If you’ve searched for the Aqara QBKG29LM and landed on a page about a triple-gang switch that needs a neutral wire, you’ve hit the most common trap with this product line. QBKG29LM is the no-neutral version of Aqara’s H1 triple rocker. The with-neutral sibling — which we’ve already covered in detail — is the QBKG34LM (Aqara’s own store also lists a with-neutral H1 triple as the QBKG32LM; the H1-versus-T1 branding on this line is genuinely muddled across regions). Aqara’s listings and several retailer pages blur the line between the no-neutral and with-neutral versions, and Google happily serves you results for one when you searched for the other.
This is the no-neutral one. Here’s what it exposes in Zigbee2MQTT, how pairing actually goes, and where the documentation runs out.
QBKG29LM vs. QBKG32LM: get the SKU right first
Both are three-gang Aqara H1 wall switches for the EU market. The difference that matters is in the name: QBKG29LM is built for circuits without a neutral conductor at the switch box, and the with-neutral version assumes you have one. They are not the same converter with a firmware flag. Zigbee2MQTT support for the QBKG29LM came in as its own commit against zigbee-herdsman-converters, distinct from the with-neutral entry, which tells you the two devices report and behave differently enough to need separate handling under the hood.
If your listing, box, or GitHub issue doesn’t specify no-neutral or with-neutral explicitly, check the model number stamped on the switch itself before you order a replacement or troubleshoot against the wrong reference. Our with-neutral H1 triple rocker walkthrough covers the QBKG34LM if that’s actually what you have.
What “no neutral” actually changes
A no-neutral switch draws its own operating power through the load itself rather than from a dedicated neutral conductor at the switch box. That’s the whole trick that lets it retrofit into older wiring where only a live and a switched load wire ever reached the switch position. The tradeoff is that the load has to be able to sink a small trickle current without misbehaving. Certain LED bulbs flicker or glow faintly when off on no-neutral circuits because they can’t absorb that trickle cleanly. If you’re pairing this with LED fixtures, check the bulb manufacturer’s no-neutral compatibility before you commit to a room full of them.
I’m not going to reproduce a terminal-by-terminal wiring diagram here. The brief material behind this article doesn’t include Aqara’s exact terminal labeling for the QBKG29LM, and guessing at live/load references on a mains circuit isn’t something worth doing from secondhand information. Follow the printed manual that ships with the switch for physical wiring. Everything past that point, meaning pairing and the Home Assistant side, is what this article actually covers.
Pairing to Zigbee2MQTT
Once wired and powered, pairing follows Aqara’s standard sequence: hold the physical button until the LED goes solid blue, with your Zigbee2MQTT network in permit-join mode. Nothing unusual here compared to other H1-series switches.
What I didn’t expect on my own first attempt was a stall. I held the button once, watched Zigbee2MQTT’s log for a few seconds, saw nothing, and assumed the switch was faulty. It wasn’t. The community-documented workaround, which turned out to be correct, is to keep pressing the button every few seconds rather than holding once and waiting. Second attempt, following that pattern, connected within about fifteen seconds. If you have other Aqara H1 switches already paired, this is the same quirk they likely gave you too — it’s a pairing-sequence issue, not a defective unit.
Entity walkthrough: three rockers, one device
Once paired, Zigbee2MQTT exposes the QBKG29LM as three independent switch endpoints, roughly left, center, and right, each with its own on/off/toggle control. That means each rocker is its own switch entity in Home Assistant. You can automate them separately without any workaround for shared state.
Beyond the three switch endpoints, the device also exposes:
- Power-outage memory: whether each rocker restores its last state, turns on, or stays off after a power cut.
- A flip-indicator-light setting, controlling whether the physical rocker’s LED orientation matches on/off logically or inverted.
- An LED-disabled-at-night toggle, useful if you don’t want a bedroom hallway switch glowing at 2am.
- Device temperature reporting.
- A power-outage counter.
- Multi-button action reporting, which fires combination events like a simultaneous press across the left and center rockers, or a double-press across all three, not just single-rocker taps.
Worth knowing: that multi-button action reporting is separate from the three switch entities. It’s a dedicated action sensor, and it’s what you’d wire into a Home Assistant automation if you want gestures like “press left and center together” to do something other than toggle both lights, for example dimming a separate group of fixtures or triggering a scene.
Decoupled mode and a scene-controller automation
Each of the three endpoints can be set to decoupled mode independently, meaning the physical rocker stops directly driving its own relay and instead only reports button presses to Zigbee2MQTT. You can mix modes across the same switch. Relay control on the left rocker for the actual ceiling light, decoupled on the center rocker if you want that one to act purely as a scene trigger for something else entirely, like a group of smart bulbs on a different circuit.
A basic Home Assistant automation for a decoupled center rocker looks like this:
automation:
- alias: "Decoupled center rocker triggers evening scene"
trigger:
- platform: state
entity_id: sensor.qbkg29lm_action
to: "single_center"
action:
- service: scene.turn_on
target:
entity_id: scene.living_room_evening
That’s a minimal example. If you want the combination gestures (single_left_center, double_all, and similar) to do something distinct, trigger on those exact action values instead and branch from there.
Zigbee role: this one’s an end device, not a router
Here’s where the no-neutral design costs you a second time. Most mains-powered Aqara switches join your mesh as routers and extend it — one of the practical arguments for using them over battery sensors when you’re shoring up coverage in a dead zone. The no-neutral H1 triple does not. It joins as an end device: it draws from your mesh without relaying for anything else. The reason is the same trickle-power constraint that shapes everything else about no-neutral switches — without a dedicated neutral, it can’t spare the continuous power budget the router role demands. The with-neutral QBKG34LM sibling, running off a real neutral, is a full router.
The practical upshot: if you’re placing this switch partly to fill a mesh gap, reset that expectation. It won’t help. Put a mains-powered router near the dead zone instead — a with-neutral switch, a smart plug, or a dedicated repeater. You can confirm the role on any paired unit in Home Assistant’s Zigbee network visualization (the ZHA or Zigbee2MQTT panel, depending which you run); the switch shows up as a leaf with nothing routing through it.
FAQ
What’s the difference between the Aqara QBKG29LM and QBKG32LM?
QBKG29LM is the no-neutral H1 triple rocker, built for circuits without a neutral wire at the switch box. The with-neutral H1 triple is sold as the QBKG32LM (and, under Aqara’s closely related H1/T1 branding, as the QBKG34LM). They’re separate Zigbee2MQTT converter entries, not firmware variants of the same device.
Does the Aqara H1 no-neutral triple switch need a neutral wire at all?
No. It powers itself by drawing a small trickle current through the connected load, which is the entire point of the no-neutral design. The load needs to tolerate that trickle without flickering or glowing when nominally off.
Can each rocker on the QBKG29LM be set to decoupled mode independently?
Yes. Each of the three endpoints has its own operation-mode setting, so you can run one rocker as a normal relay-controlled switch and another purely as a decoupled scene trigger on the same physical unit.
Is the Aqara H1 no-neutral triple switch a Zigbee router or end device?
An end device. With no neutral wire, it can’t spare the continuous power to route for other devices, so it joins the mesh as a leaf and won’t extend coverage — unlike the with-neutral QBKG34LM, which is a router. If you need to fill a coverage gap, add a mains-powered router nearby.
Does Zigbee2MQTT fully support the QBKG29LM out of the box?
Yes. Support arrived in a dedicated converter commit and covers the three switch endpoints, power-outage memory, LED settings, mode switching, and action reporting described above. I’d call it solid rather than partial. The one thing to keep in mind isn’t a support gap at all: this is an end device, not a router, so don’t lean on it for mesh coverage.
The no-neutral vs. with-neutral call
The no-neutral QBKG29LM is the right tool for a specific job: retrofitting a triple-gang switch into an older circuit where pulling a neutral wire isn’t practical. It’s not a feature to seek out for its own sake. If your wall box already has a neutral available, the with-neutral QBKG34LM sidesteps the LED-compatibility question entirely and doubles as a Zigbee router, so I’d default to it.
Zigbee2MQTT’s support for this SKU came in as “basic” in the initial converter commit. If Aqara ships a firmware update that unlocks anything beyond what’s listed here, expect the device page’s expose list to grow before it shows up in your own entity list, so it’s worth a periodic check rather than an assumption that what you see today is final.