Touchpad *mostly* non-functional

[Ozwald]

Hi all; Just recieved my G1G1 XO which is very cool. However the touchpad only works in the up/down axis albeit a little herky-jerky. Not well at all in the left/right axis, only zooming from one to the other while intermittently stopping along the way.

I've tried the de-power and four-finger salute as suggested by the folks at support@… but no joy. If it was consistentl in its badness i'd be more inclined to suspect the h/w but since the behavior varies somewhat after the attempted fixes i'm inclined to believe in s/w voodoo.

The service@… folks replied that they'd recieved 3 or so similar complaints but directed me to the laptop.org folks who have now directed me here.

Any thoughts?

[rbwjrw]

I am experiencing the same flaky touchpad issues on one of my two G1G1 XOs -- reboot and four-finger salute ineffective at resolving the issue.

[jg]

We're working on kernel driver improvements and other improvements to the situation. We hope to have something you might test in the next week or three.

[bernie]

Also see #2804, #4554 and #5268.

From the bug summary, this one seems like a separate issue. Questions:

• does this bug happen all the time?

• does this bug also happen when the laptop is on AC power?

• does it still happen after a chill boot? (chill boot = unplugging the battery and AC power for 10 seconds)

Tests you could do to help us:

• If you have some experience with Linux, please try installing gpm (yum install gpm) and then starting it (service gpm start) and see if it also happens in the console

• Try attaching a USB mouse. Does the problem happen anyway?

[bernie]

Oops, I accidentally changed the owner for this bug.

[Ozwald]

Replying to bernie:

Also see #2804, #4554 and #5268. From the bug summary, this one seems like a separate issue. Questions: - does this bug happen all the time? - does this bug also happen when the laptop is on AC power? - does it still happen after a chill boot? (chill boot = unplugging the battery and AC power for 10 seconds) Tests you could do to help us: - If you have some experience with Linux, please try installing gpm (yum install gpm) and then starting it (service gpm start) and see if it also happens in the console - Try attaching a USB mouse. Does the problem happen anyway?

- USB mouse works - Chill boot no impact - not *yet* that familiar with Linux so i'll have to snuffle around a bit to figure out how

[bernie]

We're working on a procedure that will allow us to get one of these "lucky" units back the the R&D headquarters for further analysis. So far, 2 have been identified.

[TedCH]

Greetings,

I just received my XO. I also have noted a very jumpy touchpad. My box looked opened (all cellophane bags a bit dinged up), the bag with the XO was clearly untaped. I figured maybe someone tested the unit prior to shipping?

In any event, the touchpad is very tempermental. I was using the unit on battery. It took multiple finger strokes to get the curser anywhere I wanted within the XO system (from bottom of screen---activity up to the top (Communities etc.)

If this is some sort of manufacturing defect that we've caught early, I'd also want to swap out the unit I received for one that may not have this idiocyncracy. My box also included no registration information inside, just what looks like a worn fan fold pamphlet.

Cheers,

T

Replying to Ozwald:

Hi all; Just recieved my G1G1 XO which is very cool. However the touchpad only works in the up/down axis albeit a little herky-jerky. Not well at all in the left/right axis, only zooming from one to the other while intermittently stopping along the way. I've tried the de-power and four-finger salute as suggested by the folks at support@… but no joy. If it was consistentl in its badness i'd be more inclined to suspect the h/w but since the behavior varies somewhat after the attempted fixes i'm inclined to believe in s/w voodoo. The service@… folks replied that they'd recieved 3 or so similar complaints but directed me to the laptop.org folks who have now directed me here. Any thoughts?

[andrey]

Hi. I have what appears to be the exact same problem on a G1G1 unit. The touchpad works vertically but it's very jumpy horizontally (it took me a while to realize/notice that vertical movement actually works fine).

The 'chill boot' doesn't make any difference, nor does the four-finger solute. For what it's worth, I updated to joyride-1472 and that didn't change anything (vs. build 650). Firmware is at Q2D07.

I installed gpm and tried that out, and I have the same problem with the console mouse.

Is there anything I can do to help? Any more information I can provide? Also I'll be happy to mail in my XO for investigation, if that would help.

[bernie]

It would be nice if you could examine the log from the touchpad driver:

(become root) # echo 1 >/sys/module/psmouse/parameters/tpdebug # cat /proc/kmsg

Works best in the console (CTRL-ALT-Neighborhood). I bet you'll see the x coordinate not changing while y varies normally.

Another possibility is that your resistive TP is very sensitive and swtiches to resistive mode easiliy. Try gliding your finger very lightly, or try interposing a sheet of paper over the touchpad.

[garlick]

Just a "me too" here. One of my two g1g1's has this touchpad issue.

[andrey]

bernie -- kmesg output attached. Placing a piece of paper over the touchpad didn't seem to change anything.

[ebotee]

http://www.youtube.com/watch?v=Q5UJQB2Is6c This is a video of someone experiencing the same problem.

[bernie]

Andrey, could you watch ebotee's nice video and see if it looks like your problem? Reading your logs, it seems a little different.

I'd also be interested in seeing /var/log/Xorg.0.log to ensure evdev is using the glide sensor in relative mode.

[wvbailey]

Symptom: G1G1 Touchpad is very sensitive in X-direction but only in center of pad; if you use a USB mouse to start pointer in dead center, Y-motion is smooth and seems calibrated to ends of touch-ad but any X-motion sends pointer to far left or far right (where Y-motion is okay).

Experiments:

*(1) Do 5 times: use optical mouse to position pointer at dead-center then do 4-finger salute. Result: 4 of 5 times no improvement. 1 time pad behaved like #2804 but after a minute or so "reverted" to #5575 (actually, this is the 2nd time I've had this happen).

*(2) Put tin-foil over pad (held down with a christmas candle) and connected tinfoil to (presumably) system ground with a ball-clip connected to an audio jack plugged into the "mic" input; positioned pointer in center of screen with optical mas, then did 4-finger salute. No improvement.

*(3a, b) Plastic lid over pad: No improvement. Paper over pad: No improvement

*(4) Ceramic dish over pad: No movement at all. Only very light pressure+proximity causes motion. Pure pressure (with pencil eraser): no motion.

*(5) aluminum foil over entire surface, either not grounded or grounded: reduces Y-axis sensitivity by a factor of about 5 and causes erratic Y-behavior; X-axis still very sensitive but now is harder to evoke the phenomenon.

Conclusions: touchpad is very X-sensitive, but only in center of pad. No experiments were successful excepting (2) which seemed to invoke #2804-like behavior for a while. Conjectures:

* All these touchpad issues are manifestations of the same problem.

* Pad behaves as if there's an X-gain (sensitivity) that is about 10x-50x too high. This has the feeling of either (i) defective ALPS hardware, or (ii) a hardware implementation problem. More likely (ii) than (i). If the "code" is virtually the same for both X- and Y-, why would X-axis not work but Y-axis does work. If the code is the same from machine to machine then you'd expect the code to be faulty in all instances. Can the X- and Y-axes be flipped (in both hardware and software) to see if the problem goes with the axis? * Is there an ALPS interface-IC? If so, has the circuitry around it been implemented correctly -- any accidently-floating pins? Open etch? Bad IC?

* Can someone direct me to a hardware spec sheet? thanks.

* If you need another machine, I can return mine to you. Lemme know, Bill

[andrey]

Replying to bernie:

Andrey, could you watch ebotee's nice video and see if it looks like your problem? Reading your logs, it seems a little different. I'd also be interested in seeing /var/log/Xorg.0.log to ensure evdev is using the glide sensor in relative mode.

No, my problem is different from the video. In the X axis, the cursor jumps around (as my kmsg snippets show) while in the Y it is smooth. It does jump around in the X somewhat predictably, getting more erratic toward the sides, I think, and sometimes being almost usable. Here's the tail end of Xorg.0.log, 'not sure about relative mode, can you tell from this?

(**) Option "CorePointer"
(**) Option "CoreKeyboard"
(**) fake: always reports core events
(II) evaluating device (fake)
(II) XINPUT: Adding extended input device "fake" (type: Void)
(**) OLPC ALPS GlideSensor: always reports core events
(II) OLPC ALPS GlideSensor: Found 2 absolute axes.
(II) OLPC ALPS GlideSensor: Configuring as pointer.
(II) OLPC ALPS GlideSensor: Found 2 relative axes.
(II) OLPC ALPS GlideSensor: Configuring as pointer.
(EE) OLPC ALPS GlideSensor: Unable to parse 'RelAxis 0' as a map specifier.
(EE) OLPC ALPS GlideSensor: Unable to parse 'RelAxis 1' as a map specifier.
(**) OLPC ALPS GlideSensor: Configuring 2 absolute axes.
(II) OLPC ALPS GlideSensor: Checking button DIGI_STYLUS (74)
(II) OLPC ALPS GlideSensor: Checking bit 330
(EE) OLPC ALPS GlideSensor: Button: 74.
(EE) OLPC ALPS GlideSensor: state->btn: 0x823b420.
(EE) OLPC ALPS GlideSensor: Unable to parse 'null' as a map specifier string.
(EE) OLPC ALPS GlideSensor: Unable to parse 'null' as a map specifier string.
(II) OLPC ALPS GlideSensor: Found 4 mouse buttons
(II) OLPC ALPS GlideSensor: Configured 4 mouse buttons.
(II) XINPUT: Adding extended input device "OLPC ALPS GlideSensor" (type: MOUSE)
(**) OLPC ALPS GlideSensor: 2 valuators.
(**) OLPC ALPS GlideSensor: Configuring in Relative mode.
(**) OLPC ALPS GlideSensor: Registering 4 buttons.
(II) OLPC ALPS GlideSensor: Init
(II) OLPC ALPS GlideSensor: On
(**) AT Translated Set 2 keyboard: always reports core events
(**) Option "xkb_model" "evdev"
(**) Option "xkb_layout" "us"
(II) XINPUT: Adding extended input device "AT Translated Set 2 keyboard" (type: KEYBOARD)
(II) AT Translated Set 2 keyboard: Init
(II) AT Translated Set 2 keyboard: On
(II) AT Translated Set 2 keyboard: Off
(II) OLPC ALPS GlideSensor: Off
(WW) Open ACPI failed (/var/run/acpid.socket) (No such file or directory)
(II) No APM support in BIOS or kernel
(II) AT Translated Set 2 keyboard: On
(II) OLPC ALPS GlideSensor: On
(II) AT Translated Set 2 keyboard: Off
(II) OLPC ALPS GlideSensor: Off

[andrey]

er, I guess from the above:

(**) OLPC ALPS GlideSensor: Configuring in Relative mode.

so I suppose it is using relative mode.

[bernie]

I've been told rbwjrw's laptop just arrived and is on my desk for inspection. I'll see it soon, hold on!

[bernie]

#5698 marked as duplicate of this bug.

[bernie]

After extensive testing, I'm convinced it's a hardware or firmware problem in the ALPS side.

Specifically, swapping the touchpad assembly with another one cured the problem on the laptop donated to science by rbwjrw.

We may want to ship this unit to ALPS for further investigation.

[kimquirk]

Shipping this to Quanta to manage the alps relationship and we may need another test in mfg to weed out the bad touchpads.

[wad]

Failure analysis shows that this is a failure due to a microscopic crack appearing in the flex cable connecting the touchpad/keyboard electronics module to the touchpad. The manufacturer is working on increasing the reliability of the cable to eliminate this problem.

Final testing has been added instructions to the touchpad test to ensure that both directions are tested.

We expect that this is an infant mortality problem, and shouldn't appear in the field on units received in working condition.

[kweidner]

FYI, my trackpad started misbehaving as described here after I had repaired my keyboard. I disassembled it again and noticed that the metal foil near the bottom right (4-lead) touchpad ribbon cable had two small ears on the wrong side of the ribbon, apparently touching and short circuiting it. After moving those two foil parts out of the way, it's working fine again.

So I guess this may be due to misassembly also, which is much easier to fix than cracks...

[cjb]

Replying to kweidner:

FYI, my trackpad started misbehaving as described here after I had repaired my keyboard. I disassembled it again and noticed that the metal foil near the bottom right (4-lead) touchpad ribbon cable had two small ears on the wrong side of the ribbon, apparently touching and short circuiting it. After moving those two foil parts out of the way, it's working fine again. So I guess this may be due to misassembly also, which is much easier to fix than cracks...

Adding Wad and Richard to CC so they know to keep an eye out for this, thanks.