Arrow keys intermittently doubling

[TravisE]

Has anyone else ever experienced a single press of the arrow movement keys sometimes moving the cursor two positions instead of one? I'm getting this quite frequently on my Prime, and it's becoming a nuisance. It happens with all four directions but not any other keys that I've seen. Roughly one in every few dozen times I press an arrow key (it seems random with no discernible pattern), it will move the cursor in a menu or in the editor twice for one press instead of once.

For example, in the program editor while I was trying to indent a block of code by repeatedly pressing down, space, down, space, this caused a line to be skipped every screen or two of code. I can also end up selecting the wrong item in a menu because the cursor did not move the correct number of times.

I can't be certain exactly when I started noticing this, whether it was like this since the beginning or whether it started more recently. I'm currently running firmware 10637.

[cyrille de brébisson]

Hello,

These keys are repeating. Do the selection move twice because of a timed repeat or because of a double detection in your opinion?

Have you tried using the backspace key (for deleting characters)? do you see the same pattern?

Cyrille

[TravisE]

It feels like a double detection because I'm not holding the key down long enough to repeat when it happens, just tapping it.

Yes, it appears that problem does occur with the backspace key as well. It looks like I hadn't noticed that.

I haven't determined yet if the time interval between key presses seems to matter, but in my testing I see the behavior fairly often when tapping the keys at roughly 0.5-second intervals.

[teenix]

(12-08-2016 10:45 AM)TravisE Wrote:  It feels like a double detection because I'm not holding the key down long enough to repeat when it happens, just tapping it.

Yes, it appears that problem does occur with the backspace key as well. It looks like I hadn't noticed that.

I haven't determined yet if the time interval between key presses seems to matter, but in my testing I see the behavior fairly often when tapping the keys at roughly 0.5-second intervals.

Perhaps certain key contacts might be dirty, corroded or worn causing noisy operation which results in a longer key [bounce] period than the calculator software can handle.

Maybe a peek inside might reveal the culprit.

cheers

Tony

[Tyann]

Bonjour
J ai déjà observé ce phénoméne moi aussi
même pendant le déroulement d'un programme
utilisant ISKEYDOWN.

Hello
I have already observed this phenomenon too
Even during the course of a program
Using ISKEYDOWN.

[DrD]

I experimented with this issue by filling the command line with alphabetically ordered text characters. Then about every third key press, arrow left or arrow right, I was noticing this double shift, as well. My testing was with a hardware version C machine, OS V0.048.635.

I was surprised to see this happen consistently, because I hadn't noticed it before, probably due to not using the arrow keys enough to pick up on that quirk. My key press cadence didn't seem to matter. It doesn't 'feel' like key bounce, more like the effect of a key buffer issue. Do the arrow keys happen to be double buffered?

-Dale-

UPDATE:

I selected the "C-F-O On" diagnostics, and then tried some of the key test choices, followed by the option 9 reset. Since then I have NOT had any further problem with the arrow key (left or right). I've tested this in the same way as originally described, several times, over approximately 15 minutes with no further skipping problem.

-Dale-

[cyrille de brébisson]

Hello,

"but in my testing I see the behavior fairly often when tapping the keys at roughly 0.5-second intervals"

Ha, this might be the key to the issue (pun intended
I would venture that the key release is not detected and that when you press the key again matches with the moment when the calculator tries to see if the key is still pressed, and it detects it as still pressed, and repeat.

This would not be a HW problem.

Cyrille

[TravisE]

Hmm, perhaps the following might be related, too, then: On graph screens, I have noticed previously to this that pressing and holding a cursor key in one direction so that it accelerates, then releasing it and immediately pressing a different direction quickly enough does not “reset” the acceleration like I would intuitively expect. This has thrown me off a few times.