Multiwire circuit for a garage?


  #1  
Old 07-13-05, 01:54 PM
G
Member
Thread Starter
Join Date: Jul 2005
Posts: 38
Upvotes: 0
Received 0 Upvotes on 0 Posts
Multiwire circuit for a garage?

Ok, I'll try this again--my first attempt ended in cyber oblivion so I'll try to be brief in case it happens again.

My mom has a two-car garage with only one duplex outlet for the laundry and one fluorescent light. There is a separate 3 wire 240v 20Amp (#12 wire) circuit that had been connected to a single 240V 50Amp electric range style plug which no longer serves any purpose and has been removed. It would be impractical at this time to run another cable back to the service panel, so it seems like a multiwire circuit would work nicely. Ideally I would like to install four duplex outlets (two on each leg) and two fluorescent light fixtures (one on each leg, controlled by a single double pole switch). My mom has also requested an outlet situated for the possible future addition of a garage door opener, but I am not sure I would want to put this on the same circuit.

My concern is that I have read that some devices, like fluorescent lights cause a phase lag in the current which can, under unusual circumstances, overload the grounded neutral wire in the circuit. (Hence my plan to run one light on each leg to keep the circuit balanced). I am hoping that someone with some experience in this area could tell me what types of loads might cause a problem and how concerned I need to be about this possibility. Is this plan safe and workable?

Thanks
 
  #2  
Old 07-13-05, 02:18 PM
I
Forum Topic Moderator
Join Date: Feb 2005
Location: Near Lansing, Michigan
Posts: 9,785
Upvotes: 0
Received 45 Upvotes on 43 Posts
Sounds like a good plan to me. I wouldn't be worried about phase lag with only one ballasted fixture; go ahead and use this circuit for the garage door opener too. Issues like that are far more relevant with commercial three phase power.

Another thing to keep in mind is that any general-purpose receptacle in the garage must be GFCI protected. Due to the nature of GFCI, you will need one GFCI receptacle for each leg of the MWC. The neutral cannot be shared after the GFCI.

If both legs are connected to the same "yoke" (i.e. switch, receptacle) the breaker protecting the circuit must be handle tied. A double pole breaker is the best way to do this; 20A breaker for the #12. The old outlet was likely for a welder where oversized breakers are allowed. If you wanted to spend the money, put in a double pole 20A GFCI breaker for this MWC and you're covered in most any situation.

One final note, you cannot use the "feed-through" capability of receptacles for the neutral of a MWC. The neutral must be pigtailed with wirenuts.
 
  #3  
Old 07-13-05, 05:45 PM
W
Member
Join Date: Jan 2004
Location: Oregon
Posts: 1,104
Upvotes: 0
Received 0 Upvotes on 0 Posts
Perhaps a silly question, but is the #12 3 wire that you have in place tree conductors plus _ground_, or just three conductors. Is it in conduit or armor?

Is this a detached garage?

-Jon
 
  #4  
Old 07-14-05, 10:37 PM
G
Member
Thread Starter
Join Date: Jul 2005
Posts: 38
Upvotes: 0
Received 0 Upvotes on 0 Posts
Originally Posted by ibpooks
Sounds like a good plan to me. I wouldn't be worried about phase lag with only one ballasted fixture; go ahead and use this circuit for the garage door opener too. Issues like that are far more relevant with commercial three phase power.

Another thing to keep in mind is that any general-purpose receptacle in the garage must be GFCI protected. Due to the nature of GFCI, you will need one GFCI receptacle for each leg of the MWC. The neutral cannot be shared after the GFCI.

If both legs are connected to the same "yoke" (i.e. switch, receptacle) the breaker protecting the circuit must be handle tied. A double pole breaker is the best way to do this; 20A breaker for the #12. The old outlet was likely for a welder where oversized breakers are allowed. If you wanted to spend the money, put in a double pole 20A GFCI breaker for this MWC and you're covered in most any situation.

One final note, you cannot use the "feed-through" capability of receptacles for the neutral of a MWC. The neutral must be pigtailed with wirenuts.


I didn't know general purpose outlets in a garage need to be GFCI, but this brings up an interesting question. Once before, I helped wire a multiwire circuit with two GFCI's, one on each leg, and one of the GFCI's kept tripping. We couldn't find anything wrong with either outlet, or the wiring (and the neutrals from GFCI's were not shared downstream). More recently I happened to read that in this situation you can only protect one of the legs of the multiwire circuit with a GFCI, but there was no explantation why.
 
  #5  
Old 07-14-05, 10:54 PM
G
Member
Thread Starter
Join Date: Jul 2005
Posts: 38
Upvotes: 0
Received 0 Upvotes on 0 Posts
Originally Posted by winnie
Perhaps a silly question, but is the #12 3 wire that you have in place tree conductors plus _ground_, or just three conductors. Is it in conduit or armor?

Is this a detached garage?

-Jon

Fortuneately, the 12/3 cable was with ground, and the ground tested ok. The original wiring was not in conduit, just Romex that was dangling across a span. I have added a junction box where it enters the garage and will probably run 1/2" EMT around the perimiter and down to the outlets--just seems like a nicer, safer installation. The garage is not detached.
 
  #6  
Old 07-15-05, 05:30 AM
W
Member
Join Date: Jan 2004
Location: Oregon
Posts: 1,104
Upvotes: 0
Received 0 Upvotes on 0 Posts
When you had that multiwire circuit with GFCI's, did you have the neutral feeding one GFCI on the 'LOAD' side of the other GFCI? This would cause the first GFCI to trip.

Everything you want to know about GFCIs: http://www.ecmweb.com/mag/electric_think_gfci/

-Jon
 
  #7  
Old 07-18-05, 12:57 AM
G
Member
Thread Starter
Join Date: Jul 2005
Posts: 38
Upvotes: 0
Received 0 Upvotes on 0 Posts
Jon,

Please excuse the delay in responding, I do not always have the opportunity to check my email on a regular basis. As I best remember it, we only had one GFCI duplex on each leg of the 240V multiwire circuit, so there were no load connections. We split the neutral wire into two branches upstream of the GFCI's, and ran one branch to the line side of each GFCI. To be honest, I only ran the wires to the outlet boxes, and my buddy connected the GFCI's later. He told me one of them kept tripping, so he replaced it with a regular 20A duplex (since neither outlet was actually required to be a GFCI), and checked its operation with an outlet tester. He swore he had made all the connections correctly, but since everything was already wired and working when I got back, I never checked it out for myself.

Thanks for the link to the info on GFCI's. I've already read and downloaded it.

Gregor
 
  #8  
Old 07-19-05, 08:01 PM
G
Member
Thread Starter
Join Date: Jul 2005
Posts: 38
Upvotes: 0
Received 0 Upvotes on 0 Posts
Hey, I was thumbing through my 2002 edition of Practical Wiring and I found the passage about GFCI's and multiwire circuits. It states: "An ordinary GFCI will not protect a three-wire circuit, or receptacles connected to two of the wires of a three-wire circuit...But if a three-wire circuit is divided into two 2-wire circuits, a GFCI can be used to protect one of those two-wire circuits..." (page 146). On second reading, it doesn't specifically say you can only protect one two-wire circuit. Guess I'll try two separate GFCI's in the garage and see what happens.

I feel more comfortable with the idea of using a multiwire circuit, after having read some of the replies here. I still would like to know if there is any ordinary equipment or combinations of equipment that I should be careful about using on this circuit to avoid overloading the neutral wire.
 
  #9  
Old 07-20-05, 08:14 AM
I
Forum Topic Moderator
Join Date: Feb 2005
Location: Near Lansing, Michigan
Posts: 9,785
Upvotes: 0
Received 45 Upvotes on 43 Posts
Originally Posted by Gregor
On second reading, it doesn't specifically say you can only protect one two-wire circuit. Guess I'll try two separate GFCI's in the garage and see what happens.
Two seperate GFCI receptacles is the way to go. The only other option is to install a double pole GFCI breaker to suppy the circuit.

I feel more comfortable with the idea of using a multiwire circuit, after having read some of the replies here. I still would like to know if there is any ordinary equipment or combinations of equipment that I should be careful about using on this circuit to avoid overloading the neutral wire.
Nope, the physics of a multiwire circuit dictate that the current in the neutral wire is the difference of the current in the two hots. For example, you're drawing 13 amps on the black leg and 8 amps on the red leg, the neutral wire is only carrying 5 amps. Therefore, the hots would be overloaded long before the neutral; that would cause the breaker to trip.

The worst thing that can happen on a multiwire circuit is a neutral coming loose. In that case your loads will get 240V instead of 120V. Make sure your neutral connections are good and tight, and there will be no problems. Good luck.
 
  #10  
Old 07-20-05, 08:42 AM
J
Member
Join Date: Sep 2000
Location: United States
Posts: 17,733
Upvotes: 0
Received 1 Upvote on 1 Post
ibpooks, I think you missed Gregor's concern. Without any phase shift in the branch circuit, there is no way to overload the neutral, and the neutral current is indeed the difference of the current in the two hot wires. But Gregor is asking about loads that might create a voltage/current phase shift. I don't know the answer to his question, but I doubt that it's a significant problem. I am interested in hearing others' comments on this very technical question.
 
  #11  
Old 07-20-05, 12:21 PM
I
Forum Topic Moderator
Join Date: Feb 2005
Location: Near Lansing, Michigan
Posts: 9,785
Upvotes: 0
Received 45 Upvotes on 43 Posts
Ahh yes, the phase lag thread.

Without putting on my calculus hat, my impression is that phase lag could only be a problem in a multiwire circuit if a perfect storm of loads were applied.

You would need a high power, low power factor, inductive (lagging) load on one leg, and a high power, low power factor, capacitive (leading) load on the other leg. You would then start to see a percentage of the shared neutral current waveforms be additive. The mismatches of lagging and leading power factors would have to be really huge and the apparent power would have to be near the the circuit capacity in my opinion.

You would need loads with power factors substantially off unity, which would be quite inefficient anyway. Perhaps if you had 20 of the cheapest flourescant lights you could find on one leg and 20 of the cheapest computer power supplies you could find on the other leg, you may see some overcurrent in the neutral conductor.
 
  #12  
Old 07-20-05, 11:05 PM
G
Member
Thread Starter
Join Date: Jul 2005
Posts: 38
Upvotes: 0
Received 0 Upvotes on 0 Posts
More Multiwire Madness?

Originally Posted by ibpooks
Ahh yes, the phase lag thread.

Without putting on my calculus hat, my impression is that phase lag could only be a problem in a multiwire circuit if a perfect storm of loads were applied.

You would need a high power, low power factor, inductive (lagging) load on one leg, and a high power, low power factor, capacitive (leading) load on the other leg. You would then start to see a percentage of the shared neutral current waveforms be additive. The mismatches of lagging and leading power factors would have to be really huge and the apparent power would have to be near the the circuit capacity in my opinion.

You would need loads with power factors substantially off unity, which would be quite inefficient anyway. Perhaps if you had 20 of the cheapest flourescant lights you could find on one leg and 20 of the cheapest computer power supplies you could find on the other leg, you may see some overcurrent in the neutral conductor.

Thanks, this is what I wanted to know.

While we're on the subject of multi-wire circuits, I have a question that I have been trying to get answered for a couple of years. I do part time maintenance sub work at a school camp that has 120v/208v three phase service. Most of the buildings only have three-wire service--two hots and a neutral. We were having trouble in the main staff office/residence building. It is almost 50 years old and some of the electrical circuits, which were run in conduit through the brick walls, were now being overloaded. When we started tracing the branch circuits back to the panel, we found several multi-wire circuits, some that were fed from two breakers on the same "leg" and one on the other "leg." After much deliberating and diagramming, we rearranged the branch circuits to better distribute the electrical load. We also arranged it so that each multi-wire circuit was fed from only one hot on each "leg." This seemed safer than having the possibility of two hots from the same "leg" feeding one neutral. But I have always wondered how well the currents actually cancel in this 120V/208V system considering that the two hot phases are, I believe, only 120 degrees apart with respect to the neutral wire. Is it legit to use multi-wire circuits with this type of service?
 
  #13  
Old 07-21-05, 05:57 AM
W
Member
Join Date: Jan 2004
Location: Oregon
Posts: 1,104
Upvotes: 0
Received 0 Upvotes on 0 Posts
Multi-wire branch circuits are legitimate in three phase circuits, however they are different beasties and need to be treated differently.

The situation that you describe, with two 'hots' from one phase and one 'hot' from another phase is explicitly not correct, a code violation, and a fire hazard. But you figured that out already

The first thing to know is that the current in the neutral will balance to zero only if you have _three_ hots from all _three_ legs running at exactly the same current. If you happen to have _two_ legs from a three phase panel, used as a multi-wire branch circuit, then the current flowing in the neutral is _roughly_ equal to the current flowing in the more heavily loaded 'hot' conductor. (Depending upon the current in both phases, the current in the neutral will vary between 87% and 100% of the current in the more heavily loaded 'hot' conductor.)

In this situation, the neutral won't overload, but must be counted as a current carrying conductor for derating purposes. So using two _different_ hots and one neutral is legitimate. Or using three _different_ hots from a _three_ phase panel is legitimate, and you only have to count three current carrying conductors.

You can do this analysis in a graphical fashion, by using vectors to represent both the magnitude of the current (how many amps) and the phase (time offset) of the current. From the origin you draw a vertical vector who's length is proportional to the current in phase A. Then from the tip of this vector you draw another vector at 120 degrees, who's length is proportional to the current in phase B. Finally, from the tip of this second vector you draw another vector at 240 degrees, with a length proportional to the current in phase C. The resultant distance from the origin is the current flowing in the neutral.

Finally, there is a separate neutral overload problem found in three phase systems. While it is true that the _fundamental_ currents will cancel out appropriately, _harmonic_ currents may not. In particular, if there happens to be any third harmonic current flowing to the branch circuits, this third harmonic will _not_ cancel in the neutral, but will instead add up. This will be a problem if there are a significant number of 'non-linear' loads on the branch circuits, which cause harmonic currents to flow. In particular, if you have many computer power supplies on a three phase multi-wire branch circuit, you can overload the neutral.

This is an issue that is not currently well addressed in the code. Various EU directives to deal with this problem are mandating power supplies that cause less harmonic current flow, however I am not up on the status of these rules.

-Jon
 
  #14  
Old 07-21-05, 08:32 PM
G
Member
Thread Starter
Join Date: Jul 2005
Posts: 38
Upvotes: 0
Received 0 Upvotes on 0 Posts
Thanks for the info, Jon. That ties together a lot of little bits I have come across on the subject.

Thanks to everyone who replied to this thread. You all are truly awesome. I am really glad I stumbled across this site.
 
 

Thread Tools
Search this Thread
 
Ask a Question
Question Title:
Description:
Your question will be posted in: