Addressable power-strip safety thought

[Peter "Sci" Turpin]

I mentioned the idea on Saturday night that perhaps caging-in these
strips (a box to contain plugs but with wire-access slots and a padlock)
would be the simplest way of stopping people from bypassing the socket
allocations, rather than soldering tools in permanently.

Something that's occurred to me is the issue of tool failure. If a tool
fails catastrophically (bursts into flames, say), or someone gets a
serious shock then the power needs to be disabled instantly.

Each addressable power strip is going to need it's own emergency-off
button. Seems the only way to make them safe and still limit a tool to a
particular power socket. Fumbling for an access card to shut it off via
RFID would take too long in such circumstances.

[Kimball Johnson]

Don't tools already have emergency off buttons separate the the access control?

[Monty]

Step 1) Kill the power to all workshop or entire strip (it's an catastrophic emergency afterall)
Step 2) Cut the power cord of the tool in question.
Step 3) Leave a note.

[Adrian Godwin]

If the power plug is captive, that would leave a bare wire when the
workshop's turned back on. Nobody will be patient enough to wait for a
keyholder to remove the cord, they'll tape it up at best.

Not really sure what problem this is solving. Why should the leads be captive ?

-adrian

[Paddy Duncan]

You can borrow my swiss army knife to cut the power cord.. :)

No virus found in this message.
Checked by AVG - www.avg.com
Version: 2012.0.2221 / Virus Database: 2441/5290 - Release Date: 09/24/12

[Peter "Sci" Turpin]

The idea is that it would enable any tool to be tied into the
access-control system. Each socket on the strip can be remotely turned
on or off apparently.

As I understand it the idea is to ensure people using mains powered
tools are actually trained in using them. So use your access card to
turn on a tool, if you're trained the plug it's connected to will come
on. Only works though if the tool can't simply be moved to another
active socket.
I doubt it'll go on every single tool, probably just the more
specialised ones.

Hmm, though it does raise the point of how to turn it off again. Timed?
Must be deactivated by initial user or loose tool privilege?

[David Murphy]

have it stay on while your card is in the reader.

[Adrian Godwin]

Sounds very constraining. There are some devices that wouldn't mind
being used in particular locations (minilathe, welder, compressor) but
plenty that would be awkward (circular saw, angle grinder, router).

-adrian

On Tue, Sep 25, 2012 at 5:17 PM, Peter "Sci" Turpin <s...@sci-fi-fox.com> wrote:

[Simon]

IIRC the conversation on saturday

One reason to reason to severly restrict access to the network switchable power strips is because we could use them downstream of the UPS for various locations in the space that would benefit from UPS'ed power supply - Coolbot and laser cutter, Network rack, babbage.  If we do that then we MUST do everything possible to ensure that nothing else gets plugged into the UPS just because there is a convenient power socket on the wall.  Hence, boxing in the whole PDU so that all sockets are inaccesable.

I must admit that the discussion didn't cover safety aspects so it is timely that Sci has raised this now.

It is worth pointing out that each socket on the 20 socket network switcable PDU has it's own individual power trip so if a connected device goes faulty then the trip is likely to isolate the socket.
Also, the power strip does have a mechanical isolation switch on it to cut the power to all 20 sockets - it is not the most accesable switch and probably not to be considered for emergency shutoff.

How about if the box that covers the power strip has enough gap such that a yank of a cable will pull it from the socket but leave the plug itself still loose inside the box cover?


I am also encouraged that these power strips are being considered for general RFID card activation for particular machines.  While we don't need to consider preventing other devices being plugged in in this case since there wouldn't be a UPS behind it, we may need to box in the power strip to prevent people from moving a cable from a socket for a tool they are not cleared to use to a socket for one they are cleared for.
The idea of making the box a little bigger so that the box doesn't hold the cables plugged in would work for this as well as the UPS fronted PDUs


Simon

[Charles Yarnold]

I have no plans on using these for the ACNode tool access system.

[Peter "Sci" Turpin]

That sounds like a good idea. If we laser-cut lids from, say, 10mm clear
polycarbonate it seems like it'd be strong enough to handle occasional
tugs but not wear on the cables passing through slots. I say clear since
there were activity indicator lights that would be needed for diagnostic
checks.

[Alex Pounds]

On Tue, Sep 25, 2012 at 05:17:41PM +0100, Peter "Sci" Turpin wrote:
> The idea is that it would enable any tool to be tied into the
> access-control system. Each socket on the strip can be remotely
> turned on or off apparently.

If this captive/remote power idea is considered a good idea (*if*), for
some tools it should be remote-off only. I'd like to avoid something like
the following.

---------- CHOOSE YOUR OWN ADVENTURE ------------------------------------

You are a young, smart, resourceful hacker taking his first few steps with
metalwork. One day you're machining some aluminium on the lathe, when the
tool loses power for no apparent reason.

• If you step away from the work area to investigate, turn to page 32.
• If you attempt to make the tool safe, turn to page 11.


-------------------------------------------------------------------------
PAGE 32:

You walk away from the lathe to the main room to find out what's going on.
Someone suggests you check IRC, but you spot someone in the kitchen who
you think you saw using the lathe earlier.

• If you hop on lovelace to ask about the lathe, turn to page 47.
• If you go to the kitchen, turn to page 19.


------------------------------------------------------------------------
PAGE 11:

Realising that the remote power must have tripped, you turn it off and
reach for the chuck key. As you start removing the tool head you hear a
click and have a split second to realise the inevitable: your fingers will
be broken by the accelerating chuck. Did you push the wrong button, not
engage it far enough, or make some other mistake? With all the blood and
screaming, you'll probably not be too bothered to find out.

- THE END -

-------------------------------------------------------------------------
PAGE 47:

You sit down at Lovelace and ask "wtf?". Someone tells you that you
haven't been trained and shouldn't be using the machine. "ive totally bin
trained" you say. "last wk." "Sorry," says the hacker. "I guess the wiki
didn't get updated." You see a robonaut message saying the power was
re-enabled. As you walk back into the workshop, you see the tool spinning
unattended and hurry back over. Fortunately, nobody else was in the
workshop at the time, but the cutting tool has shattered and your
workpiece is ruined. You realise, with a sigh, that it could have been
much worse.

- THE END -

-------------------------------------------------------------------------

PAGE 19:

You walk into the kitchen and ask some guy if he had any trouble with the
lathe earlier. "No trouble earlier," he says. "But there was something on
the mailing list about the drive slipping. Maybe it's that." Before you
can both go to investigate, though, you hear some commotion from the
workshop. You rush back in, and find a dude at the work table nursing a
cut to his hand. He was using the band saw when the lathe got re-enabled
remotely after someone realised they typed the wrong thing. The sudden
noise combined with the racket of your workpiece being mangled made him
jump while cutting. This time it was a flesh wound, but you definitely owe
him a pint.

- THE END -

--------------------------------------------------------------------------


This was a remarkably long post when all I was trying to say: remote power
on can be dangerous.

--
Alex Pounds
Web Developer & Photographer

http://alexpounds.com/ | http://ethicsgirls.com/

[Peter "Sci" Turpin]

An entertainingly long reply though!

Without wanting to sound trite, I'd counter by saying all tools capable
of being so should probably be fitted with magnetic safety on/off
switches. The type that latch with an electromagnet so they revert to
the off-state if there's a power loss. Probably a good thing to do
regardless of if access-controls are put in place.

Though regarding page 47, I'd hope the ideas for this have never
included using the wiki as the access database. :P I'd expect it to be a
similar setup to the exist door-access card system. I don't think that
can be disabled via the wiki.

[cepm...@yahoo.co.uk]

Not when the tool is fitted with a no-volt relay as all our fixed machine
tools are. They are there so that even a momentary loss of power requires
positive action to restart the machine.

> This was a remarkably long post when all I was trying to say: remote
> power
> on can be dangerous.
>


--

" et cognoscetis veritatem et veritas liberabit vos. "

[Monty]

What's that coming over the hill
Is it a bikeshed? Is it a bikeshed?

[David Murphy]

Where did the "bikeshed" term come from? it's a bit before my time, I assume some attempt to build a bike shed which got stuck in endless over-complex tech hacks while lacking any major details like overall structure or implementation.

[Mark Steward]

http://en.wikipedia.org/wiki/Parkinson's_Law_of_Triviality#When_governance_meetings_devolve_into_2-cents_worth

I really don't think this thread's going anywhere interesting.  Physically locking things in the space is bound to result in disputes, and we've never needed remote control of power, even for the switch.

Let's leave this discussion to die now.

[David Murphy]

didn't you just fullfil the last part of that?

they will end by asking the Secretary to procure further information, leaving the matter to be decided at the next meeting

[Peter "Sci" Turpin]

How can stating something's already been done (no-volt relays) be
bike-shedding? Bike-shedding's all about faffing over tiny details so
nothing actually gets done.
Additionally I wouldn't call the legitimate concern of someone getting
maimed by remotely-activated power-tools a tiny detail. This is just
public brainstorming.

[tom]

so deciding what to make the devices out of:

>> ... we laser-cut lids from, say, 10mm clear 
polycarbonate it seems like it'd be strong enough to handle occasional  ....

before anyones agreed that this is a good idea is not bikeshedding how?

[Peter "Sci" Turpin]

On 26/09/2012 14:49, tom wrote:
> so deciding what to make the devices out of:
>
> >> ... we laser-cut lids from, say, 10mm clear
> polycarbonate it seems like it'd be strong enough to handle occasional ....
>
> before anyones agreed that this is a good idea is not bikeshedding how?

You seem to have cropped the "If" off the start of that sentence, making
it sound like I'm dictating rather than suggesting.

And surely suggesting ways it can be practically implemented is part of
deciding if something is a good idea or not? The logistics have to be
taken into account. I'm not throwing a fit over it not being pink or blue.