UV and far-red are the two additions that cannabis grow light buyers ask about most and understand least. Both sit at the edge of the photosynthetically active range, both are sold with confident claims, and both behave in a way that a simple wattage comparison cannot capture. A UV bar is not a stronger light. A far-red channel is not free yield.
The useful way to think about either one is as a dose with a window. UV is a stress signal that a plant answers by building secondary compounds, and the same dose that triggers that answer can damage tissue if it runs too long or lands on a young canopy. Far-red is a shape and efficiency signal, and the same photons that improve light penetration in a dense canopy can stretch a plant in a sparse one. Neither replaces PAR, and neither fixes a room that is short of light.
This guide covers what the research supports, where the marketing outruns it, and how to specify either channel so that it can actually be used. It is written for cannabis growers and for buyers comparing fixtures where UV or far-red is an option. The short version: far-red is a conditional yes, worth about 5 to 10 percent of output where the canopy can use it. UV is a default no, worth testing only where the room can measure an endpoint. Add either one after PPFD, DLI, CO2 and climate are correct, specify them on independent channels, and measure the dose at the canopy rather than trusting a fixture label.
Most lighting decisions are about quantity: more photons, more even coverage, more efficient conversion. UV and far-red are different. They are signals, and a signal is judged by when it arrives, how strong it is and how long it lasts, not by how many watts the fixture draws.
UV is a stressor. Plants sense ultraviolet through specific photoreceptors and respond by thickening leaf cuticles, producing flavonoids and anthocyanins, and in cannabis altering the profile of secondary metabolites. That response is a defence, and defences cost energy. A dose below the threshold produces little response; a dose above it damages tissue and reduces photosynthesis. The useful range is narrower than for white light, which is why a UV channel needs a dimmer and a schedule rather than an on switch.
Far-red works through phytochrome, the same photoreceptor system that measures day length. It has two distinct effects that are often conflated. The first is photosynthetic: when red and far-red arrive together, the combined quantum yield is higher than the sum of the parts, because the two photosystems are driven in balance. The second is morphological: a high proportion of far-red relative to red signals shade, and a plant answers by elongating stems, raising leaves and, in some crops, flowering earlier. One effect is an efficiency gain. The other is a shape change that can be helpful or harmful depending on canopy density.
So the answer to "is UV worth it" and "what does far-red do" is not a yes or a no. It is a dose, a timing and a canopy condition. The rest of this guide gives those three for each.
Ultraviolet is not one thing. It is three bands with different penetration, different biological effects and different safety requirements, and a supplier who says "UV" without naming the band has told you very little.
| Band | Wavelength | What it does to a plant | Use in horticulture |
|---|---|---|---|
| UV-C | 100 to 280 nm | Damages DNA and RNA directly, sterilises surfaces | Water and surface disinfection, never as a growth light |
| UV-B | 280 to 315 nm | Strong stress response, flavonoid and anthocyanin production, DNA damage above a threshold | Research and specialised finishing, small doses only |
| UV-A | 315 to 400 nm | Milder stress response, pigment formation, some stomatal and morphological effects | The common band in supplemental UV fixtures |
| Violet | 400 to 430 nm | Photosynthetically active, affects phototropism and morphology | Part of a normal full spectrum, not a UV treatment |
The band matters because penetration differs. UV-C is absorbed within micrometres of a surface, which is why it sterilises and why it never reaches a lower leaf. UV-B penetrates a little further and is the band most associated with secondary metabolite research, but it is also the band with the narrowest safe window. UV-A penetrates deeper into leaf tissue, produces a milder response and is far easier to apply evenly, which is why most commercial UV channels sit in the UV-A range, often around 365 to 405 nm.
A common specification problem follows from this. A fixture described as "UV" may be emitting at 365 nm, which is UV-A, and a buyer who expected the UV-B effects read about in a study has bought something different. Ask for the peak wavelength and the band, in nanometres, and check that the answer is consistent with the effect being claimed.
The strongest controlled work on UV in a flowering crop compared UV-supplemented plants against an untreated control under matched conditions and measured plant chemistry at harvest. The result was not a null result in the strict sense. Leaf-level measurements moved: flavonoid concentration, leaf thickness and some pigment readings responded. What did not move was the outcome a buyer pays for. There was no commercially relevant difference in the harvested chemistry between treated and untreated material.
That gap is the whole argument, and it is worth stating precisely. A plant has UV receptors and it answers them. The question for a commercial room is not whether a response exists, but whether the response reaches an endpoint someone will measure and invoice. On current evidence, in a flowering room, it usually does not.
The trial designs also do not transfer to a catalogue fixture. One influential study used a single narrow UV-B band at 287 nm over 3.5 hours a day at five doses. Another used a broadband source with a UV-A to UV-B ratio near 99 to 1, delivered at 1.81 W per square metre for 18 hours a day. A third used broadband UV-A for 10 days before harvest. These are three different quantities in three different bands, and none matches a mixed UV-A bar run for a few hours in the last two weeks.
What this means for a buyer is a default, not a ban. UV is inert for nobody: it produces a real stress response. But a room that adds it without a defined endpoint is buying a response it cannot invoice. The default should be no UV premium, and the exception should be a measured programme with a control zone. A grower selling on appearance and terpene profile has a more defensible case than one selling commodity flower on weight, but even that case needs a measurement, not a spectrum chart.
UV is the part of the spectrum where a fixture label tells you least. LED UV emitters vary in wavelength, in radiant efficiency and in how quickly they degrade, and the number that matters is the irradiance at the canopy, not the wattage of the bar.
Two units appear in the literature and in supplier documents. Irradiance in micromoles per square metre per second describes the instantaneous photon flux at a wavelength, and it is the number a grower can measure with a UV sensor. Dose in kilojoules per square metre per day describes the accumulated energy, and it is the number used in horticultural research, particularly for UV-B. A dose is irradiance multiplied by time, adjusted for photon energy at that wavelength, which is why the same micromole count at 300 nm and at 380 nm is not the same dose in joules.
| Approach | Canopy level | Duration | Accumulated dose |
|---|---|---|---|
| UV-A trial, low | 2 to 5 micromoles per square metre per second | 2 to 4 hours, final two to three weeks | 9 to 24 kJ per square metre per day |
| UV-A trial, upper | 5 to 15 micromoles per square metre per second | 3 to 6 hours, final two to three weeks | 17 to 106 kJ per square metre per day |
| UV-B research range | Below 1 micromole per square metre per second, weighted | Short pulses, instrumented | Reported as weighted UV-BBE only |
| Continuous UV all cycle | Any level | Full photoperiod, vegetative and flower | Not recommended: stress without an endpoint |
Three qualifications belong beside that table. The doses assume UV-A near 365 nm and are unweighted, so they cannot be compared with a weighted UV-B dose without the action spectrum and the normalization wavelength. The upper row lands near the broadband dose used in the most quoted 2024 trial, which ran 1.81 W per square metre for 18 hours and accumulated about 117 kJ per square metre per day. And no standards body or extension service publishes a cross-crop, cross-band recommendation. Any figure you adopt is a planning range to be verified on your own hardware, not a setting to copy from a supplier page.
The figures above replace a much higher range that circulates informally. Ten to forty micromoles at the canopy sounds moderate, but at 365 nm it corresponds to roughly 3 to 13 W per square metre, which is two to seven times the broadband irradiance in the trial that produced the clearest published dose. Dose scales with hours as well as intensity, and a bar run at 40 micromoles for six hours accumulates more than 280 kJ per square metre per day, well past any published horticultural work.
The figures are planning ranges drawn from greenhouse and controlled environment practice, not a prescription for a specific cultivar. The important part is the shape: low to moderate irradiance, limited hours, late in the cycle. A UV bar run at full output for twelve hours from the first week of flower is not a stronger version of this plan, it is a different and more damaging one.
Measurement is the weak point in most installations. UV sensors are expensive and rarely owned by growers, so the dose is usually unknown. A supplier who can provide a measured irradiance grid at a stated height, for a stated dimmer setting, is telling you something useful. A supplier who quotes only the electrical wattage of the UV bar is not.
UV is not a purchase decision so much as an experimental one. It has three gates, and a room that cannot pass all three should keep the money in the base layer.
The first gate is a measured endpoint. Before ordering a channel, name the number you will judge it by: flavonoid or anthocyanin concentration, leaf thickness, internode length, flowering date, or a lab panel your buyer already pays for. If the endpoint is "it looks better", there is nothing to measure and no way to know whether the premium bought anything. A supplier who cannot state the measurement height and method cannot support the test at all.
The second gate is a base layer that is already correct. If the flower canopy runs at 800 to 1,000 PPFD with CO2 supplementation, the plant has the carbon and the light energy to answer a moderate UV dose. If the same room runs at 500 PPFD without CO2, the UV dose competes for resources the plant does not have, and the result is usually smaller flowers rather than a richer profile.
The third gate is a control zone and a full batch. UV effects are small enough that they disappear into room-to-room variation unless one zone is left untreated and both are harvested and measured. Running UV across a whole room and comparing it with last cycle's memory is not a test, it is an anecdote with a spectrum attached.
Where all three gates pass, the right move is one zone, one defined endpoint, and a dose log at canopy height. Where one fails, the money is better spent on the base layer: more uniform PPFD, better VPD control or a tighter light schedule. What UV should not be is a line item added to every quote because the fixture platform offers it.
There are several patterns that look like a UV feature and are not one, and they are worth recognising before a quote is compared.
The first is a low power UV bar that produces no measurable dose at the canopy. A few watts of UV spread over a wide area, hung at the height of a normal fixture, delivers irradiance below the response threshold. The fixture is real, the effect is not. Ask for the canopy irradiance at the working height and for the measurement method.
The second is a wavelength that does not match the claim. A 365 nm emitter is UV-A, and it is often sold as "UV" in a way that implies the UV-B research. It may still be useful, but the buyer should know which band they are getting and why. The choice inside UV-A is also a trade: 395 nm sits closest to the visible edge and produces the least UV-specific response, while 365 nm delivers higher photon energy at a higher cost per photon. Nothing in the dose literature supports paying for the shorter wavelength without a measured endpoint.
The third is a UV channel that cannot be controlled separately. If UV shares a channel with the main white output, it runs whenever the room runs, including vegetative and early flower. That is the pattern most likely to reduce yield, and it turns a finishing tool into a permanent stress. Independent channels are the minimum specification.
The fourth is missing lifetime data. UV LEDs degrade faster than white and red emitters, and a bar that has lost half its output after a year is delivering half the planned dose. Ask for the lifetime figure at a stated output percentage, and ask what the replacement interval is. A fixture that does not mention UV emitter lifetime is a fixture designed for a catalogue photo rather than a production room.
A UV channel is priced like a spectrum option and costs like a cooling load. That gap is missing from most quotes, and in a small room with a tight cooling budget it can exceed the price of the channel itself.
Take a 100 W UV-B bar with a wall-plug efficiency of 3 to 4 percent. Three to four watts leave as UV photons; the other 96 to 97 W become sensible heat inside the envelope. That is 100 multiplied by 3.412, or about 341 BTU per hour, before any UV photon is absorbed by a leaf or a wall and converted to heat as well. From the room's point of view the fixture is a 100 W heater that happens to emit a little UV. It buys no PAR photons at all.
Compare 100 W of 660 nm at a fixture-level 3.0 micromoles per joule. It buys about 300 micromoles per second of PAR, and a share of that energy ends up fixed as biomass rather than rejected as heat. The UV bar buys a stress signal and a cooling problem. That is not an argument against UV in every case, but the cost line is lamp plus chiller plus dehumidification, and a proposal that prices only the lamp has priced less than half the decision.
The same logic explains why a "cost per PAR photon" comparison for UV is meaningless. There are no PAR photons to divide by. The right comparison is cost per measured endpoint, which is why the endpoint gate comes before the purchase order.
Far-red is the band from 700 to 800 nm, with the horticulturally active part between roughly 700 and 750 nm. The effect fades above about 752 nm, which is why the 850 nm emitters sometimes sold as IR are near-infrared, not far-red, and not a substitute. Far-red sits outside the 400 to 700 nm window that PPFD counts, which is why a fixture can add far-red without changing its PPFD number. That single fact explains most of the confusion in the market.
The first mechanism is photosynthetic. In the 1940s and 1950s, Emerson observed that adding far-red to red light increased the quantum yield of photosynthesis beyond what either wavelength produced alone, because the two photosystems work in series and a red-only source overdrives one of them. The active window is narrow: roughly 700 to 732 nm, with no enhancement above about 752 nm. Modern white LEDs already contain enough far-red to avoid severe imbalance, but a spectrum with a small far-red addition can still improve the efficiency of the red peak. This is a real effect, it is measurable across a whole canopy, and it depends on the background spectrum, which is why it is common in high performance fixtures and absent from some daylight-like rooms.
The second mechanism is morphological. Phytochrome measures the ratio of red to far-red light, and a low ratio signals that the plant is under a canopy. The response is a set of shade avoidance traits: longer internodes, smaller and more upright leaves, earlier flowering in some crops, and less branching. In a dense canopy this response can be useful, because upright leaves let light deeper into the profile. In a sparse canopy the same response produces a taller, thinner plant with less flower sites.
A third use, end-of-day far-red, applies a short far-red treatment at the end of the photoperiod to set the phytochrome state going into darkness. It is a real technique in controlled environment research and it is sensitive to timing and to the dark period that follows. The sign of the effect is variety specific: in short day crops, a longer far-red treatment under a 12 hour photoperiod has delayed flowering, while a shorter one after a 10 hour photoperiod showed no adverse effect. For a cannabis room the practical advice is conservative: treat end-of-day far-red as an experiment on one cultivar, not a standard setting, and never let far-red light spill into the dark period.
If a supplier talks about far-red without mentioning the red to far-red ratio, the conversation is incomplete. The ratio, usually written R:FR, is what a plant actually reads, and the same far-red wattage produces a different response depending on how much red is present.
Two different numbers get called the ratio, and mixing them is a common specification error. R:FR is a photon ratio calculated from the spectrum. The phytochrome photostationary state, Pfr/P, is the fraction of the receptor in its active form: about 0.80 under saturated 660 nm, 0.67 to 0.73 under daylight, about 0.1 under pure far-red, and about 0.025 under 718 nm. The two move together but they are not interchangeable, and a specification that treats either as a constant will be wrong by enough to matter.
| Source | Typical R:FR | What a plant reads |
|---|---|---|
| Direct summer sunlight | About 1.1 to 1.2 | Open sky, no shade signal |
| High pressure sodium | About 0.9 to 1.1 | Broad, warm, with substantial far-red |
| White LED, 3000 to 4000 K | About 3 to 5 | Red heavy, shade signal very low |
| White LED with 5 percent far-red | About 1.5 to 2.5 | Closer to daylight, mild shade signal |
| White LED with 15 percent far-red | Below 1 | Strong shade signal, stretching likely in a sparse canopy |
The figures are approximate and depend on how the ratio is defined and measured, but the direction is reliable. A white LED without far-red is a much redder source than daylight, which is one reason a room can look bright and still produce compact, stacked growth. Adding a small far-red share brings the ratio toward daylight and can improve canopy light distribution. Adding a large share pushes the ratio below daylight and starts to imitate shade.
For cannabis, a far-red share of roughly 5 to 10 percent of combined output is the usual commercial target. The published ceiling is about 30 percent of 400 to 750 nm flux, which equals roughly 43 percent of PAR-only flux, and the same authors suggest staying near 20 percent for most crops because stem elongation becomes limiting before photosynthesis does. Watch the denominator: 30 percent of 400 to 750 nm is not 30 percent of PAR, and reading it as a share of PAR misjudges the headroom by about a third. The exact number should come from the fixture's spectrum report, not from a marketing claim, and it should be quoted together with the R:FR ratio it produces.
Far-red is often sold as a yield channel, and the claim needs unpacking. What far-red can do is improve the efficiency with which a canopy uses the light it already receives, mainly by balancing the two photosystems and by changing leaf angle so that light penetrates further. In a dense canopy with a high leaf area index, that can translate into more photosynthesis per unit of installed PAR.
What far-red does not do is add PAR. A fixture that reports 1,000 micromoles with far-red included and 850 without has not become a stronger light in the sense that PPFD measures. It has changed the spectrum and possibly the distribution. A supplier who counts far-red photons inside a PPFD figure is either using a broad definition or stretching the number, and a buyer should ask which.
The honest commercial case for far-red is strongest in three situations: a dense canopy where light penetration limits yield, a tall crop where lower leaves matter, and a vertical or multi tier system where every photon is expensive to deliver. In a single layer room with a moderate canopy and generous PPFD, the gain is smaller and harder to measure, and the money may be better spent on uniformity.
There is also a scheduling cost. Far-red at the wrong time stretches plants, and far-red spilled into the dark period can disrupt flowering. A far-red channel needs its own timer and its own place in the recipe, which is a control requirement rather than a spectrum requirement.
The efficiency metric a buyer is taught to compare, PPE, counts photons from 400 to 700 nm and divides by electrical input. Far-red photons land in the denominator and not in the numerator, so adding them always lowers the ratio, even when the fixture delivers more usable light. That is not a trick by the supplier. It is a definition, and the definition was left unchanged on purpose: the DLC declined to fold far-red into PAR in 2021 because the effect depends on the background spectrum.
The arithmetic is worth running once, because it turns an argument into a decision. Start with 1,500 micromoles per second from 400 to 700 nm at a PPE of 3.0, which is 500 W of input. Add 150 micromoles per second at 730 nm at a panel-level 3.39 micromoles per joule, which adds 44.2 W. Power rises 8.8 percent, flux from 400 to 750 nm rises 10 percent, and the far-red share is 9.1 percent.
| Metric | Before | After | Change |
|---|---|---|---|
| PPF, 400 to 700 nm | 1,500 micromoles per second | 1,500 micromoles per second | 0 |
| PPF, 400 to 750 nm | 1,500 micromoles per second | 1,650 micromoles per second | +10 percent |
| Electrical power | 500 W | 544.2 W | +8.8 percent |
| PPE, 400 to 700 nm | 3.00 micromoles per joule | 2.76 micromoles per joule | minus 8.1 percent |
| ePAR efficacy, 400 to 750 nm | 3.00 micromoles per joule | 3.03 micromoles per joule | +1.1 percent |
| Far-red share of 400 to 750 nm | 0 | 9.1 percent | not applicable |
The same fixture now has three true descriptions. It is 8.1 percent worse, 1.1 percent better, or 10 percent more light, depending on which number reaches the slide. All three are correct, and a quote that shows only one is incomplete. Three rules follow. Compare PPE only against PPE at a similar far-red share, otherwise use ePAR efficacy. Require PPE, ePAR efficacy, far-red share and the 700 to 800 nm PPF in every quote. And cost the morphology, because a stretched canopy needs more height, and more height changes the distribution at the bench.
Both UV and far-red are only useful if they can be scheduled independently of the main output. This is the specification that separates a real feature from a checkbox, and it is usually decided at the controller rather than at the fixture.
A three channel fixture, with white, UV and far-red on separate dimming inputs, allows a recipe that changes over the cycle: white alone in vegetative, white plus far-red in early flower, white plus UV in the final weeks, and no UV at all during the dark period. A single channel fixture with a fixed spectrum allows one recipe for the whole cycle, which is the wrong recipe for at least part of it.
The controller has to support the same structure. It needs independent dimming per channel, a schedule that can change by week, and a hard cut-off that guarantees no UV or far-red output during the dark period. On a greenhouse or a room with a curtain, the cut-off should be interlocked with the curtain, so that a stuck relay cannot leave a channel on overnight.
Ask for the channel count and the dimming protocol before comparing prices. A 0 to 10 V input per channel is the minimum, DALI or a proprietary digital bus is better for recipes that change often, and a fixture that only offers a single dimming input cannot deliver a staged UV or far-red plan no matter what its spectrum report shows.
UV is the one part of a grow light that can hurt a person, and the safety side belongs in the buying conversation rather than in a manual nobody reads.
UV-A at the irradiance used in finishing is a hazard to eyes and skin at close range, and the limit is closer than most people expect. ICNIRP publishes an eight hour effective exposure limit of 30 joules per square metre for 180 to 400 nm, which corresponds to an average effective irradiance of about 0.001 W per square metre across a shift. The working form is t_max equals 30 divided by E_eff, in seconds. At an effective irradiance of 0.01 W per square metre, the permitted exposure is 3,000 seconds, or 50 minutes. A UV bar that looks modest on a datasheet can put a maintenance worker past the limit inside one shift, so the calculation belongs in the commissioning file.
That is why interlocks, shielding and rated eyewear are engineering controls rather than accessories. A room with a UV channel should be entered only with the channel off, or with eye protection rated for the band. Fixtures should have a visible indicator that shows when the UV channel is energised, and the control system should make it easy to kill that channel before anyone opens the door. A small status light on the fixture is not enough if the room is large, and a door lockout is cheaper than a training programme.
Installation height matters for the same reason. A UV bar hung low over a work aisle exposes a person at head height, while the same bar over a canopy is directed at leaves. Keep UV fixtures over the crop and out of the walkway, and do not use them as general room lighting.
Documentation should include the wavelength, the maximum irradiance at a stated distance and the safety instructions for the band. A supplier who cannot state the peak wavelength in nanometres is not in a position to advise on safe use. Claims about certification should be checked against the listing for the exact model, because a certificate for a similar fixture does not cover the one being quoted.
Two more costs belong in the same file. Materials age faster than datasheets admit: polycarbonate lenses yellow, acrylic covers haze, cable jackets crack and reflector coatings chalk, so ask for the UV exposure rating of every polymer in the optical path, not only the LED board. And UV damage can look like a nutrient deficiency, so run the differential before changing the feed programme. A room running UV at the wrong height produces symptoms that get treated with fertiliser for two weeks while the real cause stays switched on.
The questions below separate a channel that can be used from a channel that can only be described. They apply to UV and far-red alike, and the answers should be in writing.
| Question | Good answer | Warning sign |
|---|---|---|
| What is the peak wavelength? | A number in nanometres, with the band named | "UV" or "full spectrum" without a wavelength |
| What irradiance reaches the canopy? | A measured grid at a stated height and dimmer setting | Only the electrical wattage of the bar |
| Is the channel independent? | Its own dimming input, timer and cut-off | Shared with white output |
| How long do the emitters last? | Lifetime at a stated percentage of initial output | No lifetime figure for UV or far-red |
| What is the far-red share and R:FR? | Percentage of combined output plus the resulting ratio | Far-red watts quoted as extra PPFD |
| Are PPE and ePAR both reported? | PPE at 400 to 700 nm with the test standard named, ePAR efficacy at 400 to 750 nm, and the far-red share | One efficiency number with no wavelength range |
| What is the safety documentation? | Wavelength, maximum irradiance, eye and skin guidance | No documentation for the band |
The sixth question is the one buyers skip and later regret. UV is not a hazard that can be managed by a label on a box, and a room where the UV channel can be switched on while someone is working under it is a room that will eventually hurt someone. The control and the documentation are part of the product.
Both channels are marginal tools, and a room with a weak base layer will not be rescued by either. Two or three percent of potential quality or efficiency does not compensate for a canopy that is 30 percent short of light, an uneven PPFD map or a VPD that swings at night.
The order of investment is worth stating plainly. First, get PPFD to the target with even coverage. Second, get the photoperiod and the dark period right. Third, get CO2 and climate into range. Fourth, improve uniformity and efficacy. Only then do UV and far-red earn a place, because only then does the plant have the resources to answer them. A buyer who adds UV before fixing the base layer has bought a signal that the plant cannot afford to hear.
The table below is a starting recipe for a room that already runs a high PPFD flower cycle. It assumes independent channels and a controller that can change settings by week.
| Stage | UV channel | Far-red share | Purpose |
|---|---|---|---|
| Propagation | Off | 0 to 3 percent, or off | Root and leaf area, no added stress |
| Vegetative | Off | 3 to 5 percent | Canopy development without stretching |
| Flower, weeks 1 to 3 | Off | 5 to 10 percent | Light penetration as the canopy closes |
| Flower, weeks 4 to 6 | 2 to 5 micromoles, 2 to 4 hours | 5 to 10 percent | Early secondary metabolism, monitored |
| Flower, final 2 to 3 weeks | 5 to 15 micromoles, 3 to 6 hours | 5 to 10 percent | Finishing, with leaf condition checked weekly |
| Final week | Off or reduced | Off | Let the plant finish without added stress |
The numbers are ranges, and the direction of adjustment is what matters. If leaves show bleaching, cupping or a drop in colour at the UV dose being used, reduce the hours before reducing the intensity, because the response is dose dependent and time is the easier variable to control. If plants stretch after far-red is added, raise the red share or lower the far-red percentage.
Change one variable at a time. A room that adds UV, far-red and a new feed schedule in the same week cannot tell which change produced the result, and the usual outcome is that all three get blamed or all three get kept without evidence. Run one change for a full cycle, keep a control bay if the room allows it, and judge on the metrics the market pays for.
The controlled trials have not shown a commercially relevant increase in the harvested chemistry of a flowering crop. Leaf-level traits move, including flavonoid concentration and leaf thickness, but the whole-plant outcome a buyer pays for has not moved in the work published so far. Some trials report small increases, others report none or a decrease when the dose reduced photosynthesis. No supplier can promise a percentage, and a room should default to no UV premium unless it has a measured endpoint and a control zone.
Yes, and most rooms should. Plants complete a full cycle under 400 to 700 nm light, and a room that is short of PPFD, uneven, or without CO2 has more to gain from fixing those first. UV is worth testing only where a measured endpoint, a correct base layer and a control zone all exist. Skipping it costs a possible quality edge; adding it to a weak room costs yield and cooling capacity.
Most commercial UV channels sit in the UV-A range, around 365 to 405 nm, because UV-A penetrates leaf tissue, produces a milder response and is easier to apply evenly. Inside that range it is a trade: 395 nm is closest to the visible edge and produces the least UV-specific response, while 365 nm delivers more photon energy at a higher cost per photon. UV-B around 280 to 315 nm produces a stronger response but has a much narrower safe window and needs instrumentation. Ask for the peak wavelength in nanometres and match it to the effect being claimed.
Two to six hours late in the photoperiod, in the final two to three weeks, at roughly 2 to 15 micromoles per square metre per second of unweighted UV-A at the canopy, which accumulates about 9 to 106 kJ per square metre per day at 365 nm. Start at the low end and lengthen the hours before raising the intensity. Running UV for the whole photoperiod from the first week of flower is a stress programme, not a finishing programme.
It can be, with caveats. Far-red improves the balance between the two photosystems and can increase light penetration in a dense canopy, which supports flower development. It also signals shade, so too much of it stretches stems and can reduce flower sites in a sparse canopy. A share of about 3 to 8 percent of combined output is the usual commercial range.
It can improve the efficiency with which a canopy uses the light it receives, mainly in dense canopies and tall crops. It does not add PAR, and a fixture that counts far-red photons inside its PPFD figure is stretching the number. In a single layer room with even coverage and adequate PPFD, the gain is small and hard to measure.
Daylight is about 1.1 to 1.2. A white LED without far-red can run 3 to 5, which is much redder than daylight. Adding 5 percent far-red brings it to roughly 1.5 to 2.5, and 15 percent can push it below 1, which starts to imitate shade. For planning, a 5 to 10 percent share is the usual commercial target, 20 percent is a practical ceiling for most crops, and 30 percent of 400 to 750 nm flux is the published ceiling, which equals about 43 percent of PAR-only flux. Ask for the ratio the fixture produces, not just the far-red percentage.
Only if the fixture has spare channels and the controller can schedule them. A retrofit bar with its own driver and timer is possible, but it must be interlocked so it cannot run during the dark period. Adding UV to a fixture that shares one dimming input means UV runs whenever the room runs, which is the pattern most likely to reduce yield.
UV-A at finishing irradiance is a hazard to eyes and skin at close range. Enter the room with the UV channel off or with rated eye protection, keep UV fixtures over the crop rather than over walkways, and fit an indicator that shows when the channel is energised. Ask for the peak wavelength and the maximum irradiance at a stated distance.
PPFD and DLI are defined over 400 to 700 nm, so UV below 400 nm and far-red above 700 nm fall outside both. That is why a fixture can add either channel without changing its PPFD number. The photons still affect the plant, but they are not part of the PAR budget and should not be counted as if they were.
No. The dark period is when flowering is triggered and when the plant runs processes that light interrupts. Any UV or far-red output during darkness is a light leak with extra risk, and it can delay or disrupt flowering. The controller should cut both channels at the start of the dark period, interlocked with any curtain.
PPE divides photons from 400 to 700 nm by the electrical input. Far-red photons add to the input and nothing to the numerator, so the ratio falls even when the fixture delivers more usable light. In a worked example, adding 150 micromoles of 730 nm to a 1,500 micromole fixture raises power 8.8 percent, raises 400 to 750 nm flux 10 percent, drops PPE 8.1 percent and raises ePAR efficacy 1.1 percent. All three statements are true, and a quote that shows only one is incomplete.
Send the room dimensions, PPFD at the canopy, CO2 setpoint, cultivar and the stage you plan to finish, plus your controller protocol. We answer with the channel plan, the canopy irradiance at the working height, a week by week schedule and the safety requirements for the room.
