Separate ground wire to panel to ground outlets?

This question was posted in another forum by someone that I know and I thought that I would try posting it here to get some feedback on his behalf:
He wrote,
"In a house that has ungrounded, 2 prong outlets, with wiring in good condition, is it acceptable (and allowed by current electrical code) to run a separate ground wire from the outlets back to the ground at the panel, in order to be able to replace the 2 prong outlets with 3 prong outlets? Existing wiring is in very good shape."
Thanks.

On 10/08/2014 8:39 AM, TomR wrote:
  Quoted Text Here  
Yes

On Wednesday, October 8, 2014 9:39:45 AM UTC-4, TomR wrote:
  Quoted Text Here  
*Article 250.130(C) in the National Electrical Code gives the options, but the short answer is yes.
John Grabowski http://www.MrElectrician.TV

Tom,
As the others have said the answer is yes. But check to see if the outlet boxes are grounded. You might get lucky.
Dave M.

  Quoted Text Here  
Thanks John and all for your replies. I passed on the information. I had also mentioned to him that depending on the location of the outlets, it may be almost just as easy to run new NM wiring to the outlets as it would to run separate ground wires to the same outlets. And, I mentioned the option of replacing 2-prong outlets with 3-prong GFCI outlets, without running any new ground wires etc., as long as the GFCI had a sticker that said "No equipment ground"
As it turned out, the outlets in question were first floor outlets above an open ceiling basement. So, he ended up having an electrician just run all new NM wiring to those outlets on the first floor -- basically because the electrician said that it would take almost the same amount of effort to do that as it would to run separate ground wires.

On Friday, October 10, 2014 11:31:22 AM UTC-4, TomR wrote:
  Quoted Text Here  
In that situation I would agree, that's the right way to do it. The only reason to run the separate wires and/or use GFCIs is if replacing the cables is prohibitively difficult.
nate

On 10/11/2014 9:21 AM, N8N wrote:
  Quoted Text Here  
> way to do it. The only reason to run the separate
  Quoted Text Here  
> is prohibitively difficult.
  Quoted Text Here  
It sure looks like an easy decision. I'd go with new Romex, and sockets. Might even go with new breakers, while you have the panel box open.

  Quoted Text Here  
That is pretty much true and even considering the cost of the wire, it is the labor that will be the expensive part.
As another poster said, you really need to verify that there is not a ground in the existing cable. There was a period of time in the 60s when you were required to ground the boxes but you didn't need 3 prong outlets. The grounded Romex was available in the 50s. The house I grew up in, built in 53, had 2 prong outlets and 3 wire Romex. We retrofitted the 3 prong outlets, pigtailing from the box. These days it is much easier with self grounding receptacle.

On 10/08/2014 09:39 AM, TomR wrote:
  Quoted Text Here  
Regardless of electrical code, would you buy a house where some jackleg had done that?

  Quoted Text Here  
what is the difference? a wire is a wire.

In
  Quoted Text Here  
Thanks. That's an interesting point, especially since the person who originally had the question stated that the existing wiring was in good shape. That does make me think that maybe it was the kind of wiring that included a ground wire to metal outlet boxes. But, since he said that he had an electrician rewire the boxes (I think he meant run new wire to the boxes) instead of just running new ground wires, I am assuming that the electrician would have noticed if the metal boxes were already grounded. Of course, I don't know for sure, but that's my guess.

  Quoted Text Here  
or he may have noticed, said nothing, and replaced the wire anyway.

On Fri, 10 Oct 2014 12:42:11 -0700, "Pico Rico"
  Quoted Text Here  
He could easily make the case that the 16ga ground wire in old Romex is not up to current code (true) but it is plenty to operate the over current device in a fault.

> > Regardless of electrical code, would you buy a house where some
  Quoted Text Here  
If I saw a separate wire pulled to an outlet box I'd be uneasy about where it went and whether it really tied into the proper place. (In fact I have seen that and I was uneasy because some kid did it in my old house and then put up two different ceilings that made it very hard to trace back. In that case the ground wire did NOT go back to the panel, but to a clamp on a cold water pipe. While the outlet appeared koshed with an outlet tested, a replumbing job with plastic pipe could have ended up with an open ground.)
I'm with Pyped and several other posters who advised that running new NM was the was to go.
I'd bet a competent home inspector would red flag a ground wire going to places unknown, especially if it ran under sheetrock or stapled ceiling tiles or was in some other way untraceable visually.
To be sure it was tied into the panel you'd have to hire an electrican to trace it with a fox and hound. I am surprised the NEC allows a separate ground to be run outside the main cable sheath or conduit. If it runs somewhere other than along the main cable or conduit, the chance of someone disconnecting it at some future time because it seemed unrelated to the 110VAC wiring is another risk.
GFCIs or new NM would certainly be the preferred way to do things, and I'd rate those solutions as two or three times as good as a new ground wire, especially if it didn't at least run along the old wire so that it was obvious it was related to that old, ungrounded cable.

  Quoted Text Here  
Well, that's a good explanation. My initial thinking is that it would be placed along the entire run back to the panel. But, that is how **I** would do it.

On Fri, 10 Oct 2014 22:49:14 -0400, "Robert Green"
  Quoted Text Here  
Until the 70s that was legal.

On Fri, 10 Oct 2014 20:01:52 -0700, "Pico Rico"
  Quoted Text Here  
On a stem wall house... I have seen this done by running a 4ga "bus" from the (service entrance/main) panel into the crawl space and pushing 12ga wires up to each receptacle, bugged to the #4. If you drive a rod at the far end of this #4 it becomes a grounding electrode conductor and does not need physical protection. It is also specifically allowed as a point you can attach a grounding conductor to.

  Quoted Text Here  
But stupid. And that's why they outlawed the practice. The wire went off to parts unknown (in my case) and some small amount of current leakage apparently caused galvanic corrosion to occur where it was clamped to the waterpipe. It caused a pinhole leaked to develop. All tucked away behind a stapled ceiling and hard to find without making a hell of a mess. I would not even had known about it until I took the ceiling down (revealing all sorts of other nasty surprises). I am sure the previous owner installed that stapled up ceiling to keep those sorts of issues from the eyes of a good home inspector.
It was also far enough away from the service entrance that a replacement of a section of copper with plastic where it entered the house would have broken the connection to ground.
In my mind, those are two good reasons why they changed the code and why such connections should be removed and done according to modern rules whenever they are discovered even if they are still grandfathered.
As for the OP, what was legal in the 70's is moot.

On Saturday, October 11, 2014 10:00:19 AM UTC-4, Robert Green wrote:
  Quoted Text Here  
f
d a
  Quoted Text Here  
d
of
  Quoted Text Here  
*Actually the 2014 code does still permit the connection to the waterpipe. Article 250.130(C)(1) states that is is permitted "At any accessible point on the grounding electrode system as described in 250.50". The waterpipe is part of the grounding electrode system. The water meter and water heate r bonding jumpers would have to be in place and the clamp would need to be approved for the type of metal piping.
That is what I think happened in your case. The disimilar metals of the gr ound clamp (Brass) attached to a steel pipe acted as a battery. Just like a water heater. The clamp should not have been buried in a finished ceilin g. The ground wire should have been run over to the water meter location w here it would have been accessible and could have been clamped to the water pipe or the grounding electrode conductor.

On Saturday, October 11, 2014 1:42:27 PM UTC-4, John G wrote:
  Quoted Text Here  
off
  Quoted Text Here  
he
  Quoted Text Here  
ind a
  Quoted Text Here  
uld
  Quoted Text Here  
l
d
a
t of
  Quoted Text Here  
y
. Article 250.130(C)(1) states that is is permitted "At any accessible poi nt on the grounding electrode system as described in 250.50". The waterpip e is part of the grounding electrode system. The water meter and water hea ter bonding jumpers would have to be in place and the clamp would need to b e approved for the type of metal piping.
  Quoted Text Here  
ground clamp (Brass) attached to a steel pipe acted as a battery. Just lik e a water heater. The clamp should not have been buried in a finished ceil ing. The ground wire should have been run over to the water meter location where it would have been accessible and could have been clamped to the wat er pipe or the grounding electrode conductor.
My understanding is that the water pipe is NOT considered part of the groun ding electrode system, only something that needs to be bonded *to* the grou nding electrode system, so relying on a metal pipe for a ground for a previ ously ungrounded receptacle is no longer allowed, as others have already st ated.
nate