First impressions on Casio FX-5800P

[Manolo Sobrino]

I've just received one of these brand new coming from Korea, and I want to share a few first impressions with you. People interested already know about its features (scientific, programmable, non graphing), here is the User's Guide.

The keyboard is decent, much better than the recent Casio scientifics, although it's plastic (ABS?) over membrane. It is "soft" (but not too/that much), yet provides a reasonable haptic feedback with an audible tap, no registering problems so far. Keep in mind that this is not an expensive calculator.

Pixels are on the big side, so numbers look big, pretty nice overall. LCD is quite good, a bit greenish but very readable. It has a protective (I guess Polycarbonate, maybe PMMA) protruding window, moderately thick, ~1mm. The calculator is well made, lighter than it looks, a bit tilted and a bit weight balanced towards the screen. Cover is made from ABS, (as is everything else?) looks a bit fragile but not flimsy at all, it closes nicely. Faceplate and cover are black, the rest is dark grey, and the texture of it all is matte. When closed it shows you just a silver small CASIO label... It's a fairly handsome closed calculator .

It is strongly modal: MODE key shows/hides the mode menu, so you can pick one. After you choose one mode, you can get the menu of available functions for that mode that are not in the layout via FUNCTION key. If you don't choose any of them, you get out of there via EXIT key, which also returns to the previous menu in case you're in a sub-menu. It's not that awkward after you get the hang of it.

You have the same maths keys of a 115ES, but now Ans and inverse are shifted. I don't understand very well why Ans is not primary as it has the nice feature of understanding the arithmetic minus key sign as a minus sign everywhere. It has a prominent i key. There is no complex mode, it just works with complex numbers for arithmetic, ^2, ^3 and SQRT. All the features of the 5000F (formula and alphanumerics specially), 4500P/PA (FILE) and 5500L/LA are there, and those of the 115ES too except for unit conversions, in fact it's kind of a 115ES on steroids.

In COMP mode, Natural Display mode works like a charm (you can choose between this "MathIO" or standard "LineIO"), but it has a problem with implied multiplication (syntax error) if just a number follows the fraction (put an x (*), sqrt or a parentheses and it works). As different templates are used for proper and improper fractions there's no ambiguity, default mode is improper.

The speed is reasonable, input buffering doesn't seem to make it sluggish when entering long expressions. All 4 cursor keys enable REPLAY, and allow for navigation/editing through current input and recent history. The number of recorded calculations is not fixed, it appears to be limited by some specific memory constraint.

Statistics look nice, regression too, you can work straightforwardly with your table of data and move around it. Base-N mode is a pleasure, as in BIN you can see all the 32 signed/unsigned bits.

It claims 15 digits for internal calculations, 10 for display mantissa, a precision (meaning accuracy?) of 1 ulp for those and cumulative errors. Ordinary Casio range. I haven't tested it really (numerical solver & integration, programmability etc.), but trigs apparently don't do weird things.

It also remembers where you were (and your numbers) when you power it on. It comes powered by a Panasonic alkaline AAA battery. You have to unscrew the battery cover to replace it (the screw socket is made of metal, nice touch, Casio's own mechanical engineering is usually pretty good -in watches it's outstanding-, this doesn't look like a Kinpo idea). The back cover is fixed with Phillips screws, so I guess the whole calculator can be disassembled easily.

It came in a small package with an A-6 size plain paper copy of the ~140PP English User's guide, and a folded leaflet with all the constants and formulas. The "optionally available cable (SB-62)" which allows for settings and programs transfer between calcs (only) is not included.

Overall, a surprisingly pleasant and relatively inexpensive (on sale) welterweight calculator without school looks. I'm liking it so far.


(edit) I guess a lousy picture is due. It might not be apparent but it's half as thick as a (dead) 50G:

   

[Tugdual]

Thanks for sharing this. Being allergic to TI I always considered Casio as a very decent alternative to HP. This one looks very nicely featured for common engineering work. I like the display and the overall design of Casio that definitely goes its own way and seems to be consistent over the years.
Was reading the user guide (thanks for the link) and it looks very similar to a 34s features wise, or am I going too far?

[Dave Britten]

How is the speed? I've got an fx-9860g Slim that smokes pretty much everything in its weight class.

Casio stuff is generally pretty good, especially considering their focus on lower-priced models. I think the heavily modal UI is something of an Achilles' heel, however, and artificially limits program flexibility.

On a related note, I just got an fx-8500g in the mail today. It's a quite peculiar mix of their first-gen models (fx-7000g) and pocket computers. It's also got one of the strangest text editors I've seen since vi. Pretty cool overall, but it's still annoying that there's no Print/Disp command that doesn't pause program execution. (Is the fx-5800p the same in that respect?)

[Thomas Klemm]

(04-19-2014 07:26 PM)Dave Britten Wrote:  How is the speed? I've got an fx-9860g Slim that smokes pretty much everything in its weight class.

From Calculator Speed Benchmark using the N-Queens Problem:

Quote: - 4:17 HP-35S Keystroke / RPN
- 4:08 HP-41CL Keystroke / RPN / Turbo5 Mode x4.3
- 4:02 HP-48GX UserRPL / Ver.P
- 3:47 FX-5800P Formula / Array
- 3:28 HP-42S Keystroke / RPN / Ver.C / Turbo + Fast Mode x3.5
- 3:15 DM-15CC Keystroke / RPN
- 28.7 FX-9860G Slim Formula / Matrix / OS 2.0
- 2.3 WP-34S Keystroke / RPN / Real Mode
- 0.346 HP-Prime Formula / List
- 0.000150 HP-50G Assembly / ARM9 @ 192 MHz / Fast Mode x2.7

[Thomas Klemm]

(04-19-2014 06:54 PM)Tugdual Wrote:  it looks very similar to a 34s features wise

I'm missing control structure commands:

• If ~ Then (~ Else) ~ IfEnd
• For ~ To ~ Next
• For ~ To ~ Step ~ Next
• While ~ WhileEnd
• Do ~ LpWhile
• Break
  

Cheers
Thomas

[Waon Shinyoe]

I have a 5800p too. I think its speed on calculation (especially integration) is quite fast: maybe it is the fastest scientific calculator. However, the keyboard is terribly soft. Even a 991 is better.

[Tugdual]

(04-19-2014 10:43 PM)Thomas Klemm Wrote:  

(04-19-2014 06:54 PM)Tugdual Wrote:  it looks very similar to a 34s features wise

I'm missing control structure commands:

• If ~ Then (~ Else) ~ IfEnd
• For ~ To ~ Next
• For ~ To ~ Step ~ Next
• While ~ WhileEnd
• Do ~ LpWhile
• Break
  

Cheers
Thomas

Are you saying that the FX 5800P is superior to the 34s?

[Thomas Klemm]

(04-20-2014 08:33 AM)Tugdual Wrote:  Are you saying that the FX 5800P is superior to the 34s?

No. But these control structures and arrays aren't provided by the WP-34S. Both calculators appear to be similar yet neither is superior to the other. It all comes down to how important the different features are for you.

Cheers
Thomas

[Manolo Sobrino]

Tugdual: You couldn't possibly compare two more different calcs. I don't know the 34S very well, just enough to tell that its programming model is keystroke. This Casio allows for more (modern) paradigms, as Thomas has suggested. I save programming in calcs for a last resort option, so that's the last thing I consider about them. This one apparently packs the command set of the Casio modern graphing calcs, except for drawing, IO and matrix Dim. I don't know if the graphing ones support logical operators in conditional statements, this one does.

The 34S is a fans' RPN programmable calculator and its layout reflects that. The 5800P is an algebraic scientific first, and then it can also do formula... and then programming.

Independently of RPN/algebraic and programming models, the 34S can do things that the 5800P can't. It is faster and claims to be very accurate. If you need those features that's the calculator for you. Something like this is just convenient for me. It could have full complex support and unit conversions... I use other calcs that can do that (there are superb TIs out there, you'd be surprised with what you're missing, too bad allergies can't be fully rationalised ). They do not exclude each other, you can get both.

Dave, speed would disappoint you, it's just OK for what it does. I like the Slim, but it lacks some elementary useful keys on the layout. I guess it's a much better calc for programming.

I've been always curious about the 8500G, I like its looks. But I have a 6500G and a 7500G, I fear getting the same calculator again. I'm more of a user than a collector anyway.

Shinyoe: I've had to use a 991ES and have tested the 570ES, I can't stand their keys. Missed keystrokes all the time, maybe the Plus versions are better. There are older Casios with softer keyboards (not only rubber ones), yet I rarely missed a key with them. Hard keyboards are nice for a while, but try to take a long calculator based exam with them... I guess I have soft fingers.

I've tried the N-Queens program from Xerxes:

http://www.hpmuseum.org/cgi-sys/cgiwrap/...i?read=700

And interesting things happen. I get it done (S=876) in 4:31s (!). I don't know very well what is exactly being measured here and the fact is that the code from Xerxes has some black magic going on (you really have to call that odd Deg mode). Except for creating matrices in the program, the code for the FX-9860Gs should work, I'll give it a go at some point... writing programs here is time consuming, bordering on tedious, as commands are in nested menus.

[jebem]

(04-19-2014 02:55 PM)Manolo Sobrino Wrote:  I've just received one of these brand new coming from Korea, and I want to share a few first impressions with you.

Thanks for sharing your findings, Manolo.
May I ask how long did the package took to be delivered to Iberia?
I am browsing eBay UK since a couple of months looking for this calculator, and the South Korean supplier looks like the best choice if one looks for the best price, but I have no experience doing business from there.

Saludos.

[Manolo Sobrino]

(04-26-2014 04:00 PM)jebem Wrote:  

(04-19-2014 02:55 PM)Manolo Sobrino Wrote:  I've just received one of these brand new coming from Korea, and I want to share a few first impressions with you.

Thanks for sharing your findings, Manolo.
May I ask how long did the package took to be delivered to Iberia?

Saludos.

You're welcome jebem! (and PM sent!)

[Eddie W. Shore]

Manolo Sobrino

Likewise, thank you for sharing your in-depth review. I have just acquired a fx-5800p myself and so far I am impressed.

Eddie

[Gerald H]

On comparing different things:

Either every object is different from every other object so they cannot be compared with each other OR every object is different from every other object and only for this reason can they be compared with each other.

I choose the 2nd alternative, otherwise you can't compare anything with any other object.

I prefer the WP 34S to an Aston Martin.

[walter b]

(05-13-2014 09:07 AM)Gerald H Wrote:  On comparing different things:

Either every object is different from every other object so they cannot be compared with each other OR every object is different from every other object and only for this reason can they be compared with each other.

I choose the 2nd alternative, otherwise you can't compare anything with any other object.

I prefer the WP 34S to an Aston Martin.

Thank you

Is there any post you were refering to with your statement about comparability? You weren't quoting any and I can't find any.

d:-?

[Eddie W. Shore]

Due to the lack of modulo (MOD) and signum (SIGN) functions on the fx-5800P, here are several subroutines to help compensate for it:

SIGN(x): // x is any calculation stored in a variable
...
x > 0 ⇒ 1 → x :
-x > 0 ⇒ -1 → x:
....

n MOD d :
...
numerator → N
denominator → D
Frac(Abs(N ÷ D)) × D
...

Alternative (better way for MOD):

N - D Intg(N ÷ D)

[jebem]

(05-13-2014 09:40 AM)walter b Wrote:  

(05-13-2014 09:07 AM)Gerald H Wrote:  On comparing different things:
Either every object is different from every other object so they cannot be compared with each other OR every object is different from every other object and only for this reason can they be compared with each other.
I choose the 2nd alternative, otherwise you can't compare anything with any other object.

I prefer the WP 34S to an Aston Martin.

Thank you

Is there any post you were refering to with your statement about comparability? You weren't quoting any and I can't find any.

d:-?

Ya, this is getting difficult to follow... Probably Gerald was commenting to the Manolo remarks on post#9, but who knows?

Speaking of comparing two different things I have received today my Casio FX-991DE Plus (German edition). Now, how does it compare with a FX-5800?
It only write messages in German and it uses comma as the decimal point. Cool!
I like the comma as the decimal point as that's the Portuguese standard as well.
Although I'm not so sure about the German messages. Probably that's my chance to lean some more words in German...

[Manolo Sobrino]

Thank you Eddie!

It's strange that this calc isn't more popular. I couldn't find the kind of review I like so I posted one. As you are discovering now it's a remarkable device.

For instance, the solver is pretty capable, it's fast and accurate. You can't specify the interval, though. From reading the manual you could think that it only solves equations of the type X=simple expression, but in fact it will try to solve any equation in the form f(x)=g(x) or F(x)=0. For more variables it treats them as parameters. You can change their values by pressing the EXE key, or solve the equation for any of them by pressing SOLVE again when the editing is done, then the last variable in the problem would be just another parameter. The input value is used as an initial guess and the result is kept as the value for the variable. It shows you the L-R residue and if convergence is slow (for instance x^4=0 with initial guess 2 -kind of a torture test -) it asks you whether you want to continue refining the root value.

I tried the input lag test for the 991ES from Voidware and could type that without problems. Actually it does get sluggish for very long expressions in MathIO, maybe more than 110 characters, that's why I didn't notice it at first. I don't think this has many practical consequences.

It also passes the Voidware integration torture test with better results than those of the 991ES. Accuracy is on par with that of the 4800P and 991ES.

Quibbles after a few weeks of using it sparsely:

1) You can really get weary of menus: typing programs on it or writing two-character formula variables is nothing short of a chore.

2) Shifted STO. When it's faster to do so and I don't need the expressions I just take an approach similar to some ideas in RPN, working inside-out with ANS and the rest of the stack being loosely emulated by registers (yes, there are a few ways to use algebraics). For this, I'd prefer shifted RCL instead... or two primary keys.

Gerald... it's just that IMHO apples and oranges aren't really the best subject for a detailed comparison

[jebem]

Meanwhile I have found out a nice video on "Unboxing - Casio fx-5800p" by our dedicated Eddie Shore that I believe can bring some additional information to this thread.

[Marcus von Cube]

(04-19-2014 02:55 PM)Manolo Sobrino Wrote:  The "optionally available cable (SB-62)" which allows for settings and programs transfer between calcs (only) is not included.

I've two of these calcs and the cable of course. I got them in order to find out more about the communication. Some preliminary checks with a digital scope didn't reveal too much about the communication scheme. Has anyone more info? Is it a standard protocol like I²C?

I've added a zipped binary from my scope. You'll need the logic analyzer software from Owon to view the file.

[BruceH]

This image* precisely sums-up, for me, exactly what has been wrong with nearly all Casio calcs for the last decade or more: they used to be so neat and elegant... <sigh>



[*] Apologies but I don't understand enough Japanese to pick-out the name of the blogger to give proper credit. The image comes from this site.