Prev: linux booting issue
Next: OMAP DSP starting vector
From: Hans-Bernhard Bröker on 24 Jun 2010 12:16
Am 24.06.2010 17:46, schrieb D Yuniskis:

> But there is no way to *acknowledge* these alarms OTHER THAN
> to recharge (in the case of the phone) or replace (in the case
> of the smoke detector) the battery.

For the smoke alarm, that's the only sensible thing to do. A smoke
detector without a sufficiently charged battery is, for all intents and
purposes, nothing but an ugly, useless brick of plastic screwed to the
ceiling.

It _cannot_ sanely stop trying to sound the alarm without a sufficiently
charged battery in place.

> I.e., it is an incredibly naive "alert"

It would be even more naive to assume that the person who acknowledged
the alarm will _really_ remember to exchange the battery before it runs
out completely, without an occasional reminder. It would be naive to
even assume the person who acknowledged the alarm as much as _told_ the
person in charge of exchanging the batteries about it.

> in both the way it announces itself *and* the
> way it is acknowledged by the user.

For the smoke detector, I think you're wrong on both counts there. It
cannot be announced any other way (because the device has no other
signalling method that is likely to attract attention in time), and
there _is_ no sane way for the user to acknowledge an out-of-power
situation other than to supply power.

> E.g., the last time my AC/DC smoke detectors (both, simultaneously)
> started chirping, the only remedy I had was to *unplug* them!
> Mind you, they are operating on AC *despite* the fact that they
> have *detected* their backup batteries to be low.

So, if you don't care about their backup battery serving any useful
purpose --- why did you by devices with backup batteries in the first place?

From: Hans-Bernhard Bröker on 24 Jun 2010 12:18
Am 23.06.2010 23:08, schrieb d_s_klein:

> Stanley tools makes a tool just for silencing cell-phone alarms.

And sane cell-phones support a mode to silence them completely
(including the battery-off alarm). For somewhat obvious reason, it's
called "airplane mode".
From: AZ Nomad on 24 Jun 2010 12:33
On Thu, 24 Jun 2010 18:18:37 +0200, Hans-Bernhard Br?ker <HBBroeker(a)t-online.de> wrote:
>Am 23.06.2010 23:08, schrieb d_s_klein:

>> Stanley tools makes a tool just for silencing cell-phone alarms.

>And sane cell-phones support a mode to silence them completely
>(including the battery-off alarm). For somewhat obvious reason, it's
>called "airplane mode".

no it's not. airplane mode has to do with connecting to the cell phone
network. It'll stop ringing, but it'll also stop all incoming and
outgoing calls as well as any other cellular network access.

From: D Yuniskis on 24 Jun 2010 12:50
Hans-Bernhard Br�ker wrote:
> Am 24.06.2010 17:46, schrieb D Yuniskis:
>
>> But there is no way to *acknowledge* these alarms OTHER THAN
>> to recharge (in the case of the phone) or replace (in the case
>> of the smoke detector) the battery.
>
> For the smoke alarm, that's the only sensible thing to do. A smoke

I disagree. I would consider a smarter algorithm to begin
periodic "alerts" before the battery is depleted. At some
*long* interval (15+ minutes... maybe even an hour!).
This allows a user to act on the alarm before it becomes
annoying.

As time progresses, the alerts can become more frequent.

Or, the device could allow you to silence the alerts for
~12 hours (by pressing the TEST button) for the first
48 hours (?). After that, it can become more insistent.

> detector without a sufficiently charged battery is, for all intents and
> purposes, nothing but an ugly, useless brick of plastic screwed to the
> ceiling.

What happens if the battery dies while you are away for
the weekend? Or, on a week long vacation? You come home
and there is no annoying chirp -- nor are the detectors
operating! Until your next monthly (manual) test, you
are living without protection.

> It _cannot_ sanely stop trying to sound the alarm without a sufficiently
> charged battery in place.
>
>> I.e., it is an incredibly naive "alert"
>
> It would be even more naive to assume that the person who acknowledged
> the alarm will _really_ remember to exchange the battery before it runs
> out completely, without an occasional reminder. It would be naive to

-----------------------------^^^^^^^^^^^^^^^^^^^^

A reminder every 60 seconds is not an "occasional reminder".
It is an annoyance.

(imagine if your kids asked you when you're going to depart for
the amusement park EVERY 60 SECONDS -- just an occasional reminder!)

The scheme I outlined is more fitting. When you first encounter
such an alert, you *probably* aren't ready to "drop everything"
to attend to it. If it came back and reminded you an hour later,
you'd probably think, "Oh, yeah. I forgot, I was going to go out
and buy some batteries for that thing!".

Eventually, a procrastinator encounters a more insistent alert.
But, at least he is *ready* for it. He *knows* he has been
previously warned and has been procrastinating (for whatever
reason).

E.g., we have a little travel alarm clock that behaves in
a similar fashion. The first time the alarm goes off, it
is very unobtrusive. If you are lying in bed, awake, you
will hear it and act accordingly (whatever that entails).
After a while, it will gradually increase in volume -- become
more insistent.

At any time, you can silence it with the "snooze" bar.
This is only temporary as it will resume once the "snooze"
interval has expired.

The smoke alarm (and cell phone, etc.) could behave similarly
with the additional condition that they would disallow
"snooze" after a certain point in the battery discharge
cycle (note that this cn be controlled by monitoring
battery voltage instead of some fixed timing algorithm)

> even assume the person who acknowledged the alarm as much as _told_ the
> person in charge of exchanging the batteries about it.
>
>> in both the way it announces itself *and* the
>> way it is acknowledged by the user.
>
> For the smoke detector, I think you're wrong on both counts there. It
> cannot be announced any other way (because the device has no other
> signalling method that is likely to attract attention in time), and
> there _is_ no sane way for the user to acknowledge an out-of-power
> situation other than to supply power.
>
>> E.g., the last time my AC/DC smoke detectors (both, simultaneously)
>> started chirping, the only remedy I had was to *unplug* them!
>> Mind you, they are operating on AC *despite* the fact that they
>> have *detected* their backup batteries to be low.
>
> So, if you don't care about their backup battery serving any useful
> purpose --- why did you by devices with backup batteries in the first
> place?

AC powered detectors are mandated in the latest Code.
Precisely because people would remove the batteries
from chirping smoke detectors and *never* replace them
(out of sound, out of mind?). Apparently, lots of
fires (including fatalities) are found to have smoke
detectors present but disabled -- hence the reason for
the Code change.

Given that you have AC power available, a smoke detector
*could* implement the additional smarts outlined above.
At least insofaras its on-line testing of the backup battery.
While on battery, one *could* argue that the cost of those
smarts (some ultralow power logic??) could impede the
operation of the device. I contend that the cost of
emitting a chirp probably exceeds the cost of a simple
state machine to implement the above algorithm.

I believe smoke detectors behave the way they do because they
don't have any sense of state/memory. I.e., they run a
fixed cycle: sample environment, make decision, alarm
if required, sample battery, alert if required, sleep
(to conserve battery). Lather, rinse, repeat.

There is *no* reason why a cell phone is as stupid as
it is! :-/
From: Hans-Bernhard Bröker on 24 Jun 2010 16:08
Am 24.06.2010 18:33, schrieb AZ Nomad:

> no it's not. airplane mode has to do with connecting to the cell phone
> network. It'll stop ringing, but it'll also stop all incoming and
> outgoing calls as well as any other cellular network access.

.... which is exactly the state a phone _should_ be in if its owner's
only reaction to the low-battery alarm is annoyance. Because that
person obviously couldn't care less about whether their phone works or
not, so for the sake of general energy economy and bandwidth
conservation, their phone should be completely off. Arguably, there
shouldn't be a battery in it in the first place. Or they shouldn't even
have a cell phone.


First  |  Prev  |  Next  |  Last
Pages: 1 2 3
Prev: linux booting issue
Next: OMAP DSP starting vector