Red Light Therapy Panels: Wavelengths and Irradiance Explained
A red light therapy panel is judged by three numbers you can actually measure — the wavelength it emits, the irradiance it delivers at the distance you sit, and how evenly that irradiance covers the panel face. Wattage, LED count and price are not among them, and a listing that leads with those three is telling you what is easy to count rather than what decides whether the panel does anything.
This page walks through each of the three, what a credible spec sheet states, and the handful of questions that separate measured hardware from a box of diodes.
Wavelength: the bands the research focuses on
The photobiomodulation literature concentrates on two windows of the spectrum. Consumer panels are built around those bands, and the peaks usually specified sit at the top of each window:
| Band | Peak usually specified | Common alternative | Typical depth in tissue |
|---|---|---|---|
| Red | 660 nm | 630–640 nm | a few millimetres |
| Near-infrared (NIR) | 850 nm | 810 nm | a few centimetres |
Both sit inside what is often called the optical window — the range where skin, fat and water absorb and scatter comparatively little light, which is why these two bands and not others keep reappearing in the research. Near-infrared penetrates deeper largely because it scatters more than it is absorbed, while red is deposited closer to the surface.
Red and NIR are not interchangeable, and the literature does not regard them as one thing: panel designs usually pair one red band with one NIR band and put a switch on the NIR array so you can run red alone or both together. A panel offering only one of the two is offering half the spectrum that consumers are looking for, whatever its wattage.
Why peak accuracy matters more than the band label
A diode sold as “red” can sit anywhere in a window a few tens of nanometres wide, and diodes are sorted into bins during manufacture. That is the whole reason a spectral report exists: it tells you where the peak actually landed, not where the marketing says it lands.
Three things are worth asking about directly:
- The measured peak, not the nominal one. The published report should show the spectrum with a wavelength axis. A graphic with no numbers on the axes is decoration.
- Binning tolerance in nanometres. A panel whose diodes drift outside your band across a production run is not the same product on the last unit as on the first. Tight tolerance is an engineering decision and it costs money, which is why cheap panels do not specify it.
- Whether the report is per model or per lot. Per-model data tells you what the design intends. Per-lot data is what tells you what was actually shipped.
One caution: “full spectrum” is a useful phrase for grow lights and a meaningless one for a therapy panel. A therapy specification should name bands and tolerances, because there is no therapy-relevant reason to emit the whole visible spectrum and no way to verify a claim that broad.
Irradiance: the number that has to come with a distance
Irradiance is measured in mW/cm² and describes how much optical power lands on a square centimetre. Two panels with identical diodes can differ by a factor of several here, because irradiance depends on drive current, optics, and how the light is spread rather than concentrated.
The part that catches buyers out is distance. Irradiance falls off steeply as you move away from the panel. For a point source it follows an inverse-square relationship — double the distance and you get a quarter of the irradiance. A panel is an extended source, so its falloff is gentler than that, but it is still steep enough that a number quoted without a distance is not a specification at all.
That is exactly what most listings publish: the brightest point on the face, measured right at the surface, at a distance no one occupies.
For consumer panels, published figures commonly land in a range of roughly 30–100 mW/cm² at the distance the device is used. The practical reading of that range:
- Well below it, you are relying on very long sessions to accumulate a given amount of light, and the device instructions usually will not allow it.
- Inside it, follow the session length the device ships with. The protocol and the panel were designed together.
- Well above it, check that the device’s own protocol accounts for it. High output with instructions copied from a low-output product is a design that has not been thought through.
Every claim on a listing, and what it really tells you
| What the listing says | What it actually tells you | What to ask instead |
|---|---|---|
| “600 W panel” | Usually wall power drawn, sometimes a sum of diode ratings that never run at full current | Optical output, and irradiance at a stated distance |
| “100 LEDs” | A component count | An irradiance map across the whole face |
| “660 nm + 850 nm” | Intended bands | A spectral report showing the measured peak and the binning tolerance |
| “150 mW/cm²” | A number with no reference point | The measurement distance and the instrument used |
| “Low flicker” | A claim about driver behaviour | The driver specification, or a measurement at the stated distance |
None of these rows is a reason to reject a panel on its own. Together they are the difference between a specification and a shopping page.
Coverage: even across the face beats a bright centre
A panel that reads 100 mW/cm² in the middle and 20 at the edges forces you to move during a session, and a routine you cannot keep consistent is not a routine. Uniformity is therefore a specification in its own right, and it is an optical and mechanical design problem — reflector geometry and diode layout — not something a quality check at the end of the line can fix.
This is the same principle growers apply to light over a canopy. In a grow tent the metric is PPFD uniformity, and a fixture rated at high peak PPFD across a coin-sized spot is worse than one rated lower but spread evenly; our PPFD and DLI calculator explains the measurement, and the planted aquarium guide makes the same point about tank lighting. A therapy panel should be bought the same way: ask for an irradiance map at stated distances, and read the edges, not the centre.
Brightness is not a measurement
Here is the trap that catches most buyers: near-infrared light is invisible. A panel emitting at full power in the 850 nm band looks, to your eye, almost dark. A red-only panel looks dramatically brighter than a red-plus-NIR panel of identical optical output.
So visual brightness tells you almost nothing about what the device is emitting, and comparing two panels by eye is a comparison of two different things. The only way to compare is the spectral report and the irradiance map. This is also why a well-made panel ships with protective eyewear: your eye cannot warn you that it is being exposed.
Session length and distance: the two knobs you control
Dose is irradiance multiplied by time. If you move closer, irradiance rises steeply and the session should shorten; if you move further back to cover a wider area, the session has to lengthen to accumulate the same amount of light. Because the relationship is nonlinear, changing distance by a few centimetres changes the dose by more than it feels like it should.
Published session lengths for consumer panels typically run 10–20 minutes, and the instructions that ship with the device are the right starting point rather than a general number from the internet. Two habits worth keeping:
- Start at the low end. Consistency over weeks does more than intensity in one session.
- Fix your distance. A repeatable distance makes every other variable easy to reason about. Most people are better served by marking a spot on the floor than by experimenting session to session.
Safety and positioning: wellness, not medical equipment
Just as important as the engineering is being honest about what the category is.
- A consumer red light panel is a wellness product, not medical equipment. If you have a health condition, ask a healthcare provider before using one — and read any listing that promises otherwise as a warning sign about the seller rather than a feature of the panel.
- Never stare into an emitting panel. Use the eyewear supplied, and remember the NIR point above: the panel may be brighter than it looks.
- Follow the device instructions on session length. More light is not automatically better, and the protocol is part of the product.
- Sensible category framing is about general wellness and routine. Claims that go further are regulatory claims, and they belong to brands that hold the corresponding clearances, not to a spec sheet.
What to demand from any panel supplier
Whether you are buying one panel for yourself or sourcing a container for a brand, the same eight questions do the work:
- The spectral report, with a wavelength axis and the measurement method stated.
- Binning tolerance in nanometres, and whether reports are per lot or per model.
- An irradiance map at stated distances, measured across the full face — not a single peak figure.
- The drive current and whether output is held as the panel heats up.
- The thermal design: heat sinking, fan curve, and the temperature the diodes are specified to hold.
- Whether the driver can be dimmed, and whether dimming is linear and repeatable.
- Which standards the unit was tested to, and by whom.
- A sample unit you can measure yourself, or a report you can have verified.
Anyone who cannot answer the first three is assembling boxes. Anyone who can answer all eight is a manufacturer.
How GoldenBuds builds therapy panels
The engineering behind a therapy panel is the same engineering behind a grow light: wavelength-controlled diodes with a stated binning tolerance, measured irradiance maps rather than marketing graphics, thermal design that keeps output stable across a session, and constant-current drivers that hold that output as the panel warms up. We build red light therapy panels for brands and distributors on that basis, and we publish the same class of data we publish for our grow lights.
Buying for yourself, or sourcing for a brand?
We are the factory, and we say so plainly, because it changes what is useful to you here.
Buying a panel (or a cap) for yourself? The most useful thing we can hand you is the specification to demand from whoever you buy from — and then the routine. Our companion guide covers how long a session should run and how far to sit, including the eye-safety point most people get wrong because near-infrared is invisible. If you are asking whether red light for hair loss is even worth trying, what the evidence actually says about red light caps is the page for that — mechanism, one honest real-world protocol, and the questions to ask before you buy. Browse the therapy light range for what we build.
Sourcing panels for a brand? That is the other conversation, and it starts with photobiomodulation panel manufacturing — how the manufacturing relationship works, what we can and cannot do on compliance, and what we need in order to quote. The red light therapy panel OEM page covers the specification and the data package. Quantities are quoted per project rather than published, because the real answer depends on whether you are rebranding an existing platform or building a new one.
Either way, the test is the same: measured irradiance, a spectral report, and a stated distance. Contact us for a spectral report on any panel.
FAQ
What wavelength should a red light therapy panel use?
The two bands the research focuses on are red at 630–660 nm and near-infrared at 810–850 nm, with 660 nm and 850 nm the peaks most panels specify. What matters at least as much as the nominal number is the measured peak and the binning tolerance, which is why the spectral report is the document to ask for.
What irradiance should a red light therapy panel deliver?
Consumer panels commonly publish figures in the range of roughly 30–100 mW/cm² at the distance the device is used. The figure only means something with that distance attached: irradiance falls off steeply as you move away, so a number measured at the panel surface describes a position no user occupies.
Is 660 nm better than 630 nm?
They sit in the same red window and most panel designs specify 660 nm as the peak because it is the wavelength the bulk of the literature uses. A panel that specifies 630–640 nm has not done anything wrong, but it should show the actual measured peak in a spectral report rather than relying on the label.
Can I judge a panel by how bright it looks?
No. Near-infrared light is invisible, so a panel emitting at full power in the 850 nm band can look almost dark while a red-only panel of identical output looks far brighter. Brightness is also not a measurement of irradiance. The spectral report and the irradiance map are the only way to compare two panels.
Do I need a spectral report to compare panels?
Yes, if you want to compare anything other than price. A report shows the measured peak, the width of the band, and carries the binning tolerance with it, while an irradiance map shows what reaches the user rather than what leaves the diodes. Ours ship with both, per model and per production lot.