Prime crashes upon fsolve

[emax]

To calculate the sides of a display with a given diagonal of 7.112 and an aspect ratio of 4:3, I tried the following (in CAS):

Code:


fsolve({(a^2) = (b^2+c^2),b = (abs(4/3*c)),a = 7.112},{a,b,c})

This works on my linux virtual calculator:

   

with

Software Version: 2.1.14288 (2019 10 22)
Hardware Version: Emu
CAS Version: 1.5.0

Operating System: Ubuntu 20.04.1 LTS


My physical Prime however crashes: "... had a problem and will reboot in 3 seconds".

I can reproduce this behavior.

Software Version 2.1.14433 (2020 01 21)
Hardware Version: D
CAS Version: 1.5.0
OS: V2.060.650

Is there something newer which I could install?

Anybody to try/reproduce this on his own Prime?

[DA74254]

If I just copy/paste your expression from here to my Virtual Prime on win10, it crashes.

When dialling in by calc keyboard on all three versions, it works.
The Android Pro version 2.1.143346 (2019 11 21)
Prime G2, I have the exact same version on my Prime as yours.
Virtual Prime 2.1.14425 (2020 01 16)

[emax]

Thank you!

I will change my language settings and the other settings to factory-defaults and report back as soon as I have time.

[roadrunner]

Try giving fsolve a guess to start with:

fsolve({(a^2) = (b^2+c^2),b = (ABS((4/3)*c)),a = s},{a,b,c},{7,6,4})

or

fsolve({(a^2) = (b^2+c^2),b = (ABS((4/3)*c)),a = s},{a,b,c},{7,6,-4})

when i do this it doesn't crash and results in {7.112,5.6896,4.2672} and {7.112,5.6896,-4.2672}

-road

[trojdor]

EDIT: I see roadrunner beat me to the draw, while I was typing.

I suspect you'll find that giving the solver an initial guess of 1 for the variables will stop the crashing on all version of the Prime.

Either:

Code:

fsolve([(a^2) = (b^2+c^2),b = (ABS((4/3)*c)),a = 7.112],[a,b,c],[1,1,1])
or
fsolve({(a^2) = (b^2+c^2),b = (ABS((4/3)*c)),a = 7.112},{a,b,c},{1,1,1})

depending on your notation preference.

Conversely, if you want to watch them all crash, use an initial guess of 0 for the variables.

(Of course, you could/should also make the initial guess for 'a' equal to it's actual value of 7.112, but I'm just lazy, lol....and just using 1 works, so...)

[emax]

Thank you for your answers.

I have now set the Prime to the standard settings, language English and Radians mode - just to get ahead of such suggestions.

This didn't help.

To your suggestions: It is of course a workaround to define a guess. But this is not a real option.

With such a ridiculously simple task like the one showed here, that may be possible. But with more complex problems, those where you really can't guess the answer, that's exactly what the calculator should do: Relieve me of guessing games. That's what I bought it for. And even a TI89/V200, which admittedly needs some seconds for the solution, doesn't need guesses. A year ago I bought one on ebay for 20 Euros for occasional messing around and so could try this out.

On the Prime, there is a setting of "Recursive Evaluation" which could be a cause. This is by default set to 25, and I didn't try other settings. And even if 25 was to small, the calculator should (after having been released seven years ago) never crash. Instead, you should see a message.

But the Prime just rebooted. :-o

For me, this is the equivalent of a windows blue screen. :-|

I had another problem before this happened, but didn't post it here as I could not reproduce it: The Prime (again in CAS mode) all of a sudden showed a totally garbled display completely covered with blueish and whiteish tatters. Keys didn't work, shift-off didn't work, no responses to whatever I did. And the display didn't dim after the dimming time had expired. The only way to get back to operation was a hard reset. I had this just once, but haven't used the Prime much since then.

Another thing came to my attention: When I edited a longer formula (again before the problem occurred which I described in this thread), I got strange error messages which contained unreadable characters. They contained characters like \@\@ or ^@^@ or so, which reminds me of binary zeroes being displayed sometimes in the output of faulty programs. As I am new to the prime, and as I could not reliably reproduce this I was reluctant to post this here.

It seems, that my hardware-version is quite a new one. I don't hope, that this is the reason for the problems.

One thing is certain: I have lost some confidence.

As there wasn't (yet?) an answer from an HP member here, I'd like to know whether there is an official contact at HP to ask for help.

Can anybody point me to the right place?

[roadrunner]

You can use solve and it get's both solutions without a guess:

solve([(a^2) = (b^2+c^2),b = (ABS((4/3)*c)),a = 7.112],[a,b,c])

returns:

[a = 7.112,b = 5.6896,c = 4.2672],[a = 7.112,b = 5.6896,c = −4.2672]

-road

[emax]

Hello roadrunner,

the point for me is actually not how to solve this particular problem. The point is, that I can not rely on this machine.

And I do not want a device which is prone to crash next moment.

Perhaps one of the augurs will chime in.