Blue dimmer, no neutral, Hue lights = Won't work

[makahlow]

Hue lights aren’t working with non neutral, smart bulb configuration with bypass. Also - 50VDC offset on switch output (two of three switches)?

I’ve got a blue series dimmer powering two Hue bulbs in parallel. It’s a house built in the 1930’s, so no neutral for me. I just bought five of the switches to install around the house.

OK, the first problem - The third switch I tried works fine (one or two bypasses) so long as there are no bulbs connected. No DC offset. Everything works fine, I can pair the switch, it turns on and off both remotely and from the paddle. And with the Hue bulbs in, it works. It works until I put the switch in smart bulb mode and turn off the bulbs remotely. Then the switch loses power. I remove the bulbs, the switch comes back to life. I put the bulbs back in, turn the switch out of smart bulb mode, and it works (except of course that I’ve turned my smart bulbs off when I flip the switch).

What am I missing here?

Why was it the third switch I tried? The first two switches I tried were a disaster. using one or two bypasses (and five different bulbs). In smart bulb mode*, the bulbs do nothing but flash (OK, sometimes one of the two works while the other flashes), even with one incandescent bulb in parallel. When I checked the voltage, I get 120VDC (or thereabouts)… but with a 50VDC offset? It’s not my meter, the power reads clean on the supply line. I’m open to the idea that there’s some high or low frequency component that my meter doesn’t read correctly, but the Hue bulbs apparently don’t like it either. Is this a setting problem, or have I gotten a bad lot of switches?

* I did it with smart bulb disabled, but I can’t remember if it worked or not. I didn’t spend nearly twice as much on these switches as other options just to use them with dumb bulbs.

Thanks in advance.

[Bry]

So non-neutrals with smart bulbs are inherently problematic, even with a bypass or two. Some get them to work, so not so much.

One solution is to wire the light box hot, unswitched. You already have a hot and neutral there. This will power the bulbs full-time. The only downside is that you have to throw the breaker to be safe when changing the bulbs.

Then, using the 2-wire running between the switch and the light box, send the hot and neutral to the switch box. This will power the switch, making it a scene controller. You can then bind the switch to your bulbs.

Regarding your voltage measurements. House wiring is AC, so you should be using the AC setting on your meter, not the DC setting (your post refers to VDC measurements). The voltage measurements won’t really help here anyway. You’re just fighting a typical non-neutral problem, which the wiring modification will resolve.

[makahlow]

I read the documentation on the bulbs, and purchased bypasses because I knew they would be needed. If they’re not adequate, and the most common bulbs are “inherently problematic” requiring a wiring modification, then maybe they shouldn’t be sold as being able to operate without a neutral.

As for the VDC measurement - I was using the AC on my meter (the voltage at the socket measured ok). But at one point I flipped it to DC by accident and read 52 VDC at the socket, then tested the other two switches (one gave the same, the other gave a zero DC reading). The DC offset was on the two switches that I couldn’t get to drive the Hue bulbs at all (without flashing) in smart bulb mode, even with an incandescent bulb in parallel.

As far as my limited understanding of circuits goes, there shouldn’t be a 50VDC offset on an AC signal. When that correlates with two switches causing the Hue bulbs to flash even in parallel with an incandescent, I think that could be an important data point?

[makahlow]

Update on the two issues:

1. I replaced the switch which works in dumb switch mode with smart bulbs, but does not work in smart mode (the switch turns off when the lights are off), and it does the same thing as the third. So - two switches work in dumb bulb mode, even the two Hue bulbs turned off. Works perfectly fine - except that in this mode, it cuts power to the bulbs. When I put the switch into smart bulb mode and turn off the bulbs via Zigbee (powered on but lights turned to 0%), the switch turns off. When I remove the bulbs from the sockets, the switch comes back on (and works perfectly fine so long as there are no bulbs in the socket). This occurs with one or two bypasses in parallel with the two Hue bulbs. That’s two switches that show the same behavior.
   Conclusion - there’s some kind of interference between the Hue bulbs when turned off (powered, but turned off via Zigbee) which interferes with the switch. This doesn’t occur when the bulbs are powered off (no current flowing) in dumb bulb mode, or in smart bulb mode when the bulbs are removed (empty sockets). In this case, the switches work perfectly fine. At no time do I get the Hue lights blinking - it’s the switch that cuts out when the bulbs are in the circuit and turned off.

2. I took one of the two switches that wouldn’t work even in smart bulb mode with two bypasses (Hue lights blinked, measured 50VDC along with 120VAC at the socket) put that into a different circuit with a neutral wire, also driving two Hue bulbs in parallel. Result - same damned thing, one of the Hue bulbs invariably flashes. (I’m not going to try the second 50VDC offset switch in this circuit; I’ve already spent two days swapping out bulbs and switches. I also can’t try to see what happens when I turn the bulbs off via Zigbee, because they’re never on long enough to properly connect.)

[rohan]

I think you’re missing a pretty key piece of information about how these switches work in non-neutral configuration. Apologies, if this is information you already know, but I want to make sure it’s explicitly clear.

• A normal dumb switch simply interrupts the hot wire. The switch itself consumes no power (unless it has an LED or something built in).
• A smart switch constantly needs power so that it can operate (run it’s radio, lights, etc). Normally it does this by having it’s own connection to neutral.
• When you operate a smart switch in non-neutral mode, it’s constantly trickling a little bit of power through the dimming circuitry into the bulbs so that it can use that power to power itself. This works perfectly fine with incandescent bulbs (since they need a lot of power before they light up) so the switch leaches the power and the incandescent bulbs don’t light up. With dumb LEDs, they use a lot less power before they light up, so you need one or more bypasses to let that current leak without any flashing or the bulbs dimly glowing.
• Smart bulbs are different. They expect to always be powered (like the smart switch) and when they are off, they are completely off besides the power they are using for their radios. When a smart dimmer in non-neutral mode tries to leach power through them, there’s not enough trickling through for it to be powered up at all. Stacking bypasses (I’ve seen people add 3 or 4 in this forum) often fixes the problem. But an easier solution is the one that @Bry already suggested of rewiring at the fixture to send a line/neutral back to the switch and convert it to a scene controller with a neutral installed. You wouldn’t need any bypasses then.

[makahlow]

Rohan - This is exactly what’s confusing me. My circuits class was a long time ago (pre-digital) but I mostly remember Ohm’s Law.

I’ve been visualizing this as two resistors in series. One resistor is the switch (call it Rswitch) the second is the load - which in this case happens to be several resistors in parallel, light bulbs and bypass(es). Call it Rload = Rbypass (maybe x2) + Rlight1 + Rlight2.

If the load resistance is infinite, then the voltage drop across the first resistor (the switch) is zero. Ohm’s Law, no current, open circuit.

With a bypass “resistor” (resistor & capacitor) in place (I’ve read that they present approximately 5.6kohm resistance at 60 Hz), and the light bulbs removed, now there’s about 20 mA of current flowing through the circuit, which is enough to power the switch. I know this because it does for my circuit. That number seems way low - let’s just call it “x” - whatever it is, it is sufficient to operate the switch. It works, it connects to my Zigbee network, I can control it, turn it off/on, change parameters, etc.

I add two light bulbs across the bypass in parallel. With the LED’s on, the two lights are drawing somewhere around 15 - 20W between the two of them, an effective load of something around 7 ohms. The contribution from the bypass is negligible, and we’ve got way more than enough voltage drop across Rswitch for it to operate. I can set the switch to dumb bulb mode, and it still operates.

Now I change the switch mode to smart bulb and use Zigbee to set the output of both bulbs to zero. If this were a simple Ohm’s Law circuit, each bulb might each draw a half a watt in standby (so says the Google). Since they’re in parallel with the bypass, that means that the total current draw is now iload = 0.5 + 0.5 + x… whatever x is, that’s greater than x. The switch should see a larger current (larger voltage drop) than it does with the bulbs unscrewed. But the switch turns off - not enough current to operate (?).

There’s my confusion. I add two light bulbs in standby mode, in parallel, across the bypass. That should give me a larger current & larger voltage drop across the switch. But instead the switch turns off. When that happens, I unscrew the bulbs, and the switch turns back on. I put the switch into dumb switch mode, which needs to supply some small voltage to the load (otherwise the circuit can’t operate), put the bulbs back in and voila, the switch works. (That just doesn’t keep my smart bulbs powered, which is why I bought these switches.)

Bottom line - since Ohm’s Law hasn’t been repealed, this series/parallel resistor Ohm’s Law visualization is too simple to explain what I’m seeing. There’s something about the interaction between the lights and the switch which causes the switch not to operate… and now I’m in handwaving territory. The light bulb load might have a substantial inductive component, for instance (and probably does). There are several voltage converters with a lot of funky frequencies which can bleed back and forth. And I know that my picture of the switch as a simple resistor (or even a second load parallel across a resistor) is too simplistic. All of which is beyond my background, although I might be able to follow someone else’s explanation.

So here’s the heart of my question. Either I’m seeing an engineering/manufacturing flaw (I’ve got bad switches), a design issue (I’m working outside of the parameters for which the switch is designed), or a configuration issue (there’s some little box setting a parameter that I need to set). Which one (or more than one) of those three am I looking at?

(And this is for the two switches that kinda sorta work. The ones with the DC offset, that cause the lights to flash even when connected ta a neutral wire… that’s a whole ‘nother thing.)

edit - Rewiring to bypass the switch isn’t really a good option. My goal with these switches is to set the house up so that my spouse & friends can come in and operate everything with the wall switches. I want my smart house to be able to work like a dumb house if needed - it shouldn’t need a Ph.D. in home automation, or be dependent on remembering the names of every one of 50+ light bulbs.

[rohan]

@makahlow

Bottom line - since Ohm’s Law hasn’t been repealed, this series/parallel resistor Ohm’s Law visualization is too simple to explain what I’m seeing. There’s something about the interaction between the lights and the switch which causes the switch not to operate… and now I’m in handwaving territory. The light bulb load might have a substantial inductive component, for instance (and probably does). There are several voltage converters with a lot of funky frequencies which can bleed back and forth. And I know that my picture of the switch as a simple resistor (or even a second load parallel across a resistor) is too simplistic. All of which is beyond my background, although I might be able to follow someone else’s explanation.

Ohm’s law as you’ve described only works for DC circuits. AC circuits are a lot more complicated to model mathematically because of impedance. V = I x Z (Impedance) is the correct formula for an AC circuit. I personally don’t want to go through the math and reasoning in further depth.

@makahlow

So here’s the heart of my question. Either I’m seeing an engineering/manufacturing flaw (I’ve got bad switches), a design issue (I’m working outside of the parameters for which the switch is designed), or a configuration issue (there’s some little box setting a parameter that I need to set). Which one (or more than one) of those three am I looking at?

I think you’re operating at an edge case of the what the switch is designed for.

1. It will operate best with a neutral.
2. With dumb bulbs, it will operate fine with one or multiple bypasses.
3. With smart bulbs, you’ll need to add more bypasses until it starts working.

The options moving forward are as follows:

1. Install additional bypasses with your smart bulb setup until the dimmer remains powered when the bulbs are off in Smart Bulb mode.
2. Rewire at the light to pass a hot/neutral between the switch box and the fixture.
3. Add a new wire between the switch box and the fixture so you have 3 conductors (line/load/neutral).
4. Switch from Smart Bulbs to dumb bulbs and disable smart bulb mode.

I don’t recommend option 3 since it’s going to be considerable rewiring and potentially making holes in drywall.

Option 4 isn’t ideal since you lose your smart bulbs.

That leaves you with Option 1 and Option 2.

@makahlow

edit - Rewiring to bypass the switch isn’t really a good option. My goal with these switches is to set the house up so that my spouse & friends can come in and operate everything with the wall switches. I want my smart house to be able to work like a dumb house if needed - it shouldn’t need a Ph.D. in home automation, or be dependent on remembering the names of every one of 50+ light bulbs.

If your smart bulbs are Zigbee (Hue are) and directly controlled by the same Zigbee hub as the one controlling your light switches (which I assume is true since you are running Home Assistant). You can directly bind your switch(es) to individual or groups of bulbs. This will mean that they will operate perfectly normally with the wall switches without any dependencies on remembering smart bulbs or not. It just works, whether your hub is online or not.