Boot loop - what's wrong?

[geekmaster]

Another thing that annoys me about this PW3 (apparently a lab125 design decision) is that the USB net drops every time screensaver mode starts. I was in the middle of editting a file when it dropped this time -- I wonder if I can recover my session (probably not, unless I install and use "screen")...

On my K5 and earlier, usbnet login sessions survived screensaver mode. Apparently not so on newer firmwares. Perhaps turning off USB is a power saving feature. I will need to get in the habit of disabling screensavers (lipc command) in my usbnet sessions, I guess...

[knc1]

Mine does not.
As long as the USB cable is connected, the system has a "wake lock" set -
It will show a screen saver when the time-out tells it to, but it does not go to sleep and it will not drop USB networking (or the sshd or telnetd processes).

Now this is on the original 5.6.1.1 firmware - I haven't upgraded my device.
(I could give a rat's ass if it can use Amazon services or not. That is incredibly low on my 'features to support' list.)

= = = =

See:
https://www.mobileread.com/forums/sho...5&postcount=14

You can see that the CRxxx devices are not installed.
In that pic - I sort of made the engineering level tag unreadable with the glue blobs.
But there is a second "PW-3" (tagged) thread with close-up pictures (of both PW-3 and PW-2) if you want to look for differences in mounted parts.
Oops - I was thinking of:
https://www.mobileread.com/forums/sho...d.php?t=249979

but the on/off and reset connections work on that device also.

[geekmaster]

Originally Posted by knc1

Mine does not.
As long as the USB cable is connected, the system has a "wake lock" set -
It will show a screen saver when the time-out tells it to, but it does not go to sleep and it will not drop USB networking (or the sshd or telnetd processes).

Now this is on the original 5.6.1.1 firmware - I haven't upgraded my device.
(I could give a rat's ass if it can use Amazon services or not. That is incredibly low on my 'features to support' list.)

= = = =

See:
https://www.mobileread.com/forums/sho...5&postcount=14

You can see that the CRxxx devices are not installed.
In that pic - I sort of made the engineering level tag unreadable with the glue blobs.
But there is a second "PW-3" (tagged) thread with close-up pictures (of both PW-3 and PW-2) if you want to look for differences in mounted parts.
Oops - I was thinking of:
https://www.mobileread.com/forums/sho...d.php?t=249979

but the on/off and reset connections work on that device also.

Hmm... I suppose I will need to downgrade my firmware then, if dropping USBnet remains an annoying issue (and if it is a "power saving feature" of newer firmware. Or get in the habit of disabling screensavers while "in session".

In the case of the missing switches, I bridged them in various ways (wires, a pair of jeweler screwdrivers, etc.) and they seemed to do nothing. Perhaps I could try soldering wires (or real switches) to them, but my PW3 is reassembled now. No need to open it (until it gets bricked, perhaps).

Ideally, for regular access to the PCB , I would install an access panel (rectangular hole and duct tape?) in the back cover. Or, if switches installed, drill small holes in the back cover (i.e. "paper clip" ports).

EDIT: I just realized that I probably DO want a serial port on this PW3, because the good stuff in diags seems to want you to use the serial port. And I want to experiment with custom u-boot variations, and they are serial-port dependent (at this time) as well. I want to use u-boot as a base for a custom "bare metal" OS I want to play with...

[knc1]

First link of the post you quoted - shows how I cut away my PW-3 to install a Jim.sh/ftx adapter board inside the case.

[NiLuJe]

Nothing has *ever* dropped the USBNet connection for me (on a wide range of devices/firmwares), except telnet on FW 2.x.

So, don't bother downgrading, it works perfectly well on 5.7.4 here .

[geekmaster]

Perhaps the USBnet drops are related to the touchscreen drops. Perhaps those are both symptoms of a bigger problem. Sometimes touchscreen activity returns on its own (or rather, it worked all along, and the DISPLAY is delayed). I did notice while typing a search command that the keypresses seemed to be ignored (i.e. touchscreen not working), then suddenly after about 20 seconds all of those "missing" kepresses squirted into the search bar. They were just stuck in a buffer for awhile...

Strange that factory reset and firmware update did not fix this problem, unless (as knc1 suggested) it could be a voltage regulator (or other hardware components) out of spec.

I suppose it makes sense that the touchscreen and usbnet problems could well be related, in which case it might suck to have a lockup while flashing a new u-boot, eh? If that happened, I would need to learn to boot an image from serial (or wherever), if I could figure out how to set the boot-mode pins on the SoC...

EDIT: A third (probably) related symptom: this PW3 had a dead LiPo battery with an inflated foil pouch, a symptom typical of severe overcharging (to death). The reason this could be related to USBnet and touchscreen dropouts (and USB state changes causing touchscreen dropouts) is that in the kindles, battery charging is under firmware control, and firmware is highly dependent on hardware reliability. That means the current battery borrowed from the PW2 may now be at risk as well. I hope it does not catch fire inside the PW3 (especially while in my pocket). Though more energy-dense (higher power-to-weight ratio) than other lithium battery technologies, LiPo batteries can be dangerous (as seen in LiPo-laden flaming "hoverboards").