-
First, Which Grower Are You?
-
Scenario A: Seedlings and Clones (18/6, Low Intensity)
-
Scenario B: Vegetative Photoperiods (18/6, But You Have Options)
-
Scenario C: Flowering Photoperiods (12/12, No Exceptions)
-
Scenario D: Autoflowers (Where "More Light" Stops Working)
-
How to Figure Out Which Scenario You're In
-
The Short Version (And Why I Still Use a Mechanical Timer)
I've been running indoor grow operations since 2016 — mostly small commercial setups in the 8-24 light range. In that time I've fried roughly 200 seedlings (not all at once, which is somehow worse), replaced three controllers, and — according to my old electricity bills — wasted somewhere close to $4,200 running lights longer than they needed to be on. So when someone asks me how long should I leave a grow light on my plants, I don't give them a number. I ask which of four situations they're in.
Because there is no single answer. Anyone who tells you there is has either never grown more than one type of plant, or is selling something. The right schedule depends on plant stage, plant type, and how much money you want to spend on electricity. Here are the four scenarios I've personally screwed up, in order of how bad the mistake was.
First, Which Grower Are You?
Quick self-check before the details:
- Plants are still in their first 2-3 weeks from seed or clone — Scenario A.
- Photoperiod plants in vegetative growth, no flowers yet — Scenario B.
- Photoperiod plants that have started flowering on their own or on a flipped schedule — Scenario C.
- You bought seeds labeled "auto" or "autoflowering" — Scenario D.
Read the one you belong to, or read all four — but don't mix schedules. That's how I killed my second batch in 2019.
Scenario A: Seedlings and Clones (18/6, Low Intensity)
Seedlings are fragile in a way that isn't obvious until you've watched a tray of 48 healthy-looking clones turn into pale, crispy little toothpicks over a single weekend. In March 2019, I put a fresh tray of rooted clones under full intensity on a 24-hour schedule. I figured more light = more growth. Three days later, they looked like they'd been in a toaster oven. Roughly $320 in clone stock — gone.
The answer for seedlings and clones is 18 hours on, 6 hours off, at 25-40% intensity. If you're just running a small veg tent, the Spider Farmer G1500 EVO at that reduced setting is honestly more light than a seedling tray needs — I measured about 180-220 µmol/m²/s at 18 inches with a dimmer set to 30%, which is right where you want to be. Some growers do run 24/0 at very low PPFD and it works. But if you don't have a PAR meter (and most beginners don't), do not gamble on 24/0. That six-hour dark window isn't strictly required for photosynthesis, but the plants seem to recover better and stem elongation is more predictable.
One thing I had to unlearn: a grow light is not a spotlight LED that you flip on when a room looks dim, and it is not downlight lighting. It's a scheduled, measured input. Treating it like a lamp — "looks dark, turn it on" — is exactly the mindset that killed my first two runs.
Scenario B: Vegetative Photoperiods (18/6, But You Have Options)
This is the one area with genuine flexibility. 18/6 is the standard and it works. But 16/8 also works, and slightly better than people admit. In summer 2022, when our utility rate jumped to $0.24/kWh during peak windows, I dropped two veg tents from 18/6 to 16/8 to save money. The growth was maybe 8-10% slower over a 4-week veg — but I saved roughly $180/month across 12 lights.
Honestly, I'm not sure that 8-10% number is fully reliable. The two tents I compared weren't perfectly matched (one had a zipper leak I didn't fix until later), so the comparison wasn't clean. What I'm confident about: 16/8 works, 18/6 works slightly better, and 20/4 hits diminishing returns fast. Pick based on your electricity rate, not on grow-forum bravado.
On the SE4500 320W specifically — I bought two in early 2024 and metered them. At 100%, each pulls around 4.9 kWh/day on an 18/6 schedule, which lines up with the 320W driver rating after accounting for efficiency losses. If you use a smart controller like the GGS, map the schedule in the app and verify with a kill-a-watt for the first week. My GGS fell out of Zigbee sync once (August 2024) and left lights on for 19 hours. Everything survived because the plants were in late veg, but it could have been worse.
Scenario C: Flowering Photoperiods (12/12, No Exceptions)
12 hours on, 12 hours off. Every single day. I mean that literally. In September 2022 a cheap mechanical timer failed on me and one tent got 14 hours of light for four consecutive days. By week five of flower, three of the six plants in that tent had hermaphrodited. I lost about 30% of the harvestable yield — roughly $700 in lost revenue, plus $120 in replacement hardware. That's a roughly $820 lesson in why you buy a second timer as a backup, not a nicer primary timer.
If you're asking "how long should I leave a grow light on" and your plants are flowering, the answer isn't just 12 hours. It's exactly 12 hours, and you need to verify that with a timer you didn't get for free in a grow kit.
Also: dark means dark. If you've got a hallway light bleeding through a zipper or a small LED indicator light on a power strip glowing inside the tent, you're chipping away at that yield without knowing it. I use black electrical tape on every LED indicator in my flower tents. Cheap insurance.
Scenario D: Autoflowers (Where "More Light" Stops Working)
This is the counterintuitive one. Common advice says autoflowers can run 24/0 — and they can. They won't die. But here's what I learned after running both schedules side by side for a full cycle in 2024: 24/0 and 20/4 produced basically identical yields, but 24/0 nearly doubled electricity cost and definitely caused more leaf canoeing and tacoing, especially in the last three weeks of flower.
The numbers said 24/0 → more photons → more photosynthesis → more yield. My gut said the difference was within measurement noise. I ran 8 plants per tent (4 White Widow autos, 4 Northern Lights autos, identical media and feed) and got 402g dry from the 20/4 tent and 411g from the 24/0 tent. That's roughly 2%, which — given my scale and trimming variability — is essentially a tie. I went with my gut, dropped to 20/4, and my power bill thanked me.
There's a caveat that matters more than the hour count: total DLI. If you're running high-PPFD lights like the SE4500 at full power, 24/0 will push some autoflower cultivars past 45 mol/m²/day of daily light integral, which is where light stress — not growth — becomes the limiting factor. I measured PPFD with an Apogee MQ-500 (calibrated in March 2024), and at 18 inches dead center the SE4500 was producing around 1,050 µmol/m²/s. Multiply that out over 24 hours and you're way past what most autos want. 20/4 or even 18/6 keeps you in the safe zone without sacrificing yield.
How to Figure Out Which Scenario You're In
If after reading the above you're still not sure:
- Check your seed packet first. If it says "auto," you're Scenario D. No exceptions.
- Look at the plant. First true leaves = Scenario A. Multiple nodes, no flowers = Scenario B (if photoperiod). Visible flower sites = Scenario C.
- Check your light's dimmer setting. Whatever mode you're in, if your lights are at 100% and dimmable, you're probably overdosing the plants in A and D, and probably fine in B and C.
- Get a cheap kill-a-watt meter. $20. Real money well spent. You'll find your "320W" light pulls 340W at the wall, or your "150W" pulls 145W. Those numbers matter when you're mapping schedules.
The Short Version (And Why I Still Use a Mechanical Timer)
I have mixed feelings about smart controllers. On one hand, being able to see my tents from my phone while traveling literally saved a full crop in October 2024 when I noticed a light had stayed off for 6 hours. On the other hand, they're another failure point — and they fail quietly. My GGS controller is honestly a game-changer for schedule management, but I still keep a $15 mechanical timer wired in parallel on every flowering tent as a backup, because I learned the hard way what one bad timer costs.
Bottom line: match the schedule to the plant's stage, verify your light's real output before you trust the label, and stop assuming more hours equals more growth. The plants don't care what you paid for the fixture. They care whether the schedule matches what their biology expects — and whether the light is actually reaching them at a usable intensity.
There's something genuinely satisfying about a harvest where every variable was dialed in on purpose. After two years of guessing, that first calm, predictable run — that's the payoff that keeps me doing this.
