Will dimming incandescent bulbs with a rectifier diode cause noticeable blinking

[Ian.M]

Assuming they are currently all in parallel, all you need to do is split the live at the mid-point and run both lives back to the switch.  On the diagram above's wires to the lamps, top and bottom wires are the two lives after splitting them and the middle wire is the existing neutral.  Same number and type of bulbs each side of the split of course.

Try two lights in series before you get into this.  If you aren't happy with the brightness, you'd be out your time and the wire. 

If I had to do this without going X10, I'd use a series dimmer module, with a relay to short it out for full power, so I could set the brightness required.  Leave off the knob and put it completely inside the wall box if you don't want anyone changing it. 

 Read this: http://www.standardpro.com/product-information/halogen/faqs.  There's a critical power level below which the glass capsule wall doesn't get hot enough for the deposited Tungsten to get scavenged by the halogen and operating in this region will vastly shorten bulb life by blackening and premature filament failure.  You certainly should make your control system go to full power for several minutes after dimming before switching off if you want to avoid having to stock spare bulbs in case lots.

[Jay112]

Assuming they are currently all in parallel, all you need to do is split the live at the mid-point and run both lives back to the switch.  On the diagram above's wires to the lamps, top and bottom wires are the two lives after splitting them and the middle wire is the existing neutral.  Same number and type of bulbs each side of the split of course.

This is really interesting to me. I had to concentrate for a long time before being able to understand it enough to ask questions.

I already have a spare unused wire running through the first half of the conduit, because there used to be 2 light switches, 1 for each side of the room, and I combined them into 1 switch a few years ago.

Forgive me for breaking this down and simplifying it, but could you please check over my reasoning here?
-When the lights are turned on full power, the existing neutral acts like a regular neutral wire. So we have 2 hot wires (one for each side of the room) and 1 neutral wire.
-When the lights are turned to half power, the neutral wire is disconnected from the box, and doesn't act like a regular neutral anymore. Instead, the neutral acts as a connector to create the series pairs. One of the previously-hot wires now acts as the new neutral, completing the circuit back at the box.

Was I far off?

Thanks for the info about the halogens. I had never heard that before, and it was good to learn. I'm actually planning on using regular incandescents here. The reason I had mentioned halogens earlier in thread is because I got the "rectifier diode as a dimmer" idea from a halogen fixture I had opened up to repair (it was this fixture here, which used a 1N4007 diode for the "Low" setting: http://www.amazon.com/gp/product/B0006SUZJM). The bulbs on those halogen fixtures seem to have good life spans to me, even though I'm running them on Low most of the time.

I'm also still considering the idea to use a series dimmer module inside the box. But I'm really enjoying learning all these useful techniques, and I would like to understand them before I make a decision.

[Ian.M]

Assuming they are currently all in parallel, all you need to do is split the live at the mid-point and run both lives back to the switch.  On the diagram above's wires to the lamps, top and bottom wires are the two lives after splitting them and the middle wire is the existing neutral.  Same number and type of bulbs each side of the split of course.

This is really interesting to me. I had to concentrate for a long time before being able to understand it enough to ask questions.

I already have a spare unused wire running through the first half of the conduit, because there used to be 2 light switches, 1 for each side of the room, and I combined them into 1 switch a few years ago.

Forgive me for breaking this down and simplifying it, but could you please check over my reasoning here?
-When the lights are turned on full power, the existing neutral acts like a regular neutral wire. So we have 2 hot wires (one for each side of the room) and 1 neutral wire.
-When the lights are turned to half power, the neutral wire is disconnected from the box, and doesn't act like a regular neutral anymore. Instead, the neutral acts as a connector to create the series pairs. One of the previously-hot wires now acts as the new neutral, completing the circuit back at the box.

Was I far off?

I think you've got it.

I'm also still considering the idea to use a series dimmer module inside the box. But I'm really enjoying learning all these useful techniques, and I would like to understand them before I make a decision.

Plug in mains stuff: do what you like as long as it doesn't violate basic electrical safety.  Hard-wired (permanently installed) mains stuff:  Keep it as close to code compliance as possible and if its non-compliant, keep the weird stuff to a minimum in as few locations as possible so you can easily revert it to full code compliance.   Otherwise, if you ever need an electrical inspection and get a hard-ass inspector, you are in for a world of $expensive$ grief.

[Jay112]

Plug in mains stuff: do what you like as long as it doesn't violate basic electrical safety.  Hard-wired (permanently installed) mains stuff:  Keep it as close to code compliance as possible and if its non-compliant, keep the weird stuff to a minimum in as few locations as possible so you can easily revert it to full code compliance.   Otherwise, if you ever need an electrical inspection and get a hard-ass inspector, you are in for a world of $expensive$ grief.

Excellent advice! I would never do something like this to my regular house. I do these modifications to a unique rural cabin-like structure that nobody would ever live in and nobody would ever inspect.

I was thinking about using a wireless relay like this for switching between off, low, and high power: http://www.amazon.com/gp/product/B00D56IX00, mainly because it's easy for me to send the wireless commands to the relays via my Arduino units.

When that relay is set to "latching" mode, only 1 of the 4 relays is allowed to turn on. So if "A" is on and I press "B", "A" would automatically turn off while "B" is switched on.

I'm not sure if it waits for the previous one to turn off before switching on a new one. The change happens so fast that I can't imagine they'd both be on for more than a few ms. I'm not sure what a good way is to test this (I don't have an oscilloscope yet). I'm planning on asking the manufacturer eventually.

But either way I'd still be curious about the answer: IF 2 of the relays happen to both be on at the same time for a few milliseconds, would that be enough time to cause a short on my line when switching from "High" to "Low" power?

This is my rough drawing of how I'm thinking about wiring it up (sorry I'm not good with schematics or diagrams):

[Ian.M]

Why have you turned it around?  You've now got the feed on the right, and its a disaster as if they are ever both on, its a dead short across the supply.

I've annotated IanB's diagram for two separate relays:

Note that A is shown ON and B OFF.

With both off all lamps are off.
Turn A on for dim.
Turn B on for full,
Turn both on by mistake and half the lamps light with full power.

I suggest spitting the neutral as then there is less work taping wires to identify them as hot and keep it near code compliance + it keeps the switching on the hot side.  Put an override switch actoss B and you will get at least *SOME* light in an emergency.

[Jay112]

Why have you turned it around?  You've now got the feed on the right, and its a disaster as if they are ever both on, its a dead short across the supply.

I've annotated IanB's diagram for two separate relays:

Thank you!! The annotations were very helpful. I'm sorry I needed things to be explained to me so simplistically.

I'm going to make a small test circuit using the same relay switch and small incandescent bulbs, before trying it out on the ceiling lights.

[Jay112]

With both off all lamps are off.
Turn A on for dim.
Turn B on for full,
Turn both on by mistake and half the lamps light with full power.

I suggest spitting the neutral as then there is less work taping wires to identify them as hot and keep it near code compliance + it keeps the switching on the hot side.  Put an override switch actoss B and you will get at least *SOME* light in an emergency.

Thanks for this additional information!

[Zero999]

A series capacitor of the right value would dim without introducing any flicker. Motor run capacitors are rated suitably and usually cover roughly the right range of values.

That's a good idea.

An inductive fluorescent lamp ballast would also do the trick.

[Jay112]

What are your opinions about using short (1"-2") 20-gauge wires to make the connections to the relays? Is that generally considered acceptable because of the short distance? The max amps passed through the wires would be about 4 amps when the lights are on full power.

[Ian.M]

Don't even think about using 20 AWG wire.  *NEVER* use any conductor rated for less current than the breaker or fuse feeding it.  What's your local fire department's response time out at your cabin?

[Jay112]

Don't even think about using 20 AWG wire.  *NEVER* use any conductor rated for less current than the breaker or fuse feeding it.  What's your local fire department's response time out at your cabin?

Thanks! Yes, everything is slow out here, even the fire department.

I was pleased to find out that my relay screw terminals will handle 12 AWG. I'll probably use stranded so that there's less force on the parts.

[Jay112]

I did a small test using two 60-watt bulbs on a table, and everything seemed to work fine.

Then I hooked it up to the ceiling lights, and at first everything seemed fine. I was surprised it worked on the first try. I loved the brightness of the low-power mode.

But then when I was switching between low power and off, there was a small explosion and the fuse tripped.

This is how the back of the relay board looks:


It looks like the 2 contacts that are blown were both the NC and the NO lines on the A relay (A meaning the same A as in Ian.M's annotated image above).

Do you guys have any theories as to why this happened? The individual relays say they are rated for 12A at 125V. This is what the specs on the board say (http://www.amazon.com/gp/product/B00D56IX00):
"-Max.Switching Voltage: 110VDC 240VAC. Max switching current:15A
-Rated switching power: 1,875VA 360W for each relay"

Is it possible that while the A contact is switching from NO to NC, somehow it creates a short?

[Jay112]

I should mention too that I was switching between the different modes (full, low, and off) a few times before it blew. It definitely didn't blow on the first try.

[IanB]

It looks like the 2 contacts that are blown were both the NC and the NO lines on the B relay (B meaning the same B as in Ian.M's annotated image above).

NC on the B relay is not connected to anything. How could it have current flowing through it?

[Jay112]

NC on the B relay is not connected to anything. How could it have current flowing through it?

Oops, I'm sorry! I meant the A relay. I'll try to edit my previous post.

[Jay112]

I have some new data that might be helpful:
When I was testing it with only two 60-watt bulbs on the tabletop, I heard a small "pop" when switching between the modes, and the dim setting didn't work anymore. I was thinking maybe one of the relays was defective, so I moved it over to a different relay (since the unit has 4 relays, and I only need 3), and it worked fine.

Now I just tested it and found out that the original relay I was using for the dim setting (the one that broke earlier) when testing on the tabletop (for the "A" relay wiring in Ian's diagram above) doesn't switch anymore, and always stays in the NC position.

I hope I explained that clearly enough. Basically there are 4 total relays on the board, and 2 different ones that were used as "A" broke.

[IanB]

From what you have described, it sounds like the NC contacts welded together. This might happen if the relay was used beyond its rated specification, or if the relay was not capable of the switching duty given in the product details.

From your product link it states that the device can switch 15 A, and yet your picture shows the relay sitting on a circuit board with solder joints and copper traces intended to carry this current. My bullshit detector is rumbling at this point.

[IanB]

Here's an example of the kind of relay that would handle this duty:

http://www.newark.com/omron-industrial-automation/ly2-12vdc/relay-dpdt-120vac-28vdc-15a/dp/42M2297

[Jay112]

From what you have described, it sounds like the NC contacts welded together. This might happen if the relay was used beyond its rated specification, or if the relay was not capable of the switching duty given in the product details.

From your product link it states that the device can switch 15 A, and yet your picture shows the relay sitting on a circuit board with solder joints and copper traces intended to carry this current. My bullshit detector is rumbling at this point.

Very interesting! Thanks for the response.

I'm using 2 of these same relay modules in other places, but only for low-power DC (less than 100mA each relay).

I'm curious about this question: Is there any possibility that some relays aren't supposed to have a 120v hot on the NO while there is a neutral on the NC (or vice versa)? Is it possible that that's too close when the relay is switching between on and off, and somehow there's a short from arcing or anything like that?

[Jay112]

Here's an example of the kind of relay that would handle this duty:

http://www.newark.com/omron-industrial-automation/ly2-12vdc/relay-dpdt-120vac-28vdc-15a/dp/42M2297

I have ones just like that. Maybe that's what I'll try next. I can use a different way of switching it wirelessly, by having a wireless outlet apply 12v power to the coils.

I think with that method I would have 1 (wireless) switch to turn the lights on and off, and a separate one to activate that relay, which would switch between Hi and Low modes. Does that sound feasible? I think I already have all the parts here.

[IanB]

I'm curious about this question: Is there any possibility that some relays aren't supposed to have a 120v hot on the NO while there is a neutral on the NC (or vice versa)? Is it possible that that's too close when the relay is switching between on and off, and somehow there's a short from arcing or anything like that?

The thought occurred to me, but I'm not sure that's the problem. Relay contacts are supposed to be rated to interrupt a certain current at a certain voltage. I don't think it should matter if there is a voltage across the NC and NO contacts at the time of switching. When buying a branded relay I suppose it may be possible to call product support and verify if there is any problem using the relay in that configuration.

[Jay112]

I tried using a large DPDT relay to switch the line between hot and neutral. I put the hot on one side, and the neutral on the other side of the relay, so that they were farther apart when switching.

I was using the same wireless relay that I was using before, only I used it to turn the larger DPDT relay on and off. The DPDT I was using had a 120v coil. The only power going through the wireless relay was to energize the coil.

At first everything seemed to be working wonderfully again. But after toggling the lights on and off a few times, I heard another "pop", and another relay blew on my wireless module. So now 3 of the 4 relays are blown!

I'm wondering if it was just a faulty module. I have another similar one by a different manufacturer that I could test, but I would hate to blow that up right away because I'm overlooking a different problem.

I also have a 12v and 5v DPDT relay that I could use instead of the 120v one, but I wonder if that would've made any difference.

[Ian.M]

I agree with IanB.

Don't believe the ratings on cheap relay boards.  Neither the tracks, nor the contacts are heavy enough for the claimed rating.   If you reinforce the tracks with wire, they will typically switch an incandescent load of no more than 20% of their nominal rating and if you want them to last, 10%!.   It sounds like your wireless relay module isn't even rated to switch inductive loads like a heavy duty relay coil.

I believe the ratings are entirely bogus:
"-Max.Switching Voltage: 110VDC 240VAC. Max switching current:15A"
"-Rated switching power: 1,875VA 360W for each relay"

Its extremely rare to see so little derating for DC.  I certainly wouldn't try to switch 15A at >24V DC with it.   Relays don't care about the difference between Watts and VA.  When closed, its the current that counts, when open its the voltage so the 360W 'rating' implies the relays cant handle more than 3A, possible only 1.5A if that was supposedly for a 240V AC circuit.

I've run some numbers through a trace width calculator.  The contact traces look to be about 0.1" wide.  If the board is 1oz copper, they'll cook themselves off the board long before the current reaches 15A.  If its 2oz, there'd be approximately a 45deg temperature rise.  I doubt its 4oz copper which would be required for an acceptable temperature rise.

Your relay with welded NC contacts is characteristic of a cheap relay trying to close on an excessive current.   A 60W 120V bulb has a cold resistance of approximately 24 ohms, giving an inrush current of 5A.  Contact bounce while closing + 5A current melted the contact faces enough for the to weld.  Game over.

The best thing to do with that P.O.S. relay board is return it as "not fit for purpose - failed while switching a 60W bulb".  However if you are stuck with it, remove *ALL* the relays, and patch the coil drive connection to the terminal block so you can use good quality off-board relays.  Measure the resistance of the coils of the relays you remove so you can determine the max coil current capability so you don't blow out the relay drivers.   The removed relays should be given the big hammer treatment and trashed!

Now you see why I recommended an X10 dimmer module.  UL rated so it wont burn your cabin down, and if it craps out, warranty replacement is a phone call away.

[Jay112]

Thanks for the help and the info, guys. @Ian.M, I'll definitely be taking a look at those X10 modules. But I really like to learn as many techniques as possible, and the process is just as important for me as the end goal. Already today I learned many useful things that'll probably be helpful to me in the future.

I just blew 2 different large DPDT relays, like the Omron one that was linked above. One was 120V and the other had a 12V coil. Each one had a relay socket with screw terminals, and all the connections were clean. I was able to flip the lights between full power and half power successfully a few times, and just like previously the relays blew between toggles.

Are you guys feeling certain that it's okay to put a hot and a neutral on each end of a relay's contacts? They seem awfully close to me. When the relay is switching between the 2, there's probably less than a millimeter of "empty" space in between the contacts.

[Ian.M]

It depends on the relay.  If it comes from PRC, they'll print whatever rating you like on it, but the guts will be the same low voltage low current crap.  If its a brand named relay from a reputable manufacturer, it will actually live up to its ratings.  However, if your circuit has too much inductance, you'll probably need to add snubbers across the contacts to prevent arcing.   

If you use a heavy duty contactor it will certainly be rated for changeover switching line to line or line to neutral.  However the one's I've used had a snap action with about 5mm travel on the moving contacts, with a double break (i.e. both ends of the moving contact opened) giving 10mm clearance.  They also had separate moving contacts for NO and NC so that any arc couldn't blow across the gap between the fixed NO and NC contacts and cause a line short.