solution used in the uWatch. this saves a key and keeps complex out of your face if you dont want them.

A very good suggestion.

I would also request that pressing the point twice between numbers could be used to implement a fraction input as used on the HP-33S and HP-35S models.

From the HP-35S User's Guide:
To enter a fraction, press "." twice: once after the integer part of a mixed number and again between the numerator and denominator of the fractional part of the number. To enter 2 3/8, press 2.3.8 . To enter 5/8, press either .5.8 or 0.5.8 .

This should not interfere with your suggested use of pressing "." twice (without numbers between the presses) to enter "i".

#23

Quote:
Having forgotten most of the previous discussions, I'm assuming this will have complex number capabilities. Does it have an imaginary "i" key, or will that be obtained from the MENU? It'd be nice to have it next to pi instead of the // (but I'm not a programmer.)

CHUCK


It will have anything someone programs into it, just like the uWatch. It will be open source, so will evolve into whatever the community wants.

It already has an "I" alpha key.

Dave.

#24

One more remark: I dunno how you did program the hyperbolics - it looks like you have to press SHIFT ARC COS HYP for ACOSH. That would deviate from the expected sequence SHIFT ARC HYP COS. If I got this right, some label rearrangement may be favourable to take care of this.


#25

The labels don't imply any actual entry order. That's up to what you want to do in software.

Dave.


#26

So how shall ACOSH be entered?

#27

I think Walter is suggesting that HYP should be on the 0 key and SIN, COS and TAN on the 1, 2 & 3 keys.

Or better, HYP on X-Y, ARC on 0 and SIN, COS and TAN on 1, 2 & 3.

Makes the entry order more natural.

So to enter ACOSH, you press these keys:

    SHIFT X-Y 0 2

- Pauli


#28

Fixed!, as per your suggestion.

Dave.

#29

Thanks for your understanding d:-)

#30

How do you control the display decimal digits? Is fix vs scientific notation automatic?


#31

Anything not directly available on the keypad gets done in the menus.

I decided to add base number stuff on the 2nd shift, so no room left for FIX, SCI and other stuff I'm afraid.

Dave.


#32

Even though I'm an engineer (electronics), I'd rather have SCI than <ENG & ENG>, I never use them. I take the value as per SCI setting and mentally convert to milli, micro, pico or units, Kilo, Mega.

And the only time I've used FIX is on the 35s because of the stupid scrolling display.

Just my opinion.


#33

I find it abhorrent to have to do a mental calculation or conversion after a calculator has given you a result, kind of defeats the purpose of having a calculator I think. If I'm using a calculator then I expect it to give me the result I expect. So I use the ENG key daily, although I hate calcs with those fixed ENG modes, I want the choice of when to convert to ENG display and when not.

SCI and FIX are more "mode" like than ENG in this instance, so don't deserve a primary key IMO.

But each to their own!

Also remember, it's trivial to make a stick on overlay to have any key layout you like.

Dave.


#34

If you do the mental calc daily it becomes trivial, order of magnitude convertion is soooo easy anyway :).

As I said it's my opinion, I don't mind if SCI is a menu item, it's not as if I change it for every calculation (like you said, more a mode change)- unlike those that use ENG.

As you say, it's easily reconfigurable and that's what I like most about it. Even though it's not "programmable" like traditional programmable calcs, I assume I can put in my own routines (that I would normally have programmed into the RAM of a programmable calc) and assign them to keys I don't normally use, or the F keys or add them to the menu.


#35

maybe you shake it and it changes the result display format!


#36

Brilliant idea - 6 extra inputs on the tilt sensor!! :)

#37

Turn it over and it turns itself off?! All sorts of possibilities here.


#38

More like right side down does ENG> and lefgt side down does <ENG. :)

#39

Quote:
Turn it over and it turns itself off?! All sorts of possibilities here.

It's mainly in there for those reasons you *can't* think of right now!

Random number generator for starters. Menu scrolling, mode switching, games, a crude spirit level sensor...

Dave.


#40

I have an HP-21 emulator on iPhone (GO-21) which uses shake to CLx the machine - kinda cool!


#41

Quote:
I have an HP-21 emulator on iPhone (GO-21) which uses shake to CLx the machine - kinda cool!

Noooooo!!!

Imagine: a long calculation, put the calc down on the desktop a bit too hastily and the answer gone before you can write it down. :-(

#42

Quote:
If you do the mental calc daily it becomes trivial, order of magnitude convertion is soooo easy anyway :)

Yes, it's trivial, but that doesn't help you on those bad days when your mind goes, err, wonky. Before you know it your Mars probe crashes into the planet :->

Dave.


#43

Quote:
but that doesn't help you on those bad days when your mind goes, err, wonky. Before you know it your Mars probe crashes into the planet

IIRC, that wasn't due to wonky minds but due to stubborn minds keeping wonky units, wasn't it? d;-)

#44

Mistakes I've come across in aerospace equipment (which although not fully certified, were in advanced stages of development, delivered for fitment to flight test aircraft):

Multiplying instead of dividing by 1.94 when converting from knots to m/s.
Using the wrong sign for ascending/descending.
Swapping the reference axes around.

And these are just some that I can recall off hand.

#45

Quote:
I find it abhorrent to have to do a mental calculation or conversion after a calculator has given you a result, kind of defeats the purpose of having a calculator I think.

Well said!

That is one reason I am so drawn to the 48 series - the units management, at least in theory, makes mental checking of the result less necessary. If only it weren't so difficult to learn ...


#46

So you believe everything your calculator spits out, not even caring to do a quick mental check as to the order of magnitude? (That's what you seem to be saying).

By the way, all I was saying was it's easy enough to mentally go from say 1E-8 to 0.01E-6 or 10E-9 without pressing buttons.


Edit: but I caved in: keep the <ENG and ENG> buttons, I'll leave them as unused as on all my other calcs that have them :)

Edited: 17 Dec 2009, 5:27 a.m.


#47

Quote:
So you believe everything your calculator spits out, not even caring to do a quick mental check as to the order of magnitude? (That's what you seem to be saying).

That's what everyone says when this subject is bought up, but it always misses the point.

The calculator is a calculating tool, if it doesn't easily give you the result you expect from it, formatted the way you want it, then it's not as good a tool as it could be.

If you have to do extra conversion in your head then your chance of error is greater, regardless of how easy it is. Mentally checking your result and knowing what to expect etc, is really a separate issue.

Quote:


By the way, all I was saying was it's easy enough to mentally go from say 1E-8 to 0.01E-6 or 10E-9 without pressing buttons.

Sure, but that's not the point.

I bet you'd get a tad annoyed if your scientific calculator gave you "0.0000000123" all the time instead of scientific notation, or engineering notation all the time when you wanted scientific etc. There is nothing wrong with wanting your tool to give you the exact result you expect.

BTW, a good engineering calc will have a primary ENG key, so it's only one key press. And when you get used to hitting it subconsciously it really is a valuable function you don't want to live without. But YMMV.

Quote:
Edit: but I caved in: keep the <ENG and ENG> buttons, I'll leave them as unused as on all my other calcs that have them :)

I will keep the ENG keys, thanks, the joy of being the product designer :-P

And that weird parallel key stays too, ain't I just evil!

Dave.


Edited: 17 Dec 2009, 6:30 a.m.


#48

Quote:
formatted the way you want it,

And this part of the thread started because I merely said ENG format is not MY kind of format. (The mental thinking bit was only a justification, not the point in itself). But I concede because I understand there is a large contingent out there that does use it. That's the reason you should keep it on there - not just because you're the product designer. As I said early on in this thread, I'm happy with what I see, let's see some hardware.


Quote:
That's what everyone says when this subject is bought up, but it always misses the point.

A statement which you negate by:
Quote:
if it doesn't easily give you the result you expect from it,

How can you expect anything if you haven't done some mental reasoning?

By the way, in another post in this thread I mention errors in equipment, I found those errors because whilst doing on-aircraft checking (on ground of course) other equipment dependent on the information were reacting, and whilst seemingly correct, I was always asking myself "is this more or less what I was expecting?". Some were obvious (wrong direction) others more subtle (over or under reaction).

#49

Quote:
How can you expect anything if you haven't done some mental reasoning?

Not the actual quantitative number but the *formatting* of the number.

Like I said, as an example you might get pretty annoyed if your calc always gave you 0.000000123 instead of scientific notation.

Just like as an engineer I get annoyed sometimes when I don't get my result in engineering notation, because that's the *formatting* I've come to expect and work more naturally with, even if it is a small difference I can easily convert mentally. But note that I don't always want engineering notation, which is why I like my ENG key and having the choice, instead of an ENG mode which forces it down my throat when I don't want it.

Dave.

#50

How do the buttons work? Clicky membranes? If so, they should be easily swapable. I have had such clicky membraned break after not so much wear, and that mostly means tossing the object in the bin.


#51

Quote:
How do the buttons work? Clicky membranes? If so, they should be easily swapable. I have had such clicky membraned break after not so much wear, and that mostly means tossing the object in the bin.

They are capacitive touch sensors, so nothing to wear out apart from the silkscreen PCB overlay. It's the only risky part of the design, it actually may not work reliably or at all, that's what the first prototype is designed to suss out. If I'm lucky it'll work first go and I should have something close to a finished product, if not then it's tweaking time, or on to plan B.

Dave.

#52

Thunderbirds are GO!

I should have a physical proto just after new year.

If it doesn't work as intended then I'll quietly slither away... :->

Dave.


#53

Will there be an additional protection for the display to reinforce the calculator and making it more dust-proof and scratch-resistant?


#54

Quote:
Will there be an additional protection for the display to reinforce the calculator and making it more dust-proof and scratch-resistant?

I'm afraid not. Unless you know someone who can custom cut quality thin (0.5mm) scratch resistant mineral glass cheaply? I'd be very interested to know of anyone who does.

Dave.


Possibly Related Threads...
Thread Author Replies Views Last Post
  uCalc Video DaveJ 9 1,006 12-01-2010, 01:18 AM
Last Post: DaveJ
  uCalc DaveJ 6 765 07-02-2010, 04:21 AM
Last Post: Bill Zimmerly
  OT: The uCalc is born DaveJ 21 1,765 01-07-2010, 03:25 PM
Last Post: Michael Plant
  HP Classic with new electronics (a la uCalc/uWatch?) Mark Henderson 26 2,150 12-18-2009, 05:54 PM
Last Post: geoff quickfall
  OT: uCalc Processor DaveJ 9 986 12-09-2009, 12:51 AM
Last Post: Steve Leibson
  uCalc DaveJ 126 8,846 11-21-2009, 11:12 PM
Last Post: designnut
  HP41 MLDL2000 proto's arrived! Meindert Kuipers 12 1,001 11-25-2004, 01:10 AM
Last Post: Eddie Shore
  Museum and MLDL2000 proto pictures online Meindert Kuipers 0 265 07-25-2003, 03:39 AM
Last Post: Meindert Kuipers
  Charlemagne Proto Kim 1 284 07-23-2002, 10:13 PM
Last Post: James M. Prange

Forum Jump: