Short answer, in plain words. The driver is the box that feeds the LEDs. It decides how many watts actually reach the diodes, how long the light lasts, and how much of your wall power turns into wasted heat before it ever makes light. Most North American home growers run on 120 V, and most drivers are a little less efficient there than at 240 V. Read the driver's datasheet, not the light's box: in the founder's experience you often get only about 80% of the advertised watts once you do the math.
The deeper version. Mean Well's HLG-600H runs about 94.7% efficient at 115 V versus 96.5% at 230 V; the XLG-320 about 92% versus 95% [1][2]. Drivers also derate at low input voltage and high temperature, and constant-current or constant-power drivers only deliver their full rating inside a set voltage or current window [1][2]. The brand comparison below is from the makers' own datasheets.
Part of the series: LED grow lights, properly. 1. LED vs HPS: the heat and power-bill myth · 2. Samsung LM301B vs LM301H, bins and fake chips · 3. LED drivers explained · 4. How much grow light your space needs
Why are advertised watts not real watts?
Founder's field experience: advertised watts and real delivered power often differ; you typically get about 80% of the claim once you do the math or read the driver datasheet's fine print [3]. Common forms: a model number that sounds like a wattage ("1000"), wattage quoted as "HPS equivalent", or the sum of the diodes' maximum ratings rather than what the driver delivers [3]. We found no public dataset measuring this across fixtures; the worked example below shows how to check your own.
How do you read an LED driver datasheet?
The driver turns wall power into the DC power the diodes run on, and its datasheet tells you how much power the diodes can actually get. Read five things: the rated power and how the driver regulates (constant current, constant voltage or constant power), efficiency at your input voltage, the static characteristic (input-voltage derating), load versus temperature, and life expectancy. Mean Well's HLG and XLG series are good examples [1][2][4].
Constant current, constant voltage, constant power
- Constant current (CC): the driver holds the current steady; power = string voltage × current. On an HLG-600H-48, the CC region runs 24–48 V at a rated 12.5 A [1]. If the diode string sits at 44 V, the diodes get 44 × 12.5 = 550 W, not 600 W.
- Constant voltage (CV): the driver holds voltage steady, and the board's own circuitry sets the current. Common on boards with built-in resistors or secondary drivers [1].
- Constant power: the XLG-320-H delivers its full 312 W only when the current sits in its "full power" window, 5.57–7.42 A, which means a string voltage of about 42–56 V [2]. Below that window the current is capped at 7.42 A, so a 36 V string gets only 36 × 7.42 ≈ 267 W (86%).
Efficiency depends on input voltage
Most North American home growers run their lights on 120 V household circuits, while driver datasheets are measured at 230 VAC unless stated otherwise [1][2][4].
| Driver | Efficiency at 115–120 VAC | Efficiency at 230 VAC | Source |
|---|---|---|---|
| Mean Well HLG-600H-48 (full load, Tcase 80 °C) | about 94.7% (115 V curve) | about 96.5% (curve); 96% typ. in table | [1] |
| Mean Well XLG-320-L (full load, Tcase 85 °C) | about 92% (120 V curve) | about 95% (curve); 93.5–94.5% typ. in table | [2] |
| Mean Well HLG-480H | not tabled; see curve | 94–95% typ. | [4] |
Curve values are read from the published graphs and are approximate.
The difference is modest on a good driver: a 600 W HLG-600H load draws about 634 W at 115 V versus 622 W at 230 V, about 12 W (41 BTU/h) more heat [5][1]. Budget drivers with no published curve may do worse (founder's field experience)being reviewed [3].
Read the derating curves
- Static characteristic (input voltage): the HLG-600H keeps 100% load from 115 V up, but derates toward 80% at 90 V [1]. The XLG-320 keeps 100% from 120 V and falls toward 80% at 100 V [2]. A sagging 120 V circuit (long extension cords, a loaded circuit) can push a driver into its derating zone.
- Load vs ambient temperature: on 120 VAC the XLG-320 holds full load only to about 35 °C (95 °F) ambient, versus about 45 °C (113 °F) on 230 VAC [2]. A driver mounted inside a hot tent may be running past its full-power rating.
- Life expectancy: stated at a case temperature, for example above 62,000 h at a Tcase of about 75 °C or less for the HLG-600H and HLG-480H [1][4], and above 50,000 h at 75 °C or less for the XLG-320 [2].
- Input voltageBelow ~115 V: down to 80% at 90 V
- Ambient temperatureXLG-320 full load to ~35 °C on 120 V
- Output windowConstant power only between 42 and 56 V
- Case temperatureLife rated at ~75 °C or less
- Watts to the diodes
Mean Well vs Inventronics vs Sosen vs no-name drivers
The three brands below publish full datasheets, so you can check them. Figures are typical values from each maker's datasheet for the model named; sizes differ, so compare like with like. Supported (manufacturer datasheets).
| Mean Well HLG-600H (600 W) | Mean Well XLG-320 (315 W) | Inventronics EUD-600SxxxDV (600 W) | Sosen SS-200EP-56B (200 W) | Generic no-name | |
|---|---|---|---|---|---|
| Efficiency at 115–120 VAC | ~94.7% (48 V model, curve) [1] | ~92% (curve) [2] | 91–93% typ. by model [6] | 88% [7] | Usually unpublished |
| Efficiency at 220–240 VAC | ~96.5% (curve); 92–96% typ. by model [1] | ~95% (curve); 92.5–94.5% typ. [2] | 93–95% typ. by model [6] | 91% [7] | Usually unpublished |
| Low-voltage derating | 100% from 115 V; 80% at 90 V [1] | 100% from 120 V; 80% at 100 V [2] | 100% from 115 V; 80% at 90 V [6] | Derating curve referenced [7] | Unknown |
| Rated life | >62,000 h at Tcase about 75 °C or less [1] | >50,000 h at Tcase 75 °C or less [2] | 90,000 h at 70 °C case (220 V, 80% load); about 120,000 h at 65 °C or less on the curve [6] | 62,000 h at Tc 72 °C [7] | Unknown |
| Warranty | 7 years for HLG made from 2015 on [8] | 5 years in reseller listingsbeing reviewed | 7 years, at the warranty case temperature [6] | 5 years [7] | Often short or none (founder's field experience) [3] |
| Surge immunity | 4 kV line–earth, 2 kV line–line [1] | 6 kV line–earth, 4 kV line–line [2] | 10 kV common mode, 6 kV differential [6] | 10 kV CM, 6 kV DM [7] | Unknown |
| Dimming | 0–10 V, PWM or resistance (B type) [1] | 0–10 V, PWM, resistance (AB type) [2] | 0–10 V, PWM, timer modes; programmable current [6] | 1–10 V, PWM, resistor [7] | Often a crude potentiometer |
| Safety listing and IP | UL 8750 (type HL), CSA C22.2 No. 250.13; IP65/IP67 [1] | UL 8750 (type HL), CSA; IP67 [2] | This variant lists ENEC, CE, CB and others; check the North American variant for UL [6] | UL/cUL 8750; IP rating per model [7] | Often none, or a printed logo with no file number |
Fair pros and cons.
- Mean Well: highest published efficiency of the group, easy-to-find datasheets with curves at 115, 230 and 277 V, UL "type HL" listings for damp locations. But its HLG surge rating is lower on paper than Inventronics or Sosen, and the life figure is quoted at 75 °C, so it can't be compared directly with a figure quoted at 70 °C.
- Inventronics: high surge immunity, a long published life curve, a 7-year warranty and programmable current. But 120 V efficiency is a little lower than the Mean Well HLG, and you must check that your exact suffix is the UL-listed variant.
- Sosen: high surge immunity, UL 8750, a 5-year warranty and a 62,000 h rating. But the datasheet we could find covers a 200 W model with lower published efficiency (88% at 120 V); we found no public Sosen 600 W datasheet to compare.
- No-name drivers: cheap, and sometimes fine. But there's no datasheet, so there's no way to know efficiency at 120 V, the life at your temperature, surge protection or listing (founder's field experience) [3].
How drivers usually die. Heat ages the electrolytic capacitors, roughly halving life for every 10 °C hotter [9]; all four makers rate life against case temperature for that reason [1][2][6][7]. Power surges kill drivers with low surge immunity. Water gets into drivers without an adequate IP rating. Drivers run near or past their derating limits (low input voltage, hot tent) fail early. A driver hung outside the tent runs cooler and keeps its heat out of the canopy.
Worked example: where does "650 W" go?
Here is the founder's "about 80%" in numbers. The two builds below are illustrative, not tests of any product.
| Step | Better build | Budget build | How to find it |
|---|---|---|---|
| Advertised power | 650 W | 650 W | Listing |
| Driver rating | 600 W (HLG-600H-48) | 600 W (unbranded) | Driver label and datasheet |
| String voltage × driver current | 44 V × 12.5 A = 550 W | 41.6 V × 12.5 A = 520 W | Board spec, or measure DC volts and amps |
| Driver efficiency at 120 V | ~94.7% [1] | ~89% (assumed; founder's field experience) [3] | Efficiency vs load curve |
| Wall watts (diode watts ÷ efficiency) | 550 ÷ 0.947 ≈ 581 W | 520 ÷ 0.89 ≈ 584 W | Plug-in power meter |
| Power to diodes as % of claim | 550 ÷ 650 ≈ 85% | 520 ÷ 650 = 80% | |
| Heat into room | 581 × 3.412 ≈ 1,982 BTU/h | 584 × 3.412 ≈ 1,993 BTU/h | [5] |
Same wall watts, same heat, and nearly the same power to the diodes. What differs is what the diodes do with it: bin, drive current, temperature and authenticity decide how many µmol/s come out.
- 650 WAdvertised
- 550 W (85%)Better build: to the diodes
- 520 W (80%)Budget build: to the diodes
- ~584 WAt the wall, both builds
Source: Worked example above; HLG-600H efficiency [1]; budget driver is founder's field experience [3].
Plain-words version: the number on the box is a wish. The number on your watt meter is the truth.
Common mistakes
- Trusting the model name ("1000") instead of the driver label and a watt meter.
- Reading the 230 V efficiency when you run on 120 V [1][2].
- Mounting the driver inside a hot tent, past its full-load ambient rating [2].
- Running on a long 120 V extension cord that sags the input voltage into the derating zone [1].
- Buying a no-name driver with no datasheet, so nothing about it can be checked [3].
