workshop feeder circuit questions

Reply

  #1  
Old 04-02-06, 08:09 PM
Member
Thread Starter
Join Date: Apr 2006
Posts: 13
workshop feeder circuit questions

I am wiring up my workshop and have a couple of questions. I am putting a 100 AMP load center. I am planning on feed it off of a 100 amp breaker off my main panel. I fall under the 2005 NEC. questions are:

1) it is a 320' run. can I use 2/2/4 (no ground) direct burial alu wire? I think I am fine with the amps, but do I have to worry about voltage drop?

2) If I have less then 6 breakers, do I need a main disconnect in the workshop load center?

3)I will be connecting the ground to the rebar in the workshop. Should the neutral and ground be bonded in the workshop load center? My understanding of the NEC is that this should only happen at a service entrance, but I am not sure if this would be considered a service entrance (since it is a separate building) or a sub-panel.

Thanks in advance for the help
Jon
 
Sponsored Links
  #2  
Old 04-02-06, 11:17 PM
bolide's Avatar
Member
Join Date: Jan 2006
Location: PA
Posts: 1,909
>1) it is a 320' run.

That's a long way.

> Can I use 2/2/4 (no ground) direct burial alu wire?
Why not pay the extra $32 dollars and have a ground?
At that distance you won't have any other of any practical value.


> I think I am fine with the amps, but do I have
> to worry about voltage drop?
Yes. You minimize the voltage drop by increasing the ampacity of the feeder.


> 2) If I have less than 6 breakers, do I need a main disconnect in the workshop load center?

No. But it's a good idea.


> 3) I will be connecting the ground to the rebar in the workshop.
> Should the neutral and ground be bonded in the workshop load center?

That depends.


> My understanding of the NEC is that this should only happen at a service entrance,
> but I am not sure if this would be considered a service entrance (since it is a separate building) or a sub-panel.

That depends.
Will you ever have any other metallic services to the premises (telephone, cable, etc.)?

If so, you must run four-wires and not bond your panel.
If you run three wires, you must have no other metallic pathways and you must bond and earth your neutral.


You really should run four-wires now and avoid any issues later.
 
  #3  
Old 04-03-06, 04:54 AM
Member
Join Date: Sep 2003
Location: Central New York State
Posts: 13,973
A four wire run is not only a good idea, it may be required. If you have any other metal between the buildings, such as water pipes, telephone line, computer cable, television cable, etc.), then a four run is required.
 
  #4  
Old 04-03-06, 11:07 AM
Member
Join Date: Jan 2004
Location: Oregon
Posts: 1,219
With a 320 foot long run, you will need to give strong consideration to voltage drop.

The first thing that you need to do is determine how much load will be connected to this feeder. A simple rough guide is to assume the full rating of the feeder as your design current flow, but this is very often a serious overestimate of the current flow. Will you be running large sporadic loads such as a welder or large compressor? Electric heaters? Etc.

Next you must determine how much voltage drop you can tolerate. A common suggestion is a maximum of 5% at the end of your branch circuits connected to this supply, but this is only a rule of thumb. If you are only going to run large heaters (say a kiln) and a laptop (with a universal 100V-240V power supply), then you could actually tolerate more than a 10% voltage drop. On the other hand, if you are setting up a music studio, then you might need less than 1-2% drop.

Finally you do a voltage drop calculation. You take your supply voltage and multiply by the allowed percentage drop. This gives you the voltage lost in the wires. You then divide this voltage by your design current, which gives you the allowed wire resistance. Divide again by 2 to get the allowed resistance for a single wire in your circuit, and then divide by the length of the circuit (in 1000's of feet) to get the allowed resistance per thousand feet. Then select your conductor.

For example, presume a design current for 80A, and an allowed drop of 3% with a 240V supply:
240V* 0.03 = 7.2V
7.2V/80A= 0.09 ohms per circuit
0.09 ohms / 2 = 0.045 ohms per conductor
0.045/0.320 = 0.140 ohms per 1000 feet.

To meet this voltage drop requirement, you need to use 3/0 aluminium conductors.

Note that when you increase the size of your supply conductors to reduce voltage drop, you must also increase the size of your equipment grounding conductors proportionately. This can make the use of some cables problematic, because the equipment grounding conductor in the cables may be too small.

You must provide a grounding electrode system (ground rods, etc) since this is a detached structure. This is true even if you run an EGC from your main service.

You have a choice about bonding the neutral. If you have no other _bonded_ metallic paths, (metal water pipes, certain communications lines, etc) then you are permitted to make this connection 'like a service entrance', that is by bonding ground and neutral at the panel. The general advice on this board is to run a separate EGC and then connect the panel like a normal subpanel, with separate ground and neutral. However give the length of this feeder, I believe that treating this feeder as a service is a better approach if possible. You will need to evaluate if there are any other metallic paths between these two structures.

-Jon
 
  #5  
Old 04-03-06, 11:43 AM
bolide's Avatar
Member
Join Date: Jan 2006
Location: PA
Posts: 1,909
> However given the length of this feeder, I believe that treating
> this feeder as a service is a better approach if possible.

I'll bite.

Why?

The only rationale I have ever heard is lower cost for the wire.
What's yours?
 
  #6  
Old 04-03-06, 01:06 PM
Member
Join Date: Jan 2004
Location: Oregon
Posts: 1,219
The long feeder means increased impedance between the supply conductors and the surrounding bonded metal. To make an accurate determination of which approach is 'best' would require a detailed engineering analysis, which I have not done.

As long as the requirements of article 250 are properly met, either approach will meet the NEC requirements.

-Jon
 
  #7  
Old 04-03-06, 02:41 PM
Member
Thread Starter
Join Date: Apr 2006
Posts: 13
Originally Posted by winnie
For example, presume a design current for 80A, and an allowed drop of 3% with a 240V supply:
240V* 0.03 = 7.2V
7.2V/80A= 0.09 ohms per circuit
0.09 ohms / 2 = 0.045 ohms per conductor
0.045/0.320 = 0.140 ohms per 1000 feet.

To meet this voltage drop requirement, you need to use 3/0 aluminium conductors.
Thanks for all the info. In the above calculation, where are you getting the .320. I assume that is the resitance for alu 2/0 wire? Can you point me to a table where I can look this up?

Also, if I do run an equipment ground from the main panel, how big does it need to be (100 amps)? Is there a place I can look this up? Do i use bare copper that I can bury direct with the service wire, or does it need to be insulated (otherwise it seems to me it would be just like having a seperate gound at the workshop panel)

thanks agian!!
 
  #8  
Old 04-03-06, 03:13 PM
Member
Join Date: Jan 2004
Location: Oregon
Posts: 1,219
The 0.320 is the distance of the run in 1000's of feet. 320 feet = 0.320 1000feet

So in the above calculation, I determined a maximum of 0.045 ohms per conductor. Then I divided by the length (0.320) to get a maximum allowed resistance of 0.140 ohms per 1000 feet.

Properties of conductors may be found in Table 8 of Chapter 9 of the NEC. Among other things, this table lists the conductor resistance in ohms per 1000 feet. 2/0 aluminium has a resistance of 0.159 ohms/1000 feet; 3/0 has a resistance of 0.126 ohms per 1000 feet.

Note that I made an error in the above: the resistance is listed at 75C; but the wire will be cooler than this. Table 8 has a note which provides the temperature correction. In the above example calculation, 2/0 would probably be fine rather than 3/0.

Also note that the above calculation is based upon the stated assumptions about load and permissible voltage drop; to use the calculation you need to substitute in the correct values.

As to size of a 'ground wire': The table to use is NEC table 250.122. This will tell you that for a 100A rated circuit, the equipment ground conductor needs to be a #8 copper or a #6 aluminium conductor. _HOWEVER_ you are required to increase the size of this conductor in proportion to any increase in size of your supply conductors. Since you would ordinarily use a #1 conductor for a feeder of this rating, if you calculate that you need to use #2/0 conductors for reasons of voltage drop, then you would need to use a #6 copper or a #4 aluminium equipment ground conductor.

I believe that you would need to use insulated conductors rated for direct burial for the equipment ground conductor. However I have seen a suggestion that if you use a thick enough copper conductor, then it can be bare, and would serve both as an equipment grounding conductor _and_ as a grounding electrode. In this case the wire would probably need to be #2 copper, as specified in 250.52(A)(4) for 'ground rings'. I would not use a bare conductor without first confirming such use with your local electrical inspector, who will be familiar with soil conditions and corrosion issues.

If you use an aluminium feeder _cable_, then you can get 4 conductor cable which will include the current carrying conductors as well as the equipment grounding conductor.

-Jon
 
  #9  
Old 04-03-06, 08:56 PM
bolide's Avatar
Member
Join Date: Jan 2006
Location: PA
Posts: 1,909
Originally Posted by winnie
The long feeder means increased impedance between the supply conductors and the surrounding bonded metal.
Correct.


> To make an accurate determination of which approach is 'best'
> would require a detailed engineering analysis, which I have not
> done.

I thought you stated: " I believe that treating this feeder as a service is a better approach..."


> As long as the requirements of article 250 are properly met,
> either approach will meet the NEC requirements.

Later you explained, "you are required to increase the size of this conductor in proportion to any increase in size of your supply conductors."

Of course, upsizing the ECG has precisely the effect of decreasing the impedance from the supply conductors to the fault-current return path.


At some distance the impedance will be so great that the OCPD will not open even for a bolted fault.
(I mention this to establish an upper boundary at which both fail.)

Below this I hope that you agree that an ECG of the same size as the neutral must always perform at least as well as the neutral itself.

If you don't agree, please give some numbers as a counter example.

If you do you agree, then I respectfully suggest that the only "engineering study" that is needed is to calculate the ECG size that will perform as well as required to operate the largest OCPD on the circuit within the required response time with a wide margin for safety.

Of course, for a 100A circuit, the ECG needs to be able to handle what? 200A? So it can be 1 ohm/kft and still carry all that current. Therefore, #6 Al is more than adequate.
There's my "study". From this study I infer that the #4 Al (at around 0.5 ohm/kft) easily carries 200A fault current. In fact, it can carry 400A for the 320', so the margin of safety is there.

I do see how local neutral and local earth can diverge at the distant location.
An electrode can do little to hold them together.

Local neutral will rise and fall as the loads change.
Local earth and the fourth wire ECG will stay at a constant level.


Even considering this, I still don't see how omitting the ECG could be "a better approach" than having it.
 
  #10  
Old 04-04-06, 04:13 AM
Member
Join Date: Jan 2004
Location: Oregon
Posts: 1,219
I _believe_ that for this long feeder to a detached structure, bonding ground and neutral at the detached structure will be 'better' than keeping ground and neutral separate as for a normal subpanel. To firmly establish this one way or the other would require a detailed engineering analysis which would take into account the preferences and requirements of the end user, local environmental conditions, and available engineering materials. Either approach is accepted as being compliant with the NEC if conditions regarding bonded metallic paths are met.

Factors to consider would include: cost of the EGC, effective ground fault path considerations, transient potential differences between neutral and 'local ground' in the detached structure, etc.

I certainly agree that an EGC that was the same size as the neutral will always perform as well as the neutral itself as an effective ground fault current path.

However I stated that 'the long feeder means increasing the impedance between the supply conductors and the surrounding bonded metal'. I am comparing the 640 foot round trip path (neutral in subpanel to main bonding jumper and then back to EGC in subpanel) with a direct ground to neutral bond. 640 feet of conductor versus 0 feet of conductor. I am sure that you will agree that 640 feet of aluminium conductor will have a greater impedance than a few inches of bus bar. If you know of bus bars with greater impedance than 640 feet of aluminium conductor, then I'd appreciate knowing about them so that I can avoid buying them.

In the absence of any loads, the round trip path for the ungrounded conductor is the same in either case, but any loads, especially motor loads, will act to stabilize the voltage between these two conductors.

I would further note that in your ground fault current calculation, Soares suggests using 8 times the OCPD trip rating in order to place the device into is 'instantaneous trip' region. The actual multiple is not defined in the NEC; simply that an effective ground fault current path must facilitate the operation of the OCPD. The #4 Al EGC suggested by simply upsizing the standard #6 EGC in proportion to the ungrounded conductors will provide a bolted fault current of about 560A, with doesn't meet Saores' suggestion, but probably provides an 'effective ground fault current path'.

-Jon
 
  #11  
Old 04-04-06, 09:03 AM
bolide's Avatar
Member
Join Date: Jan 2006
Location: PA
Posts: 1,909
> I _believe_ that for this long feeder to a detached structure,
> bonding ground and neutral at the detached structure will be
> 'better' than keeping ground and neutral separate as for a
> normal subpanel.
Do you have any evidence that we could review from other installations to support your belief?


> To firmly establish this one way or the other would require a
> detailed engineering analysis which would take into account
> the preferences and requirements of the end user,
Such as?


> local environmental conditions,
Such as?


> and available engineering materials.
What does this mean?


> Either approach is accepted as being compliant with the NEC
> if conditions regarding bonded metallic paths are met.
But only four-wire always complies.


> Factors to consider would include: cost of the EGC,
Of course I already cited the tiny additional cost.

> effective ground fault path considerations,
Such as?


> transient potential differences between neutral and 'local ground'
How is this a factor?
Or more to the point, how can this factor ever weigh in the favor of three-wires?


> in the detached structure, etc.
Definitely we're employing the same local grounding in the detached structure for four-wire as for three-wire.

So there is no difference here at all.


> I certainly agree that an EGC that was the same size as the
> neutral will always perform as well as the neutral itself as an
> effective ground fault current path.
Then this is no reason not to use four-wires.


> However I stated that 'the long feeder means increasing the
> impedance between the supply conductors and the
> surrounding bonded metal'.
Of what significance to performance or safety is this?


> I am comparing the 640 foot round trip path (neutral in
> subpanel to main bonding jumper and then back to EGC in
> subpanel) with a direct ground to neutral bond. 640 feet of
> conductor versus 0 feet of conductor.

There is no 640' path for neutral current.
Why would neutral current need to return on the ECG?

> I am sure that you will agree that 640 feet of aluminium
> conductor will have a greater impedance than a few inches
> of bus bar.
I also agree that liquid water is wet. But how is that relevant?


> If you know of bus bars with greater impedance than 640 feet
> of aluminium conductor, then I'd appreciate knowing about
> them so that I can avoid buying them.
How is this relevant?


> In the absence of any loads, the round trip path for the
> ungrounded conductor is the same in either case, but any
> loads, especially motor loads, will act to stabilize the
> voltage between these two conductors.
Between which two?

> I would further note that in your ground fault current
> calculation, Soares suggests using 8 times the OCPD trip
> rating in order to place the device into is 'instantaneous trip'
> region.
> The #4 Al EGC suggested by simply upsizing the standard #6
> EGC in proportion to the ungrounded conductors will provide a
> bolted fault current of about 560A, with doesn't meet Soares'
> suggestion, but probably provides an 'effective ground fault
> current path'.

A #1 Al EGC will handle 800A. If you want more, just use an ECG that is larger.

We agree that the ECG can always be made large enough to perform as well as the neutral, so this is not a reason to use only a three-wire installation.


Advantages of three-wire:
  • Slightly lower cost.

Advantages of four-wire:
  • Can have other metallic pathways to structure.
  • EGC is pretty much flatline across all loads -- remaining at local earth potential.
 
  #12  
Old 04-04-06, 10:50 AM
Member
Join Date: Jan 2004
Location: Oregon
Posts: 1,219
>> I _believe_ that for this long feeder to a detached structure,
>> bonding ground and neutral at the detached structure will be
>> 'better' than keeping ground and neutral separate as for a
>> normal subpanel.
>Do you have any evidence that we could review from other
>installations to support your belief?

I do not have evidence in a format suitable for discussion. I can point you to basic physics, the Soares' text on grounding and bonding, and the common 'multi-ground-neutral' distribution system as the basis for my belief, but I will not write out a thesis on the topic.

In the case of two different but code compliant installations, the original poster can choose to use your opinion, my opinion, both of which are not significantly supported by the discussion so far, or can choose to hire an engineer, or could flip a coin.

>> To firmly establish this one way or the other would require a
>> detailed engineering analysis which would take into account
>> the preferences and requirements of the end user,
>Such as?

If the end use requires a communications system that introduces a parallel path, then the NEC would mandate a 4 wire feeder. If the end user for some reason wants to minimize the potential difference between the supply conductors and the surrounding earth, then a 3 wire feeder would better meet that goal.


>> local environmental conditions,
>Such as?

Local earth electrical currents, perhaps caused by natural processes, perhaps cased by power distribution systems.

>> and available engineering materials.
>What does this mean?

While you can happily hypothesize the use of a 4 wire feeder where the EGC is the same size or even larger than the neutral, a good engineering design will take into account the materials that happen to be easily available on the shelf. It may be that mobile homes are very common in the original poster's area, and that cables suitable for 4 wire feeders are less expensive than cables suitable for 3 wire feeders, or that the only 4 wire cables available have an undersized EGC for this application.

>> Either approach is accepted as being compliant with the NEC
>> if conditions regarding bonded metallic paths are met.
>But only four-wire always complies.


>> Factors to consider would include: cost of the EGC,
>Of course I already cited the tiny additional cost.

320 feet of direct bury 2/0 Al? $100 may not be significant to you.

>> transient potential differences between neutral and 'local ground'
>How is this a factor?
Or more to the point, how can this factor ever weigh in the favor of three-wires?

With a three wire feeder, the neutral conductor is directly bonded to the local ground and local bonded metal. With a four wire feeder the neutral conductor is bonded to local ground and local bonded metal via a very long path. This long path has higher impedance.

>> I certainly agree that an EGC that was the same size as the
>> neutral will always perform as well as the neutral itself as an
>> effective ground fault current path.
>Then this is no reason not to use four-wires.

An effective ground fault current path is not the only reason for grounding. If an effective ground fault current path was the only quality metric, then you would not need grounding electrodes, and a simple _bonded_ system would suffice.

If effective ground fault current path is the only metric that you are using for system quality, then a four wire system with full sized EGC is of equivalent quality to a three wire system.

>> I am comparing the 640 foot round trip path (neutral in
>> subpanel to main bonding jumper and then back to EGC in
>> subpanel) with a direct ground to neutral bond. 640 feet of
>> conductor versus 0 feet of conductor.

>There is no 640' path for neutral current.
>Why would neutral current need to return on the ECG?

Of course not. I never said that there was a 640' path for neutral current. I said that there was impedance between ground and neutral at the detached structure, and that impedance is caused by the 640' path from ground to neutral.


>> In the absence of any loads, the round trip path for the
>> ungrounded conductor is the same in either case, but any
>> loads, especially motor loads, will act to stabilize the
>> voltage between these two conductors.
Between which two?

My error. The voltage between the ungrounded conductor and the grounded conductor (and thereby to the grounded/bonded metal).



>> I would further note that in your ground fault current
>> calculation, Soares suggests using 8 times the OCPD trip
>> rating in order to place the device into is 'instantaneous trip'
>> region.
>> The #4 Al EGC suggested by simply upsizing the standard #6
>> EGC in proportion to the ungrounded conductors will provide a
>> bolted fault current of about 560A, with doesn't meet Soares'
>> suggestion, but probably provides an 'effective ground fault
>> current path'.

>A #1 Al EGC will handle 800A. If you want more, just use an
>ECG that is larger.

>We agree that the ECG can always be made large enough to
>perform as well as the neutral, so this is not a reason to use
>only a three-wire installation.

If you wish, you can make all the conductors as large as you wish; that is only a bit of money. The minimum EGC per the NEC is a #4. Spend more money, make it larger. Make it large enough and it will perform as well as the 2/0 neutral conductor.

I have modified the benefits list that bolide started:

Advantages of three-wire:
  • Lower cost.
  • Better matching of neutral potential to local earth potential
  • Better protection against externally caused transient overvoltages.

Advantages of four-wire:
  • Can have other metallic _bonded_ pathways between structures.
  • EGC is pretty much flatline across all loads -- remaining at local earth potential.
  • EGC only carries current in fault conditions.
  • Does not introduce new earth currents.
[/QUOTE]

-Jon
 
  #13  
Old 04-04-06, 12:08 PM
bolide's Avatar
Member
Join Date: Jan 2006
Location: PA
Posts: 1,909
>>> I _believe_ that for this long feeder to a detached structure,
>>> bonding ground and neutral at the detached structure will be 'better'
>I do not have evidence in a format suitable for discussion.

I don't have any to support your belief at all.

> I can point you to basic physics,
I am quite familiar with physics, DC and AC electronics.

> the Soares' text on grounding and bonding,
I read it.

> and the common 'multi-ground-neutral' distribution system
A horrible system.

> If the end use requires a communications system that
> introduces a parallel path, then the NEC would mandate
> a 4 wire feeder.

Correct. That is against three-wire.

> If the end user for some reason wants to minimize the
> potential difference between the supply conductors and
> the surrounding earth, then a 3 wire feeder would better
> meet that goal.

Give a valid reason for this goal.


> Local earth electrical currents, perhaps caused by natural
> processes, perhaps cased by power distribution systems.

And how does this favor three-wires?


> While you can happily hypothesize the use of a 4 wire feeder
> where the EGC is the same size or even larger than the
> neutral, a good engineering design will take into account
> the materials that happen to be easily available on the shelf.

How does this favor three-wires?


> It may be that mobile homes are very common in the original
> poster's area, and that cables suitable for 4 wire feeders are
> less expensive than cables suitable for 3 wire feeders

This is pretty common.
How does this favor three-wires?


> the only 4 wire cables available have an undersized EGC for
> this application.

"Available"?


>> Either approach is accepted as being compliant with the NEC
>> if conditions regarding bonded metallic paths are met.
>But only four-wire always complies.


> 320 feet of direct bury 2/0 Al? $100 may not be significant to you.

Compared to the future cost of being denied other utilities, it is trivial.



>>> transient potential differences between neutral and 'local ground'
>> How is this a factor?
>> Or more to the point, how can this factor ever weigh in
>> the favor of
three-wires?
>
>With a three wire feeder, the neutral conductor is directly
> bonded to the local ground and local bonded metal.
> With a four wire feeder the neutral conductor is bonded
> to local ground and local bonded metal via a very long path.

We all know this.


> This long path has higher impedance.

We all know this.

How can this factor ever weigh in the favor of three-wires?


>>> I certainly agree that an EGC that was the same size as the
>>> neutral will always perform as well as the neutral itself as an
>>> effective ground fault current path.
>>Then this is no reason not to use four-wires.
>
>An effective ground fault current path is not the only reason
> for grounding. If an effective ground fault current path
> was the only quality metric, then you would not need
> grounding electrodes, and a simple _bonded_ system would
> suffice.

That is false.

> If effective ground fault current path is the only metric that
> you are using for system quality, then a four wire system
> with full sized EGC is of equivalent quality to a three wire
> system.

Yes, it is equivalent or better. It is never worse.

>>> I am comparing the 640 foot round trip path (neutral in
>>> subpanel to main bonding jumper and then back to EGC in
>>> subpanel) with a direct ground to neutral bond. 640 feet of
>>> conductor versus 0 feet of conductor.
>>
>>There is no 640' path for neutral current.
>>Why would neutral current need to return on the ECG?
>
>Of course not.
> I never said that there was a 640' path for neutral current.

You did: " I am comparing the 640 foot round trip path
(neutral in subpanel to main bonding jumper and then
back to EGC in subpanel)".

> I said that there was impedance between ground and
> neutral at the detached structure, and that impedance
> is caused by the 640' path from ground to neutral.

Great. Tea is up ten cents a ton in China.
Please state the actual relevance.



>>> In the absence of any loads, the round trip path for the
>>> ungrounded conductor is the same in either case, but any
>>> loads, especially motor loads, will act to stabilize the
>>> voltage between these two conductors.
>> Between which two?
> The voltage between the ungrounded conductor and the
> grounded conductor (and thereby to the grounded/bonded
> metal).

For all anyone cares, the neutral is just another ungrounded conductor at the remote building.

This has no relevance whatsoever to safety or performance.


>>A #1 Al EGC will handle 800A. If you want more, just use an
>>ECG that is larger.

> If you wish, you can make all the conductors as large as you
> wish; that is only a bit of money. The minimum EGC per the
> NEC is a #4. Spend more money, make it larger.
> Make it large enough and it will perform as well as the 2/0
> neutral conductor.

Isn't that what I said?

Advantages of three-wire:
  • Lower cost.
  • Better matching of neutral potential to local earth potential
    (a meaningless benefit)
  • Better protection against externally caused transient overvoltages (not supported by any evidence; if anything, this is truer for four-wire systems).

Advantages of four-wire:
  • Can have other metallic pathways between structures.
  • EGC is pretty much flatline across all loads -- remaining at local earth potential.
  • EGC only carries current in fault conditions.
  • Does not introduce new earth currents.
 
  #14  
Old 04-04-06, 08:39 PM
Member
Thread Starter
Join Date: Apr 2006
Posts: 13
Thanks again for all the info. I have desided to run a ground. I am using insulated #1 alu (which may be overkill) cause I had some around.
Should I still drive a ground rod and/or tie to the rebar at the workshop?
Also, I need to do a splice in the ground run, Is ther a direct burial wire nut that will work for that big of wire? If not what are my other options?
Jon
 
  #15  
Old 04-04-06, 10:25 PM
bolide's Avatar
Member
Join Date: Jan 2006
Location: PA
Posts: 1,909
> Should I still drive a ground rod and/or tie to the rebar
> at the workshop?

Yes, absolutely to to the rebar.
If it is not a concrete-encased electrode per NEC 250.52(3), add rods, plates, or rings as needed.


> I need to do a splice in the ground run.
> Is there a direct burial wire nut that will work for that big a wire?

There are direct-burial splice kits that are listed for direct-burial use per NEC 110.14(B). No box is required.
 
  #16  
Old 04-04-06, 10:45 PM
bolide's Avatar
Member
Join Date: Jan 2006
Location: PA
Posts: 1,909
Are you using conduit or direct burial?

If you are using direct burial, you can't use some insulated 1/0 Al conductor, say, RHW. You have to use direct burial cable like USE or XLPE.

Soil conditions might also dictate that you use conduit.
 
  #17  
Old 04-05-06, 08:42 AM
Member
Thread Starter
Join Date: Apr 2006
Posts: 13
I am using direct burial (USE), which is common in this area. I am also using conduit where it runs through my driveway and parking spots.
 
  #18  
Old 04-09-06, 11:59 PM
bolide's Avatar
Member
Join Date: Jan 2006
Location: PA
Posts: 1,909
Originally Posted by winnie
I believe that treating this feeder as a service is a better approach if possible.
I have never seen such a claim made elsewhere. For if I had, I would have challenged anyone making such a claim to provide actual calculations showing how this is possible.

Grounding vs Bonding, part 5:
Originally Posted by Mike Holt, NEC Consultant
The preferred practice is to not use the grounded conductor (neutral) as the effective ground-fault current path. Instead, you should install an equipment grounding (bonding) conductor with the feeder conductors to the building or structure in accordance with 250.32(B)(1).
Note: not even a hint that the neutral could ever be a better fault current path nor that a three-wire feeder is ever to be preferred over a four-wire feeder.

So if anyone knows of an actual reason that three-wire feeders are a better approach for any 240/120V system, other than cost, I am still waiting to hear it.
 
  #19  
Old 04-10-06, 12:01 AM
bolide's Avatar
Member
Join Date: Jan 2006
Location: PA
Posts: 1,909
Originally Posted by colorado hick
I am using direct burial (USE)
Okay. May we assume that it is not visible inside your house either?
 
  #20  
Old 04-10-06, 07:39 AM
Member
Thread Starter
Join Date: Apr 2006
Posts: 13
It is not visable inside the house. I have an outside panel and the feeder (with the ground) runs through the ground, then pops up to the panel inside some schedule 80 PVC.
Looking at the NEC, it seems that I am not required to run a seperate ground at the workshop since I am tieing into the ground at the main panel, but if understand correctly it would be a good idea?'
 
  #21  
Old 04-10-06, 08:21 AM
Member
Join Date: Jan 2004
Location: Oregon
Posts: 1,219
This is the point that bolide and I have been going back and forth on. Re-read the above if you want the details.

For a _detached_ structure, _if_ you have no other bonded metal paths between the two structures (eg. metal water pipes, grounded coax, etc.) then you are _permitted_ to either use a three wire feed where the neutral gets grounded at the subpanel in the detached structure. Or you may run a separate equipment ground conductor and have an isolated neutral and ground bus in the subpanel. The NEC permits either approach in this case.

If you have any metal paths that could act to carry neutral current, then you are _required_ to run a separate equipment ground conductor and have an isolated neutral and ground bus in the subpanel.

I have a hunch that for this long feed, using the first method would be better if possible, bolide is quite certain that running the separate egc is preferable.

We've listed the reasons for and against each up above.

The only significant difference between my understanding and bolide's understanding, if I am reading him correctly, is the issue of impedance between the ground and the neutral at the detached structure. He appears to agree that this impedance exists, but states that this impedance has no significant negative impact on the installation. My _hunch_ is that this impedance will have a significant impact on how the system responds to external upsets, for example a nearby lightning strike. That is my hunch and I'm sticking to it; so far I have not presented significant evidence that this is a significant issue; similarly bolide has not presented any evidence that this is _not_ an issue; he has simply stated that he cannot see it being an issue. I will not be spending significant time researching this point; I have my real job to do.

We also have minor differences in our assessments of the costs of having the EGC and if a reduced size EGC versus a full size EGC is signficant.

I would not spend too much time worrying about this detail; there are probably far more significant issues related to the installation that you should be researching.

-Jon
 
  #22  
Old 04-10-06, 09:26 AM
bolide's Avatar
Member
Join Date: Jan 2006
Location: PA
Posts: 1,909
Originally Posted by colorado hick
Looking at the NEC, it seems that I am not required to run a separate ground at the workshop
Correct.


> since I am tying into the ground at the main panel,

You should not do this.


> but if understand correctly it would be a good idea?'

It is a great idea.
You still must have a local earth ground (whatever is present at the building, or ground rods). The local earth ground handles the issue with lightning.

The grounding article in EC&M which I quoted earlier makes it clear that four-wires is the preferred practice. (Jon's "hunch" is devoid of scientific basis and should be disregarded. The neutral conductor responds to external upsets no differently than the ungrounded conductors and clearly he does not advocate grounding the latter at the remote building. Bonding the neutral to the local ground would simply change the potential between the grounded and ungrounded conductors to something other than 120V. This is not a net improvement compared to the alternative. In fact, it is just as easily argued that not having them bonded is more stable because all three circuit conductors track only the same frame of reference instead of having one of them tracking a local earth reference. That's like riding in one car but having your arm tied a car that is following 320' behind you.)

Of course, much of your equipment could be 240V with no neutral and thus completely unaffected either way.
 
  #23  
Old 04-10-06, 06:51 PM
Join Date: Mar 2006
Location: Fort Collins, Colorado
Posts: 159
Jon, you're absolutely right about the benefit of a three-wire feeder in this installation. A NEC-compliant EGC would not open an OCPD as quickly as a full-sized neutral would, and I will take that with you to the bank. Thus, it's expense does not justify it's use in this circumstance.

The use of a three-wire feeder does not "incur expense in the form of denial of utilities." That is supremely deceptive. If a normal Cat-5 is run out to the outbuilding from the originating building, then it will not have a metallic sheath to bond, so therefore it is not a parallel neutral path. If a new independent phone service is provided to the outbuilding (which is uncommon but not unheard of) then this will be no different than the millions of phone services in use today. And there would not be any "continuous metallic paths bonded to the grounding system in each building or structure involved..." aside from through utility equipment (outside the grasp of the NEC) that has this connection made daily anyway.

I've heard a rumor that non-metallic pipe suitable for water use is for sale, so they won't be denied running water out to the outbuilding. They could even run a substantial amount of copper, and switch to plastic before creating a grounding electrode, if they so desire.

Lower impedance results in more current flow, which will open an OCPD quicker than a smaller conductor with higher impedance.

Yes, you have every right to run an oversized EGC if you have money to burn.

The four-wire system is only superior in three ways:
  • An open neutral between structures will not result in energization of the normally non-current-carrying metallic electrical equipment in the remote structure.
  • If this neutral opens, the effective ground fault current path is not compromised.
  • Future bonding will not present a parallel path for neutral.

Given these three points, Mike Holt is fully justified in advising a EGC be run with the conductors. But we are fully justified in exercising our code-compliant, cost-effective, and under all but certain conditions, electrically superior (in terms of OCPD-opening) option.

Edit to add: As for the (remote/local) (earth/neutral) "hunch" comments, I see no real relevance to the thread. Transients and surges happen all the time, and the connection of a neutral to earth at a service is a marginally effective activity at best, IMO. A TVSS would work equally well/poorly in either scenario, IMO.
 

Last edited by Rocky Mountain; 04-10-06 at 07:20 PM. Reason: fix list, and then to remove 'hunch' comment. I was referring to a different hunch
  #24  
Old 04-10-06, 07:52 PM
Member
Join Date: Jan 2004
Location: Oregon
Posts: 1,219
Rocky Mountain:

You raise a good point that the impedance of the neutral will likely be lower than the impedance of an EGC sized to NEC minimums. However, as bolide mentioned, one could choose to increase the size of the EGC to match the size of the neutral. I agree with bolide that for 'ground fault' currents a separate EGC equal impedance to the neutral if made full sized. But this was not the focus of my 'hunch'.

The point that bolide and I are debating is the benefit or lack thereof for the coupling between the building electrical system and 'local' earth. My hunch is that if there is a lightning strike or other similar transient, that it is better for the entire building electrical system to 'ride' with local earth.

If you have a separate EGC, then the current carrying conductors will 'ride' with the earth potential back at the main service, while all the bonded metal and EGC conductors will be bouncing around at the local earth potential. Of course the EGC in the feeder will act to equalize the potential between the two locations, but for lightning induced surges there will be quite a large voltage transient.

This needs to be traded off against issues such as introducing earth current, and parallel path issues if there is any other metallic pathway between the two structures. My hunch is that four wire feeders are generally better, but that at some long and not well defined distance the balance of benefits will make a three wire feeder 'better'.

-Jon
 
  #25  
Old 04-10-06, 08:50 PM
bolide's Avatar
Member
Join Date: Jan 2006
Location: PA
Posts: 1,909
> A NEC-compliant EGC would not open an OCPD as quickly as a full-sized neutral would.

The NEC specifies the minimums. Larger conductors are pretty much always NEC-compliant too, you know. Certainly an unreduced EGC conductor is.

The NEC doesn't cover every possibility.

Obviously you would have the same complaint if the run were a mere 3.2' -- a #6 EGC isn't as good as a #4 neutral. That's true. It's not as good.


> Thus, its expense does not justify its use in this circumstance.

I don't see how you put prices into your cost-benefit analysis. Could you share those?


Four-wire:
> Future bonding will not present a parallel path for neutral.

Correct.


Three-wire:
> under all but certain conditions, electrically superior (in terms of OCPD-opening) option.

Voltage drop must be considered in calculating the EGC size. A #1 Al EGC will do the job for 800A at 320' - meaning that it will do just as good a job as the 2/0 neutral in terms of magnetic trip on the OCPD.

This is no different from using the EGC on any other feeder.

Doubling the other conductors from #2 to 2/0 requires doubling the EGC from #4 (since you wouldn't be satisfied with the performance of #6) to #1.


> Transients and surges happen all the time, and the connection of a neutral to earth
> at a service is a marginally effective activity at best, IMO.

I think you are correct. It think its effectiveness falls off with the square of the impedance to the poco line.
So it is somewhat effective if you are close to the grid, but 320' away with just a 1/0 Al wire or whatever, it sure can't dissipate a lightning strike.

> A TVSS would work equally well/poorly in either scenario, IMO.
Again, I think you are correct. Small difference, but no change to efficacy (or lack thereof).
 
  #26  
Old 04-16-06, 03:50 PM
bolide's Avatar
Member
Join Date: Jan 2006
Location: PA
Posts: 1,909
Originally Posted by winnie
You raise a good point that the impedance of the neutral will likely be lower than the impedance of an EGC sized to NEC minimums.
The minimum from NEC 250.4(A)(5)?



> The point that bolide and I are debating

I apologize for the late reply. Apparently I after reading RM's response, I replied and failed to notice that you also had responded.


> is the benefit or lack thereof for the coupling between the building
> electrical system and 'local' earth.

Okay.

> My hunch is that if there is a lightning strike or other similar transient,
> that it is better for the entire building electrical system to 'ride' with local earth.

Explain "ride". Only the EGC rides.
If you bond the EGC and neutral, you still have two ungrounded conductors that are not "riding".

Furthermore, by definition, anything using a neutral will now see something other than 120V.

(See analogy about riding in a car with one arm tied to a car 320' away.
Basically equipment will have an ungrounded conductor referenced to a location 320' away with a grounded conductor referenced to its location. Having both conductors remotely referenced at least assures that they move together. All equipment cares about is the voltage between the conductors. If, with respect to local earth, one is 180V and the other is 300V, nothing bad happens.)


> If you have a separate EGC, then the current carrying conductors
> will 'ride' with the earth potential back at the main service,

This is good. However, a local electrical field will raise or lower all three equally.

So L1 to neutral is still 120V, L2 to neutral is still 120V, and L1 to L2 is still 240V.

> while all the bonded metal and EGC conductors will be bouncing around at the local earth potential.

Correct and unavoidable.


> Of course the EGC in the feeder will act to equalize the potential between
> the two locations, but for lightning induced surges there will be quite
> a large voltage transient.

So if you tie EGC to neutral locally,
L1 to neutral is not 120V, L2 to neutral is not 120V, and L1 to L2 is still 240V.

How is this any improvement whatsoever?


> This needs to be traded off against issues

Trade off?

What is there to trade? You need something of value to trade.

You didn't give a single advantage of the local earth ground tied to neutral.

Having the neutral bouncing all over the place relative to the ungrounded conductors sounds like a disadvantage to me.
I won't trade anything to have that!


> such as introducing earth current, and parallel path issues
> if there is any other metallic pathway between the two structures.

More disadvantages of three-wire versus four.


> My hunch is that four wire feeders are generally better,

Still waiting for an example where they are not.


> but that at some long and not well defined distance the balance
> of benefits will make a three wire feeder 'better'.

There is no evidence to support this.
Even at ten miles away there is no advantage.

It makes far more sense to have something that absorbs excess energy locally rather than to dump it onto the neutral where it will follow all paths including to the ungrounded conductors.


If lightning has to blow out some lamps or burn up some motors on its way from neutral to ungrounded, it will.

Remember, lightning sees ungrounded conductors and grounded conductors all as paths to be followed. So you have not explained how you are protecting your equipment by tying the neutral to the EGC.

Why not treat the neutral as any other ungrounded conductor?
How is it different?
 
Reply

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Thread Tools
Search this Thread
Display Modes
'