I run a lighting schedule that warms and dims on its own every evening. The hardware behind it cost a fraction of the equivalent Philips Hue setup, and the part that matters most cost nothing at all.
This is the companion piece to how different light colors affect your mood, sleep and focus, which covers why any of this works. Short version: a small set of cells in your eye reports ambient light to your body clock, they are most sensitive to blue-cyan around 479nm, and the published targets are at least 250 melanopic EDI lux at the eye during the day, under 10 in the three hours before bed, and under 1 while you sleep.
Nothing in that sentence requires 16 million colors.
What you are actually paying Philips for
Hue is good hardware. Being clear about what the premium buys is the only way to work out which parts you can skip.
A proprietary bridge and a polished app. The bridge is a Zigbee coordinator with Philips software on top. The app is genuinely better than most.
Reliability and ecosystem breadth. Bulbs, strips, outdoor fittings, switches, all tested together, all updating together.
Colour gamut and dimming quality. Hue bulbs dim smoothly to low levels. Cheap bulbs often stutter, flicker or cut out near the bottom of their range, which matters enormously for an evening setting.
Not needing to care how any of it works.
What you are not paying for is a special kind of light. A 2700K Hue bulb and a 2700K generic bulb put out broadly similar spectra. The physics is in the LED and the phosphor, not the brand.
Decide the spec before you shop
Most people buy the kit and then work out what to do with it. Reverse that. For a circadian setup the requirements are short:
- Tunable white, roughly 2200K to 5000K or wider. Colour is optional and mostly decorative.
- Deep, smooth dimming. Down to 5% or lower without flicker or cutout. This is the specification that separates good cheap bulbs from bad ones, and it is the one nobody prints on the box.
- Scheduling that runs without you. The single most important feature. A setting you have to remember at 9pm is a setting you abandon in a week.
- Local control, ideally. A schedule that depends on a company's cloud stops working when that company changes its mind.
Colour, music sync, and voice control are nice. None of them change what the light does to you.
Route 1: no smart anything
Two lamps and a mechanical timer.
Put a 2700K or 2200K bulb in a lamp on one side of the room, a bright 4000K to 5000K bulb in a different fixture, and a plug-in timer on the warm one. In the morning you use the bright fixture. From early evening the timer runs the warm lamp and you stop using the ceiling light.
This costs the price of two bulbs and a timer and gets you most of the benefit, because the two things that matter most are warmth and dimness in the evening, and both are achievable with a dumb bulb and a decision.
Its limit is that it cannot dim gradually and it cannot adapt. For a lot of people that is fine. If you try nothing else from this article, try this for two weeks before spending anything.
Route 2: tunable white bulbs, app only, no hub
Wi-Fi or Bluetooth bulbs that connect straight to your router or phone. Cheapest route into real scheduling.
What you get: per-bulb schedules, colour temperature control, remote access, usually voice assistant support.
What goes wrong: every bulb is a Wi-Fi client, so a dozen of them crowd your network. Most depend on the manufacturer's cloud, so an outage or a discontinued app takes your schedule with it. Bluetooth-only bulbs need your phone in range, which defeats automation. Response is slower and less reliable than Zigbee.
Reasonable for three or four bulbs in one room. Painful beyond that.
Route 3: Zigbee bulbs and your own coordinator
This is what I run, and it is the route that genuinely replaces Hue rather than approximating it.
Zigbee is an open low-power mesh standard. The Hue bridge is a Zigbee coordinator. You can buy your own coordinator for a fraction of the price, and Hue bulbs themselves are standard Zigbee devices that will pair with it, which means you can mix brands freely and buy whatever is cheapest per bulb.
The stack:
A coordinator. A USB stick that talks Zigbee. The Zigbee2MQTT documentation recommends three chipset families: Texas Instruments zStack, Silicon Labs EmberZNet, and Dresden Elektronik deCONZ. It specifically advises a locally connected USB or UART adapter over a Wi-Fi connected one for stability, and notes that only zStack and EmberZNet adapters currently support coordinator backup. Take the backup point seriously. Re-pairing twenty bulbs is an evening you will not enjoy.
Something to run the software on. A Raspberry Pi, a mini PC, an always-on machine you already have, or a NAS. Anything that stays on.
The software. Zigbee2MQTT lists support for 5,798 devices from 608 vendors at the time of writing, which is the practical reason this route works: almost any bulb you find on sale is already supported. Home Assistant's built-in ZHA integration is the lower-effort alternative if you do not want to run MQTT.
Bulbs. Any Zigbee tunable white bulb. Buy one first, test it at 5% brightness in a dark room, and only then buy the rest.
What you get: everything runs locally, nothing phones home, the schedule survives any company going out of business, and you are never locked to one brand again.
What it costs you: an afternoon of setup and a willingness to read documentation. That is the real price, and it is why Hue still sells.
Route 4: dumb bulbs on a smart dimmer
Often overlooked and sometimes the best answer.
A smart wall dimmer controlling good quality dimmable bulbs gives you the smoothest dimming of any option, keeps the wall switch working normally for everyone else in the house, and does not care what bulbs you buy.
The catch is that a dimmer changes brightness, not colour temperature. Most warm LEDs do not warm further as they dim the way an incandescent did, though some are specifically designed to. Look for bulbs sold as "warm glow" or "dim to warm" if you take this route.
It also involves mains wiring, which in many places is a job for an electrician, and some dimmers need a neutral wire your wall may not have.
Choosing a protocol, honestly
| Protocol | Needs a hub | Local control | Good for |
|---|---|---|---|
| Wi-Fi | No | Usually cloud dependent | Three or four bulbs, one room |
| Bluetooth | No | Phone must be in range | A single lamp, not automation |
| Zigbee | Yes, a coordinator | Yes, fully | Ten or more devices, whole home |
| Thread and Matter | A border router | Yes | New builds, mixed ecosystems |
Matter is the standard that is supposed to end this fragmentation, and support is improving every year. If you are buying today and want the least future regret, Zigbee with your own coordinator is the proven route and Matter over Thread is the one to watch.
The schedule, which is the actual product
Hardware is the easy part. This is the configuration that does the work:
Within 30 minutes of waking. Everything to maximum brightness and your coolest setting. Then go outside for ten minutes if you can. A bright overcast morning beats any bulb you can buy, and a dedicated therapy lamp is 10,000 lux against roughly 50,000 for a sunny day outdoors.
Through the working day. Bright and cool wherever you work. Being underlit all day is half the problem and gets almost none of the attention.
Three hours before bed. Begin a slow transition to warm. A gradual ramp over 30 to 60 minutes is far less jarring than a step change, and most software supports a transition time.
The last hour. 2200K or lower, dimmed hard. Bright enough to move around, not quite bright enough to read comfortably.
Overnight. Everything off. If you need to see, a dim red nightlight on a motion sensor is the least disruptive option there is.
Set it once. Never touch it again. That is the whole point.
Mistakes that cost money
Buying a dozen bulbs before testing one. Dimming quality is the thing that varies most between cheap bulbs and the thing you cannot check from a product page. Buy one, run it at its lowest setting, look at it for a minute, then commit.
Smart bulbs behind a dumb switch. If someone flips the wall switch, the bulb loses power and your schedule stops. Either replace the switch with a smart one or put a cover on it. This catches almost everyone once.
Paying for colour you will not use. Colour bulbs cost more than tunable white. For a circadian setup you need warm to cool, not purple.
Choosing a cloud-only ecosystem for a schedule you want to keep for a decade. Companies discontinue apps. Local control is not paranoia, it is maintenance planning.
Over-automating. Start with one room and one schedule. Every additional automation is another thing that can break at 7am.
What is still worth paying Philips for
I am not going to pretend the cheap route is strictly better.
If you want it to work without thought, if you want one app that handles bulbs, strips, outdoor lights and switches, if you value smooth low-end dimming above all else, or if the idea of troubleshooting a Zigbee mesh fills you with dread, buy Hue. It is good, and your time has value.
The argument of this article is narrower: the biological benefit comes from the schedule, not the brand. If the price is what has stopped you from setting this up, it should not. Two bulbs and a timer will get you most of the way, and a Zigbee coordinator will get you the rest.
Sources
- Zigbee2MQTT documentation, adapter selection and coordinator backup guidance.
- Brown TM, et al. Recommendations for daytime, evening, and nighttime indoor light exposure. PLOS Biology 2022;20(3):e3001571. Source of the lux targets.
Wiring work on mains circuits should be done by a qualified electrician where your local rules require it.