Re: Wake on USB event

On 2/1/2025 6:12 PM, Dave Platt wrote:

In article <vnjqe8$3ojnb$2@dont-email.me>,
Don Y  <blockedofcourse@foo.invalid> wrote:
 

Note, this *was* working correctly.  And, an identical keyboard
plugged into the same physical port STILL works correctly.
I.e., something has changed *in* the keyboard.

 Three thoughts:
 (1) Does this keyboard have an integral touch-pad or joystick?
     If so, possibly some spurious touch or motion detection in
     the electronics there might cause the keyboard to send in
     a motion event when polled.

No.  And, any sort of "action event" would also leave side-effects
visible on the screen.
E.g., it would "type" a letter/number/symbol into the Notepad
document that I described elsewhere.  Or, would move the text cursor.
Or, the MOUSE cursor.  Or, enable/disable caps/scroll/num -lock
indicators.  Or, invoke HELP (F1), Find Next (F3), time/date (F5),
move to the beginning of the line (HOME), end of line (END), previous
page (PGUP), next page (PGDN), etc.
I.e., *most* keystrokes would leave visible evidence of their
occurrence.
AND, would occur while the machine was NOT sleeping, as well.
So, any event either has to be something that has no visible
consequences (e.g., a shift/control key spontaneously claiming
to be active -- but, why not visible WHILE in normal use???)
OR, as Sylvia suggested, just some spurious notification
that serves to awaken the PC and, when *it* polls the keyboard,
it sees "nothing to report".
If it can so reliably awaken the machine EACH time it is commanded
to sleep, then you would assume it would manifest while I was actively
using the machine!
[The "spurious notification" fits the bill; I would never notice if the
keyboard was routinely saying "look at me!" -- if, when the PC
actually "looked", it saw NOTHING!  Just some extra bus activity...]

(2) I once had a problem with a Macintosh II which would spontaneously
     power up (not just come out of idle).  This would happen in the
     mornings, not long after my wife and I got up and started our
     days.  The chain of events turned out to be (cold house),
     (hot shower), (steam and moisture traveling down the hall and
     down into the den), (moisture condensing on the cold keyboard
     PC-board), (very slight current leak between ground and the
     ADB "power on!" button pads), (BONG!)

The keyboards on my first two Compaqs would (over time) exhibit
flakey behavior due to high impedance implementation.  I would
periodically disassemble the keyboards -- including removing the PCB
behind the keys -- and wipe everything down with alcohol to
restore reliable operation.
These don't seem to have a similar problem (I have 6 or 7 of these
in service, presently).

     Cleaning the PC board in the area of the power-on key's pads,
     and then dabbing the traces with some clear nail polish,
     sealed the high-impedance area away from moisture and the
     problem went away.
      Any chance you've had a fluid spill in the keyboard?  Giving

No.  And, I don't smoke (smokers' keyboards are the pits!)

     it a good flush-out with an electronics-compatible cleaner,
     a distilled water rinse, and a final rinse with isopropyl
     might eliminate the problem.

I periodically clean all of the keyboards as the keycaps inevitably
collect the oils/dirt from fingers.  Remove keycaps and soak in
soapy water.  Vacuum keyboard assembly (dust and crud that get
"behind" the keycaps -- nothing can get *into* the keyboard as the
keycaps cover the only points into the mechanisms).  Wipe down with
alcohol (soapy water would migrate into the keyboard but the alcohol
will evaporate quickly).
Keyboards end up looking "like new"  :>

(3) Every time I've used a PC with a "wake up from sleep" capability,
     its BIOS let me set which sorts of events were and were not
     allowed to bring it out of hibernation.  If you check your BIOS
     you may be able to disable the "wake up on keyboard/mouse activity"
     feature.

But I *want* the wake-on-USB feature.  I just want it to work as
intended; not as if invisible actors were at play!
I leave my workstations on 24/7/365.  I count on them "sleeping"
when I turn my attention away from them (to other activities).
Then, expect them to come back online quickly when I click mouse
or press a key (MOVING the mouse doesn't seem to cause a wake event).
It took me some time to notice this "problem" because I can never
be sure when I've been "away" long enough for a machine to sleep.
So, if it is awake when I return, I assume I've returned quicker
than the sleep timer.  If it's asleep, I figure I was "away"
for longer.
Sleeping is a costly event as all of memory has to be flushed
out to disk.  With 100MB/s disks, it takes a while to flush
100+GB!  (it's really annoying to find myself waiting for it
to finish GOING to sleep before I can wake it back up!  :< )
[Some apps apparently disable this inactivity reporting.  And,
some apps have other timeouts -- e.g., I can have a TELNET
session active and the machine can sleep but the session
won't automatically close unless the machine has been asleep
longer than the keepalive interval for the session]