Smart wall dimmer switch with small LED loads - and I did some homework, sounds like I will need to pay attention of leading edge vs trailing edge, and triac, (and dimmable LED bulbs obviously)… and on top of all that, if the LED load is small enough, I will need a small “bypass” device.
Now I’m confused - what does a “bypass” do exactly?
Hi,
Move the dimmer module to where you can access Live / Neutral and switch wiring. Two-wire or “No Neutral” devices cause endless trouble as current has to flow through the load (light bulb) to power the module, and modern LED bulbs are VERY complex with MANY different ways of operating.
Bypass devices try to give a constant current flow to power the module, but are dimmer-specific and can be different. Don’t mix brand A with brand B, even though they may be a Class X capacitor and bleed resistor.
L---MODULE-----BULB-----N // no current for the module when off
L---MODULE-----BULB-----N // small trickle for the module when off
| |
|-BYPASS-|
Shelly’s snubber isn’t the same as it designed for a completely different purpose - transient suppression.
You don’t mention US/ EU/ 110V/ 220V/ 240V/ 48VDC so it is hard to be specific, but in the UK behind switch units are a bad idea, but behind the ceiling rose works well.
I’ve decades of experience with everything from US X10 modified for the UK to the latest Matter modules. Conclusion? MOVE THE MODULE to another location to save a lot of pain.
Yes some US/ UK/ UK systems can work without a Neutral and with a “lucky” LED bulb, but many don’t with glowing when off, strobing, bad dimming, and worse.
Of course I forgot to mention this is a US install, so 120V split phase.
3 extra piece of information, hoping it would help someone down the road:
Further googling (sorry can’t find it now) inside reddit also suggested that specifically for Shelly bypass, there were 2 HW versions, and the earlier one was not suitable for 120VAC environment.
James had a good point - those seemingly the same little things are designed for the same / similar purposes, but the specifics could be different depending on what exactly being encapsulated, and they are typically designed specifically for their own brands of (no-neutral) dimmers. And obviously the specifics of each dimmer are different.
And I will update my earlier posts to include links.
So, action plan for me:
Ideally, I can get rid of the no-neutral situation for the dimmer (or FWIW, any switch)
And search for the kind of smart dimmers with better compatibilities among all kinds of bulbs, ideally pick something that would support both leading edge or trailing edge via some configuration for the dimmer itself.
and then if the above fails, it’s going to be trial and error - to find the lucky combination of dimmer / bulbs / bypass.
Did you state a problem you are trying to solve?
AC dimmers use a triac device that requires a*minimum holding current to stay latched and conduct through each half-cycle of the AC waveform. With traditional incandescent bulbs, this isn’t a problem—they’re resistive loads and draw plenty of current.
LED bulb is low current and will fall below the triac’s holding threshold and present as flickering, inconsistent dimming or the not turning on at all.
Jumping in as I have a strange issue and the company I bought the dimmers from suggest I use a bypass.
I setup a bunch of dimmers modules with neutral (so I wouldn’t expect needing a bypass) in a new place. (Australia, 240V single phase, dimmers are sold as OzSmartThings and show as Sunricher on ZigBee).
With some bulbs they work fine (basic 2.7K dimmable OSRAM). With more expensive ones , I get an occasional flicker when at 100% (Philips ‘selectable’ ones).
The bulbs are fairly high power for LEDs, the Philips are 14W and I have two of then on one of the circuits.
The flickering happens regularly, 1 to 5 times per minute, but only with the Philips bulbs and only at 100%. If I dim to 95% they do not flicker. Unfortunately those dimmers don’t seem to have a way to configure a “Max” value.
I also have some ceiling downlights that occasionally (much more rarely) flicker as well. I don’t have the brand at hand but there are 3 on one dimmer so the load should be reasonable.
I’m trying to understand why a bypass would help in that scenario and what is actually happening at the electrical level… I understand the use for a bypass on no-neutral setups but not in this use case.
It almost feels like at 100% the dimmer is still cutting a tiny bit of the wave, possibly small enough to csuse something funny bulb-side but I can’t figure out what.
Another interesting data point is if I mix the two bulbs types on one dimmer circuit then none of them flickers.
If your bulb is within the dimmer specs (W), I don’t see need for bypass.
To have flickering at 100% level sounds like the zero-crossing detection of the dimmer is tiny bit off.
We are on the same boat I think. And Steve above has captured my problem correctly.
I believe the older-style dimmers were OK for incandescent bulbs, but in the era of LED bulbs (even dimmable ones) we will be seeing inconsistencies on the light output, especially when dimmed.
(Mine would not flicker at 100%, though.)
So sounds like I need one of those bypass. So the question remains, what does those bypass do exactly? Are they essentially increase the load of the circuit, so that to be above the holding threshold for triac’s? Or how would those help address the flicker-when-dimmed issue?
Depends on your setup. Specs of the bulb, specs of the dimmer?
Generally precise zero-cross is tricky, so dimming out of 3-97% should be ignored. Occasionally ~100 can become 0 or ~0 become 100.
There are lots of LED failure modes due to the VERY different types of PSU inside a very small space. Some add smoothing (some don’t), some have small load resistors (bleed off capacitively coupled current to stop glowing when OFF), some are linear, some capacitive, etc.
Dimmers have also become complex and are no longer a simple leading phase edge (diac + triac), but but can use trailing egde, PWM, etc.
Mix the two - and you get a mess. Yes, you could explore waveforms with a mains-isolated 'scope ( ), but swapping bulb types is the simplest solution.
As linked, ballast are designed to trickle current to power a dimmer with no neutral. They will also add a load (probably capacitive), so you are seeing some change.
The circuitry in a LED and a dimmer can also interact - I’ve seen a snubber network oscillate with a LED load very badly when on at 100%.
Hacks like adding capacitive loads (safe Class X or Class Y, depending on the circuit), or a small resistive load (10k, rated for I2R heating) might work, but different bulbs is much easier. The old hack of one incandescent bulb in a fitting is very hard these days.
If you really want to explore, watch some BigClive LED reverse engineering videos on YT, and perhaps some EEVblog reviews on test gear (e.g. understand WHY you MUST use an isolated 'scope probe ).
this post!
), but swapping bulb types is the simplest solution.