Reminder: Never connect a generator to home wiring without transfer switch

Mon Aug 30 17:46:44 UTC 2021

During the February 2021 storm that swept through the US, power got knocked
out on my rural street due to a tree coming down and taking out a pole.

While they were waiting for a few more trucks to arrive with a replacement
pole, I got to ask them a few questions.  They said it's standard practice
for them to ground on both sides exactly for the reason that someone might
accidentally connect a generator.  They open the nearest switch on the
upstream side, test to make sure the line is dead, install grounds on all
the wires, then test the downstream side and attach grounds to all the
wires, effectively making the work zone an isolated segment.

I doubt it's "if you follow every step perfectly at all times and never
make a mistake".
There are usually redundancies built-in when it comes to safety.  i.e.
what's the point of installing grounds on the upstream side if you have the
switch open?  If the lines are de-energized, why wear gloves?  If you're
doing all that, why carry an AED?

-A

On Mon, Aug 30, 2021 at 10:19 AM Warren Kumari <warren at kumari.net> wrote:

>
>
> On Mon, Aug 30, 2021 at 12:47 PM Aaron C. de Bruyn via NANOG <
> nanog at nanog.org> wrote:
>
>> I've been following the thread.
>> If I'm dumb enough to back feed through the transformer into the
>> downstream side of the downed line, how is it going to be a problem if
>> linemen are grounding the phases on *both sides* of the work area.
>>
>
> I suspect that there is a non-zero amount of "in an ideal, perfect world,
> when all of the wires are simply lines on a piece of paper, and you can
> look at them from the comfort of your office chair, this is easy" - but, in
> the real world, linesmen are rushing about and trying to get the lights
> back on, cut through the big ash tree that is wedged between the oak and
> the pole, etc. Even the nice idea of "well, just take the conductos and tie
> 'em to ground" means that you need to go trudging through hedges and
> vegetation and tree limbs and lions and tigers and bears, often while it is
> pissing down with rain or baking hot.
>
> I guess I'm missing how we've moved from the "some people are putting
> their lives on the line, let's try to make their life less dangerous" into
> a "weeeeell... if they simply followed these set of steps perfectly at all
> times, and never made a mistake they'd be fine."
> This is NANOG -- I'm sure that we've all followed a set of steps perfectly
> and still managed to redistribute BGP into the IGP, or apply an ACL and
> lock ourselves out of a box, or types "show run" and watched the router
> randomly reboot. Now consider this, but with the added drama of potentially
> ending up dead...
>
> W
>
>
>> That's what Ben seemed to be implying.
>>
>> -A
>>
>> On Mon, Aug 30, 2021 at 9:09 AM Mel Beckman <mel at beckman.org> wrote:
>>
>>> Aaron,
>>>
>>> If you read back in this thread (using the NANOG mailing list archive),
>>> you’ll find this has been explained in great detail. In a nutshell, phase
>>> grounding won’t help if a generator is energized from the customer end, and
>>> this technique was discontinued in the 1970s due to the many deaths that
>>> resulted.
>>>
>>>  -mel
>>>
>>> On Aug 30, 2021, at 9:02 AM, Aaron C. de Bruyn via NANOG <
>>> nanog at nanog.org> wrote:
>>>
>>> 
>>> On Mon, Aug 30, 2021 at 7:35 AM Lady Benjamin Cannon of Glencoe, ASCE <
>>> lb at 6by7.net> wrote:
>>>
>>>> Yes, this is a real and dangerous problem.  Today.  Even with grounding
>>>> I’m afraid.  Source: I’ve been working in an engineering capacity for 27
>>>> years and I have the license you’d need to build a nuclear power plant.
>>>>
>>>
>>> Would you care to educate me on this?
>>> If you ground the phases on both sides of the work-site, how are you
>>> going to end up being a better path to ground?
>>>
>>> -A
>>>
>>>
>
> --
> The computing scientist’s main challenge is not to get confused by the
> complexities of his own making.
>   -- E. W. Dijkstra
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.nanog.org/pipermail/nanog/attachments/20210830/cf31daf8/attachment.html>