hp35s vector input bug semi-workaround?

[pkh]

Am I years too late? Has a workaround already been posted for this?

I got my 35s (and for all its faults, I have still come to like it, but I digress) and while going through the Vectors section of the manual, for some reason, entering any vector with more than one dimension gave me a "Syntax Error".

I couldn't find anything on an internet search, only vague clues that it exists. Some users mentioned that it caused them some grief. Can anyone tell me if the following helps / works for you?

(1)
If you enter a 2-D vector (while the HP35s is willing to work), press enter and it will push it on the stack.

(2)
If you now enter an expression in the equation editor (could be anything as simple as a single RCL X), then going back to the stack, entering a 2-D vector causes the syntax error when enter is pressed. (1-D vectors worked for me, but 2-D vectors cause a syntax error that sends the cursor under the first digit after the comma).

(3)
Now, if you delete that equation you just entered, it starts to behave good again.

(4)
What if you wanted to keep that equation/expression? If you enter an expression that you want to keep, but causes vector syntax errors, then put more (useless space-filler) expressions into the equation editor. Try a bigger one, like for example sin(x)*inv(sqrt(x)) and it MIGHT start to work. If it didn't, then try another one. This has worked for me.

So this half-solution is vague, but just try sticking random equations/expressions into memory. That way you may not have to erase all your programs trying to get it to work again.

All good fun.

[Dieter]

(03-17-2014 04:38 PM)pkh Wrote:  I got my 35s (and for all its faults, I have still come to like it, but I digress) and while going through the Vectors section of the manual, for some reason, entering any vector with more than one dimension gave me a "Syntax Error".

Hm... I cannot reproduce an error. Actually I am not sure if I understand which error you refer to. On my 35s it all works as advertised:

Press the "[]" key (blue-shifted brackets) and enter one, two or three numbers, separated by a comma (the one on the yellow-shifted zero-key). What happens if you do this?

[1, 2, 3] [ENTER] [2, 3, 5] [x] => 23 (dot product)
[2, 4, 6] [ABS] => 7,4833 (magnitude)

Dieter

[Thomas Klemm]

The vector input bug has been reported in the HP-35s bug list.
Dieter, did you follow step 2? Because that's probably crucial to reproduce the bug. I wasn't aware of the connection to the equation editor. But then I rarely used them.
As I don't have my HP-35s with me I can't verify that now.

Pkh, thanks for sharing the information
Thomas

[Dieter]

(03-17-2014 09:20 PM)Thomas Klemm Wrote:  The vector input bug has been reported in the HP-35s bug list.

Here is a quote from that URL:

Quote:Vector input bug has been mentioned and very weird behavior shows up but I'm not aware of this being isolated or repeatable. It might be related to memory full. I do have enough trust in the people reporting problems to believe there is something to them.

So it says that "weird behaviour" has been reported, but obviously it was not possible to reproduce the problem. The source does not even say what exactly the error was. Do you know some details? I never experienced any problems with my 35s when using vectors.

(03-17-2014 09:20 PM)Thomas Klemm Wrote:  Dieter, did you follow step 2? Because that's probably crucial to reproduce the bug. I wasn't aware of the connection to the equation editor. But then I rarely used them.

I tried the procedure as suggested: Enter a 2-D-vector - e.g. [1, 2], quit with [ENTER] (both display lines now show [1.0000, 2.0000]), switch to EQN mode and enter "RCL X" (displayed as a simple "X"). Exit EQN mode and enter another 2-D-vector - e.g. [3, 4] - and press ENTER again, or [x] for a multiplication, or something else. I could not force any error to show up.

I did this about a dozen times. Then, finally, in one case indeed an error message appeared. Tried the same procedure once again - no problem. Strange.

Dieter

[pkh]

Yes, Thanks Dieter and Thomas for your replies and trying things out.

I actually noticed the error appeared for me right after I entered a custom expression in the equations list. I don't have any programs stored in memory, but did already have 6 (maybe 7) equations in there. When I added that one more equation, got an error just trying to put a matrix on the stack.

The error shows up (in my experience) when just entering a single matrix onto the stack, with SOME connection with equation memory usage.

You know the bug is in effect when you enter a matrix and press enter, like:
[2,3] ENTER
...and on pressing enter, you get "syntax error", and cursor goes to the character after the comma (in this example, the digit '3'). If you then immediately clear the line with 'C' and enter a 1-D vector...
[5] ENTER
...then I found that to be accepted, but that doesn't cure it for other dimensions. The only half-cure I found so far is to key in more equations to push the memory around a bit. Hope it gets us or anyone closer to understanding it better, for what it's worth.

I'm thinking therefore that program storage may affect it also, but I have not tried writing programs yet, I'm still going through the manual for the first time.

I'll post back as (or if) I find out more.

[pkh]

I should also mention that the memory was anything but full when the error appears. I had plenty of available memory (~29.8k free) and variables were cleared to see if that made any difference (which it didn't). It's only when there's a certain amount of equations filling the memory, but not many.

Also, I've been saying 'matrix' when I mean 'vector', but it's all the same in these instances.

[pkh]

...and one more thing,

if you find that you suddenly start getting the error when you already have two 2-D vectors entered into the stack, then clear the entry that causes the error, roll down the stack (with R-down), and do a dot product on the stack entries, it calculates the dot product correctly, but still won't accept new vectors input.

It's as if it starts to have difficulty accepting new vector inputs, maybe difficulty understanding the comma?

[dalupus]

I just ran across this bug on the hp35s simulator.

Was trying to do an equation to take the ACOS of each element of a 3 dimensional vector.
In trying to get it to work the vector bug cropped up and it started doing syntax error on all the commas.

I did as the poster suggested and just added a few bogus equations and suddenly it all started working again.

I just ordered an actual hp35s and I really hope that mine doesn't have this issue, since vectors is one of the main reason I bought it.

I do believe the bug is related to equations and vectors being used in combination. Might be safer to just do all vector operations in programs.

[Thomas Radtke]

(06-09-2016 11:51 PM)dalupus Wrote:  I just ordered an actual hp35s and I really hope that mine doesn't have this issue, since vectors is one of the main reason I bought it.

It most certainly has this issue as there were no updates to the firmware since its release in 2007. Beautifull concept and hardware, but the firmware is a lost case.

[Marcio]

(06-10-2016 07:30 AM)Thomas Radtke Wrote:  Beautifull concept and hardware, but the firmware is a lost case.

That's the 35s, unfortunately.

[dalupus]

(06-10-2016 07:30 AM)Thomas Radtke Wrote:  

(06-09-2016 11:51 PM)dalupus Wrote:  I just ordered an actual hp35s and I really hope that mine doesn't have this issue, since vectors is one of the main reason I bought it.

It most certainly has this issue as there were no updates to the firmware since its release in 2007. Beautifull concept and hardware, but the firmware is a lost case.

Well just replicated it again and again simply adding another bogus equation (like 75+35) fixed it. Hopefully that is all it ever takes to fix the bug.

[Gerald H]

I don't consider you, Thomas, a lost cause - & in this case I advise non-proliferation of the wrong form.

As for the 35S: one of my favourite calculators, sadly hampered by dreadful defects.

[Paul Dale]

(06-10-2016 07:30 AM)Thomas Radtke Wrote:  Beautifull concept and hardware, but the firmware is a lost case.

I have to agree completely. I wish I'd had a chance to write firmware for this hardware


Pauli

[Pacazo]

I might have found a fix for the vector input bug.

When the problem shows up, go to the equation editor and move to the start of the begining of the list. That is:
1. Press EQN
2. Press up and right until the message EQN LIST TOP appears on the bottom line.
3. Input a vector normally.
That is. If it doesn't work, try turning it off and back on. This has been working for me consistently and I cannot reproduce the bug any more.

[Thomas Radtke]

(10-09-2021 06:35 AM)Pacazo Wrote:  I might have found a fix for the vector input bug.

When the problem shows up, go to the equation editor and move to the start of the begining of the list. That is:
1. Press EQN
2. Press up and right until the message EQN LIST TOP appears on the bottom line.
3. Input a vector normally.
That is. If it doesn't work, try turning it off and back on. This has been working for me consistently and I cannot reproduce the bug any more.

Nice. Should be added to the bugs-list.

[Paul Dale]

Added the workaround to the bug list.