Picture this.
It's 2:47 a.m. Your baby is crying. You swing your legs off the bed, reach for the light, and in the half-second it takes your thumb to find the switch, you already know what's coming.
That flood of white light hits your eyes like a camera flash.
Your pupils clamp shut. Your brain snaps to full alert. And somewhere behind your eyeballs, a tiny chain of biological events kicks off that will keep you awake for the next hour and a half.
But here's the thing most people never think about. It wasn't the baby that stole your sleep. It was the light you switched on to deal with the baby.
And it wasn't just any property of that light. It was a specific set of wavelengths, measured in nanometres, that your brain is wired to treat as a wake-up signal.
Which raises a question: if you could swap those wavelengths for ones your brain mostly ignores, could you get back to sleep in minutes instead of lying there staring at the ceiling?
That's what 670nm is about. And the difference between that number and a generic "red" LED is bigger than most people realise.
Your eyes have a hidden light sensor (and it doesn't care about vision)
Your retina contains specialised cells called intrinsically photosensitive retinal ganglion cells, or ipRGCs for short. They don't help you see. They don't contribute to your night vision. They have one job: to tell your brain's internal clock whether it's daytime or night-time.
These cells contain a photopigment called melanopsin. And melanopsin has a very specific sensitivity curve. It responds most powerfully to short-wavelength light around 490nm, which sits squarely in the blue part of the visible spectrum.
This is not a theory. The CIE (the International Commission on Illumination, the body that literally defines how light is measured worldwide) published a standard in 2018 called CIE S 026. It maps out exactly how sensitive melanopsin is at every wavelength across the visible spectrum.
And the numbers tell a fascinating story.
Not all red light is created equal
Most night lights marketed as "red" use LEDs somewhere between 620nm and 640nm. And for a long time, people assumed that was good enough. It looks red, it's not white, job done.
But when you look at the CIE melanopic action spectrum, the difference between 630nm and 670nm is not small. It's dramatic.
At 630nm, the melanopic weighting factor is 4.18 x 10-4.
At 670nm, it drops to 1.96 x 10-5.
That means 670nm carries roughly 21 times less melanopic weight than 630nm. Both values sit on the modelled long-wavelength tail of the melanopic sensitivity curve. But 670nm sits much further along it.
In practical terms, at 670nm you could have roughly 20 times more light in the room before reaching the same melanopic dose as a 630nm red LED. More light for the same biological cost. That's the headroom advantage, and it's meaningful when you actually need to see what you're doing at 3 a.m.
Why "red" on the box is not enough
Here's where it gets interesting. Walk into any supermarket and pick up a night light labelled "red." Look at the spec sheet (if there even is one). You'll often find LEDs rated at 620nm or 625nm. Some don't list a wavelength at all.
And some cheap red LEDs are not truly narrowband. They can have a broad emission spectrum that bleeds into orange and even yellow-green wavelengths, which carry significantly more melanopic weight.
Compare that to the blue spike in a standard white LED, which peaks around 450nm. On the CIE melanopic curve, 450nm has a weighting of 0.554. That is roughly 28,000 times more melanopically stimulating than 670nm.
So the hierarchy is clear:
- White/blue-enriched light (450nm peak): extremely high melanopic stimulation
- Generic "red" LEDs (620-630nm): very low melanopic stimulation, but still measurable
- 670nm red: the lowest practical melanopic stimulation you can get while still producing visible light
The number on the spec sheet is what counts. Not the colour your eyes perceive.
What does the research say?
A 2022 consensus paper published in PLoS Biology brought together 17 international experts in light and circadian science. Their recommendation: keep evening lighting below 10 melanopic lux, and keep the sleep environment below 1 melanopic lux (Brown TM et al., 2022).
That threshold is defined in melanopic units, not regular lux. It's the melanopsin-weighted illuminance that matters, not the raw brightness. A 670nm light can be bright enough to see your baby's face clearly and still register a vanishingly small melanopic signal.
Separately, Figueiro and colleagues compared white light and red light (630nm) during night-time exposure in a controlled lab study. White light significantly suppressed melatonin. Red light at 630nm did not (Figueiro MG et al., 2015).
If 630nm already preserved melatonin in that study, 670nm, with roughly 21 times less melanopic weight on the CIE standard curve, sits even further from melanopsin's sensitivity range.
The practical difference: headroom
This is where 670nm stops being an academic detail and starts being useful in the real world.
You're not trying to sit in darkness. You need to see the bottle, find the dummy, check the nappy, avoid the creaky floorboard on the landing. You need functional light.
The melanopic headroom at 670nm means you can have a brighter, more usable light without nudging your brain toward its daytime state. You're not squinting under a dim red glow. You're working in a genuinely useful light that happens to fall in the one narrow part of the spectrum your circadian system mostly overlooks.
That's the whole point. Not dimmer light. Smarter light.
What we built at Driftback
We designed two lights around genuine 670nm LEDs. Not 620. Not 630. Verified 670nm.
The Driftback Sensor Light (£23.99) is a wall plug-in with an automatic motion sensor. It switches on when you step into the hallway and off when you leave. No fumbling for switches in the dark, no waking anyone else up.
The Driftback Bedside Lamp (£34.99) sits on a bedside table or nursery shelf with a simple on/off switch. It gives you enough 670nm light to handle a night feed, a nappy change, or a bedtime story without flooding the room with wavelengths your brain reads as morning.
Both are designed in the UK and built around the same principle: the wavelength on the spec sheet is the thing that matters, and 670nm gives you the most headroom on the melanopic curve.
You can see both at driftback.uk.
Disclaimer
This article is for general information only and is not medical advice. Driftback products are consumer lighting products, not medical devices. They are not intended to diagnose, treat, cure or prevent any disease or medical condition. If you have persistent sleep problems, please speak to your GP.
References
- Brown TM, Brainard GC, Cajochen C, et al. Recommendations for daytime, evening, and nighttime indoor light exposure to best support physiology, sleep, and wakefulness in healthy adults. PLoS Biology. 2022;20(3):e3001571. PMID: 35298459.
- Figueiro MG, Sahin L, Wood B, Plitnick B. Light at night and measures of alertness and performance: implications for shift workers. Biological Research for Nursing. 2016;18(1):90-100. PMID: 25697165.
- CIE S 026:2018. CIE System for Metrology of Optical Radiation for ipRGC-Influenced Responses to Light. International Commission on Illumination.