I generally dislike three way switches. The reason I have frequently stated in the past is the accidents that can occur while trying to flip a switch while carrying things in both hands. Last week the power went out (strange story) and it reminded me of another reason I dislike three way switches. Due to the power situation, I was going around the house unplugging critical things (computer first, happily running on UPS, so it was the UPS that got unplugged), and switching off lights.

The problem with three way switches is that I cannot tell from the switch whether the lights are actually on or off. The switches were not wired consistently.

It would be helpful if the switches were wired so that when all switches are in the down positions, the controlled circuit is down. My house (my mother had built) was definitely not wired very well (had it been properly inspected, there should have been a dozen or more red tags). So I have a project to rewire these three ways so there is at least a way to make sure the lights are off when there is no power.

I still don't like three way switches because I don't like any instance of a switch being reversed from up-is-on and down-is-off. I'd rather use those switches that make momentary contact in the up and down positions and then have a spring pressure return to the middle, controlling a contactor that controls the light circuits (most would be magnetically held on, but the few lights I want to normally stay on would be the latching type).

Our street was having the MV feeder rewired for some reason (the crews doing the work don't know why, as the new cable is the same capacity as the old). It's all underground. Apparently they switched one of the pad transformers down the street to the new cable. But there was a problem with the new connection and it was arcing. They didn't discover it until after the crew that connected the new cable hot had left for the day. So the crew that was still on site (not authorized or trained or geared up to work MV hot, I guess) had to pull the fuse on the whole street to remake the end of the cable going into the transformer, since it was now damaged by the arcing. Once I heard part of this story while the power was still out, I decided to play it safe and unplug stuff at home, even though this work was not on the transformer feeding me (that connection may come in the next week or two). Given mistakes this crew has made already on this project (slicing through a water main and damaging a gas line), I decided that disconnecting would be better, in case they managed to cause an underground equivalent of the Harford Event (video 1 - video 2).

You have put your thumb on the issue. When 3 ways are wired it is a good idea to verify that off is all down. All you have to do is flip one of them to make this true if it isn't. The "flip" side of that is they will also be off with both switches up.
These represent a logical "exclusive or".
0+0=0
1+1=0
1+0=1
0+1=1

This gets a little more complicated when you put 4 way(s) in the middle but you can still set them up so all down is off. Dr Boole says an even combination of "up" switches is off and an odd combination is on.

It always amazes me that an electrician will make sure all the cover screws are aligned in the same direction but they don't bother to coordinate how the 3 ways are set up. A lot of them just seem happy that they work right in all 4 states.
There is no code issue here.

I'd say you have problems other than switch issues.

Sure, you can instal the switches so that 'down is off.' That will last right up to the first use. Do things 'out of sequence' just once, and all bets are off.

For example ... you enter the hallway, the lights are off, and all switches are down. You flip a switch up, and you have lights.
Now you walk the length of the hall, reach the other end ... and you flip that switch 'up' to turn the lights off.

They DO make illuminated 3-way switches, where the toggle is lit when the lights are off. They probably also make 'pilot' versions, which are lit when the lights are on.

Otherwise, there is a role for timers, motion sensors, and contactors. For example, overseas it is common for stairwell lights to operate for a timed interval from a push button at each level.

You can (should?) still have them oriented so with all switches down the light is off.

Greg ... in the example I just gave ... you can see that it's just not possible to do that!

Let's look at it once again:
You enter a dark stairwell, lights are off, all switch handles are 'down.' So far, so good.
You flip one switch up, and you have light. You proceed down the stairs.
As you exit the stairwell, you want to turn the lights off. This switch has remained 'down' from the start. To turn off the lights, you have to flip the switch 'up.'

On your return trip, you would eventually return things to a 'down is off' state. If you exited the area by another route, however, it's anyone's guess what position the switch will be in.
For example, the next person to repeat your exact trip would turn the lights on by flipping the first switch 'down,' then turn off the lights by flipping the second switch 'down.'

In short, the orientation of the switch has absolutely no connection to the way it was installed.

They do make 3 ways with indicator lights in togal and decora. Needs a load to work.

Myself I would try to put most switches in so off is all one way. On resi jobs I might even make sure the switches are left in a uniform position when the job is done. Otherwise some people think you put the switch in upside down. Proving the customer wrong is not always the best idea.

I think Greg is just talking about having them all the same orientation (i.e. both down = off, both up = off).

Anyway, what do three way switches have to do with having both hands full? What's the difference between a three way and a single pole switch under these circumstances? Yeah, you might have to try switching twice (if you expect it being down while it is actually up or vice versa), but I think that's acceptable. Spring operated switches or push buttons operating contactors just make the whole thing unnecessarily complex, adding potential error sources.
The whole stairway contactor thing only exists for power saving reasons - landlords don't trust their tenants to switch off the light after walking up or down the stairs.

I wish all I had to worry about was the orientation of the 3 ways in my house.
Install motion detectors

I like a 3 way switch that controls a light that isn't visible from the switch. You just can't tell without going for a walk.

I have eliminated a lot of this "3 way" confusion in my house with motion/occupancy detectors. You seldom turn on a light switch unless you want task lighting. The motion system makes sure you always have "walking around" light everywhere, inside and out. It is all fairly soft light. This started years ago when I figured out the girls were using the fridge for a night light. My first motion light was in the kitchen/dining area that also gets you from the front and back door. It went on from there. Now the light just follows you around the house.

Ya, I tied the occupancy sensor thing .
The library is just no fun anymore, unless you read fast .
My honey never shuts lights off anyway AAARRRGGGGG... so it doesn't realy matter at my house.

Sorry just, had to chime in!

I installed 4-way switching for a stairs with a middle landing. The homeowner told me later he'd run up & down the stairs to set the switches all in the down position. I still think he's certifiably insane for doing this. There is no proper orientation for 3-way or 4-way switching. Get a life.

Motion detectors would be a very good way to go. But I'd still want switches for extra lighting in many cases. Either way, some kind of mechanical or electrical switching of the 120V is taking place under control of something at 12V or less. If the motion detector logic can switch it, a momentary contact switch driving similar logic should be able to also do the job. There are lots of things around that can control 120V or more from 12V or less, such as all that X10 stuff (which I don't want for other reasons), and is presumably listed. So I shouldn't need to use those expensive industrial contactors.

When the power is on, and you want to be sure the lights are off, the light itself is the verification. However, when the power is off, and you have a reason to be sure the lights are off without a means to verify it, having all your switches in an expected configuration is appropriate.

Another example of when one might want to do this is when the power goes out while a lot of lights were on while at home, but you need to leave for whatever purpose before the power has come back on, and want to be sure the lights are off (and other things are desired to be on, so flipping the main breaker off is undesired).

I agree with the all-switches-down must be off. It's not hard to fix since only one 3-way needs to be reversed. You just have to do it when you can verify that it's not the way you want it to be. The only extra step beyond tagging a switch to be reversed is checking behind the plate in the case of 3 or more switches controlling a light to be sure you're going to work on a 3-way one instead of a 4-way one (more hassle). Then cut power and reverse the marked switches.

I have found that even if I am not flipping pages fast enough to satisfy the motion detector my lips moving will do it.

"Switches down must be off"

What a nice idea. Oddly enough, it hasn't been "part of code" for all that long ... maybe 20 years. Even then, the rule applies to DISCONNECT switches.

While an ordinary light switch may serve as a disconnecting means, 3-way and 4-way switches cannot. There are two reasons why they cannot qualify as disconnecting means.

The first is that it's simply too easy for you to turn things off at one switch ... while someone around the corner flips the other switch "on."

The second comes from violating a much older (maybe 40 yr. old) code rule: switching the neutral.
While no longer allowed by code, I still encounter 3-way switches that are arranged in a manner that switches the neutral. This method can present a situation where the light is 'off' only because both wires -the 'hot' as well as the 'neutral'- are hot.

Others have advocated motion sensors. Another nice idea, but limited. Fluorescent lighting ... so fondly thought of by the energy code wonks ... don't like motion sensor switching. The bulbs don't reach full brightness, the ballasts' electronics have issues with the sensors' electronics, etc.
Think of it as conflicting codes - or good intentions run amok.
Then there are unintended consequences; my cat trips every motion sensor on the street as he walks by. Just think how much fun that might be in your house at night.

I have CFLs on motion sensors in the garage and the only PAR I have, outside (the rest are softer light) and I don't see a problem. The triac should be driven to saturation, unlike some dimmers. The dog walking around does trip some of the motion sensors but they are arranged so they don't bother you in the bedroom. I masked the window of the one in the bathroom so it doesn't "see" the dog. It is under the vanity counter.
If I have the real lights on it disables the MD as does daylight.
The lights in the bathrooms are just short pieces of rope under the toe kicks and the hall lighting is rope in crown, about 4" below the ceiling for soft indirect light. In the living room, dining room and kitchen it is rope on the train tracks (another story I guess)

The purpose for being able to turn the lights off when there is no power is NOT for the purpose of formal safe disconnection so that I can work on wiring the circuits. If I had a really compelling reason to start work on the wiring while the power is out, I would switch off the main service entrance disconnect breaker.

The big reasons I can see for wanting to turn the lights off are:

1. To prevent all the lights from coming on when you are away, wasting electricity and increasing your electric bill. This applies if you need to leave home during the outage. Some people like to keep their appointments or social/family engagements even if the power is off at home. Some people like to go to work even if the power is off at home. And some people might even choose to go out for dinner because they can't cook, because ... you guessed it ... the power is off at home.

2. To prevent all the lights from coming on when you are asleep.

The situation I had with the utility maintenance going on that knocked out the power was more of a fluke. I simply took no chances and disconnected all the computers (one pull of the UPS plug since at the moment only 4 computers are in operating state), stereo system, and 4 TVs. I then instinctively went to turn off all the lights.

Perhaps I should have switched off the main entrance breaker, too.

In a previous power outage, my father had a burner element on in the kitchen. He never thought to shut it off when the power went off, despite moving the food off the burner (he decided it was done enough to eat it). After the power came back on a couple hours later, the burner stayed on, without a pan on it, for about a couple more hours. Fortunately it was on low.

So now I remember to also check the stove (which I did during this most recent event).

I did a quick survey of my 3/4-way light switches today. 2 are off when all switches are down. 3 are on when all switches are down. 1 has a push to toggle on/off rotary dimmer switch as part of a 3-way setup, so there is no way to know if it is off or not since the pushbutton state is unknown without power. So I'm going to replace that rotary with a definite up/down switch with the side-slider for dimming.

Wow, I know that it has been fairly quiet around here lately, but...........

pdh, just wire the lights initially so that when every light switch is down, the lights are off. You may end up with some up and down during normal use, but during a power failure, put them all down, and you'll know the lights will be off when power returns. And, honestly, this doesn't even sound like a problem so much as an aesthetic preference and pet peeve- for anything actually critical, you should be securing the breaker and not relying on light switches anyhow. The electrical costs for a single light are negligible in the scheme of things, especially considering the low-likelihood of the kind of thing you're talking about. I mean, you're talking about 10 cents a day for a 60W bulb left on 24/7. About $1/month for a typical CFL. I have a bunch I just leave on 24/7 intentionally because.

3-way switches are indispensible, especially for rooms with multiple entrances. I have 4 light switches controlling my dining room chandalier, and wouldn't have it any other way. One of which is a push-button, boo-yah!

I use Leviton devices for the most part but this technique works on P&S as well.

Assuming I did the rough and used a stuffing/switching convention that lets me readily identify what travelers go where, and assuming the colors of the travelers are different and consistent from end to end, at finish I do this...

I have a hierarchy of colors in place that in most of my cases goes black, red, white-phased black.

When installing the first switch in a 3 way set, after attaching the common first I then attach the highest color in the the hierarchy (present in that particular traveler) to the screw directly above the common. This leaves the lesser color to go on the screw opposite the common.

I then do the reverse of that at the other end of the traveler.

Both switches are installed into the box with the ground screws on the left ( the same way I install my receptacles).

When I have 4 ways in between I make sure that the colors match on their respective ins and outs. (both blacks on left and reds on right as an example).
These also get installed with the ground screws on the left.

Without dimmers or other devices that that need to be rung out to determine which 2 contacts connect in the "up" position, this technique gets all the switches down when the light is off.

I use Leviton "Togglers" to match standard switches and they have the opposite connection pattern as the Leviton switches. In this case connect them the same way at both ends of the traveler.

Occasionally I have to go back and flip a 3 way over after I power it up to make it perfect ( the best laid plans and all of that) but this works for me 99% of the time and I do it mindlessly (which probably accounts for the other 1%).

The reason I would turn them off has nothing to do with the reason I would secure a breaker. I never rely on switches to make a circuit "safe" for wiring work. It's all about making sure lights are off ... or on as the case may be. I can make a light be on by moving all switches down to off, then moving one back up. Then when I come back, if the power is back on, the lights should be in the desired on/off arrangement. FYI, I have 500W of nice warm kitchen task lights on one of the 3-ways (each end of access to the kitchen). CFLs are unsuitable for task lighting, at least for me. My electrical conservation is in the form of not having lights on unless I need them.


Switches that do up and down momentary contact, resting in the middle position, controlling a contactor/relay, would do the job. I would prefer this kind of setup.

Look into th4e GE RR7 relay system if you want momentary switch activation. Then all of your switch legs are run in bell wire and the number of switches is virtually unlimited.
I had the outside floods on this system at my mouse in Md. There were several places around the house with a switch that would turn them all on or off.

The RR7 are mechanically latching. Some lights I would want to work that way. Some others I would not. I want to look into some solid state switching, which I presume is how things like X-10 do it.

I have a bunch of things switched with SSRs and CMOS controllers but some day the "listing police" swat team will be storming my house.
"Come and get me copper"

They do have the U/R label but no assembly listing.
The opto 22 SSRs are a nice fit in a handy box and it is easy to rig a barrier between the line side and the LV side so I am "hold your nose" legal. If you use a 4x4 box the factory supplied barrier works if you cut a slot to fit over the SSR. Hammer the back KO down flat (if it isn't already) and apply the white heat sink goo. On the ones running my 11kw spa heater have a supplemental heat sink but that is not in a device box.
Once you get the line voltage side out of the way the sky is the limit with what you can do on the CMOS side.

If I were to run some twisted AWG#22 or such wire to the switches, with the intent to have them controlling solid state control devices, I'm wondering what the AHJ inspectors might think of that. Even if explained to them exactly what is going on (on the plans when permits are pulled), they might take the position that the next homeowner that buys the place from my estate might think the wire is OK to use after removing all the solid state and then wire it up as normal switch loops. So it might still need to be wired with AWG#14.

All of my LV stuff uses momentary contact switches so that would not be an issue here. I am not even sure I would trust regular snap switches going into CMOS. I bet they are pretty noisy and might depend on a certain amount of current just to keep them clean.

Just as with the NEC, our code in Britain does not address the issue of 3-way switch orientation, but I always wire them so that "both up" and "both down" are off (of course, our regular light switches here are up=off, down=on, but for 3-ways it works out the same on both sides of the Atlantic).

Where a 4-way is involved as well, I wire so that all switches up = off.

I wired my personal 3 ways at home and have had many customers request: both down = off. The problem is, now it bothers me if both are up for off - I catch myself turning them both down for uniformity (I think I'm not OCD.). I suspect those customers run all over their houses turning all their three ways down - completely defeating the convenience of them.

the easy trick is to reverse the orientation of the travelers, using common as reference - it seems to work no matter the switch brand.

what I really wanted to do was an RF programmable system, maybe next year.

I can't believe this topic has gone on for three (3) pages and one (1) month.

That I find more amazing than an electrician or customer worrying about the position of a 3-way.

Heck! How would you ever get the job signed off? the customer walk thru would destroy all your hard work!

I notice now a bit late, that chico gave a much more detailed description of traveler juggling - oops -oh well.

(this makes page 4)

I challenge anyone. 1 months time under NORMAL operation to maintain that condition!
It's fun reading, but........ You folks are outa your MINDS!!! I meen that in a good way!

"All down" = "off" will always be true if you wired it that way.

I suppose we could install the switches sideways, but then there would be the argument over whether both left/both right = off or one left/one right = off! Still, that's better than ground up vs. ground down I suppose!

Well, Greg, you and I are going to have to disagree about that one! I've already described -twice- how an installation the starts off as 'down=off' changes the very first time you flip any switch.

I dare say that it's simply not possible for 'down=off' to remain true after the first switch operation.

I think you two are looking at this from a slightly different perspective.

If you wire so that both down = off, then clearly as soon as you move one switch up you will be left with the situation where the other switch being down will result in the light being on.

But however many times you operate the switches, these conditions will hold true:

1. Both switches down = Off
2. Both switches up = Off
3. One switch up, one switch down = On.

Swap the travelers over at one end only and these three conditions will always hold true:

1. Both switches down = On
2. Both switches up = On
3. One switch down, one switch up = Off

For systems with an odd number of switches (an odd number of 4-way ones), then you have:

1. All switches down = Off
2. All switches up = On
3. Even number of switches up, odd number of switches down = Off
4. Odd number of switched up, even number of switches down = On

Swap the travelers over at one end only and these four conditions will always hold true:

1. All switches down = On
2. All switches up = Off
3. Even number of switches up, odd number of switches down = On
4. Odd number of switched up, even number of switches down = Off

So the design rule choices could be narrowed down to:

1. All switches down = Off
2. All switches down = On
3. All switches up = Off
4. All switches up = On

I'd prefer #1. The reason is because if I ever needed to change the switches when I did not have a means to see what the effect is (e.g. power is out or light bulb is blown) I'd be wanting to turn them to off in almost all cases. In the rare case I want to leave a light on, I can set them all down then set one up.

After reading thru 4 pages of this (which I am still laughing at), unless I missed something (highly possible at my advanced state of senility), I think I may have found the flaw. I'm one of the anal people who has all the screws north/south & all lights off if all switches are down ... then along came the wife. She wanted the old rotary dimmer on the dining room 3way. Now I'm scr... uh ... out of luck.

Is it the rotary aspect she really wants? Or the push-in to turn it on and off?

If all she wants is the rotary, take a 3-way version of the rotary control and connect its travelers together. Then connect those paralleled travelers to the common of a normal 3-way switch right beside it. Then the normal switch travelers are connected in the usual way. Then you will have a dimmer with no on/off effect, and a real switch right beside it that can be oriented with up for on and down for off. That's if you have the space to add a switch next to it.

Steve,

You're not out of luck.

Dimmers are available (even rotary) in 3 way configurations.

http://www.lutron.com/rotary/?s=17000&t=17200

Old fashion wife. She wants the push/push operation and the rotary. It's actually the end of a 4way system so it's a 3way rotary push/push dimmer. You just can't tell if it's on of off if there's no pwr.
PDH ... one of the guys at work told me the same thing but it's a single gang box and I'm too lazy to cut the wall.

To really solve this severe problem, we have 3-way-switches here now that always stay "up", push them once = on, push 'em twice = off again.

Would that help?

You would be surprised at just how neurotic people are.

Steve,

See the D-603P- model in the link that I posted.....

It's a push-on push-off rotary dimmer, 3 way.

One of my grandparents had those push-button switches that were sized to exactly fit the standard switch plate. So if the power was out, you had no idea at all about every light controlled by those, whether 3-way or regular. The only exceptions were a couple switches for a dozen outside floods, and one to control the sink disposal motor. The house was built in early 1950's and even had three phase power. As a kid visiting them often I thought they were great. I didn't think about the consequences at that time.

I couldn't tell you if a push on/off rotary single pole dimmer was on or off without power - unless it was the lighted version in which light off = power failed (in either position.)
:-))