In off-grid systems, it can make sense to use a low voltage DC light bulb, instead of having the overheads of a DC to AC inverter – in terms of capital cost, quiescent (idle) current, conversion efficiency and risk of electrocution. For example, in a shed or other de-attached structure you could have a small 12V or 24V PV Solar system and run your lighting directly from that source.
LED bulbs exist in standard E27 (Edison screw, 27mm) and B22 (Bayonet cap, 22mm) that can be powered from lower voltages – typically 12V, 24V or 48V. We review one such product below.
The Purchase
I purchased five LED bulbs from the Aliexpress YCMY LED Lighting Store in February 2024. The variant I purchased was the 9W E27 12-85V, Cold White.
My motivation for this bulb was the wider voltage range. In addition to 12V use, I was also wanting to operate the bulb from a “48 Volt” LiFePO4 system that could have a fully charged battery as high as 58.4V.
According to the following table sourced from the vendor, the 9W bulb should provide an output of 720 lumens.
The specifications indicate the brand name is Heetech and it includes Epistar LEDs in the 2835 form factor. The power tolerance is 5%.
Testing
The bulb has printed on the base “LED Bulb 12-85V 9W 6000K CE RoHS”, confirming it is in fact a 9W version. The light output seems to vary at little around the 12V mark but is stable at 13V and higher (more on this later).
At 20V, my five bulbs drew the following:
- 350mA @ 20V (7.0 Watts)
- 370mA @ 20V (7.4 Watts)
- 390mA @ 20V (7.8 Watts)
- 400mA @ 20V (8.0 Watts)
- 400mA @ 20V (8.0 Watts)
The listing indicated the bulb was 9W with a power tolerance of 5%, hence providing an acceptable range of 8.55 to 9.45 watts. Reviews for the listing did suggest this might be the case. A few savvy buyers noted lower power ratings than advertised, across the different variants.
LCD Driver PCB
The driver board, pictured below, is based on the AP5218 Step-down constant current LED driver from Shiwei Semiconductor Co, Ltd.
I’m unable to find much on the integrated circuit, outside of the following excerpt:
The device has the following specifications:
- Wide input voltage range: 5 to 100V
- Output current range: 10mA to 1500mA
- Fixed switching frequency: 130KHz
- Spread spectrum for reduced EMI
- Average current mode sampling
- 0 – 100% Duty Cycle Control
- Output short circuit protection; over temperature protection
- Function mode: Full / half brightness
The schematic for the driver looks like this:
The input bridge rectifier is made up of four discrete 1N4007 1A 1000V Silicon Diodes (Marked M7). The voltage drop on the bridge rectifier is more significant when running at lower voltages than would be if operating at mains. When the bulb is operating at 20V (~350mA), the forward voltage drop measured across the diode was approximately 870mV and is consistent with the datasheet. With two diodes in the bridge operational at a time, this means there is a 1.74V drop over the bridge. This dissipates approximately 0.6W.
The 22uF 100V input electrolytic capacitor is branded Axboom and rated for 105degC.
The PCB has a 390mΩ current sense resistor used to regulate current into the LED. I’ve been unable to find a datasheet or method to calculate this value, but placing a 10 ohm load (via a programmable DC load) appears to limit the current at approximately 550 to 560mA.
The AP5218 is powered from a filtered supply consisting of a 5.1k series resistor and 1uF ceramic cap. This VDD node, powering the chip’s logic, appears to sit at about 5.7V, suggesting it is clamped and current limited by the 5.1k resistor.
LED Board
The lamp is fitted with a 35mm diameter LED aluminum backed printed circuit board, labelled LP-08:
The LED board consists of nine 2835 LEDs (2.8mm x 3.5mm) arranged in three banks of three LEDs, with each bank having a 13 ohm series resistor. Each 13 ohm resistor appears to have 2.5V across it, suggesting it is passing 190mA each and dissipating 475mW each. The resistors would have been fitted to help with current sharing, however their values do seem high.
At full current, the voltage applied to this PCB is approximately 11.25V. The driver circuit has a voltage drop due to the forward voltage of the bridge rectifier of 1.84V. Summed together this is 13V and hence why the LED output is stable at 13V and above.
Of the roughly 8W consumed by the bulb, approximately 0.6W is lost in the bridge rectifier, and 1.4W lost in the three LED series resistors. The converter would also contribute some loses.
Fittings
One of the advantages (and disadvantages) of a low voltage bulb using the same base than their 230V/110V mains counterpart is the fittings are common and cheap. (The disadvantage, is the low voltage bulb can be inadvertently plugged into a mains socket).
From a safety perspective, I generally wouldn’t recommend purchasing fittings intended for mains voltages from Aliexpress, however as this installation is extra low voltage (ELV), any lack of rating or certification is less of a problem.
I purchased some E27 bulb holder/sockets from here. If you want some cable strain relief, you can get these. Alternatively the socket has a M10 female thread, and you can purchase hollow plastic riser (tubes) to fit.
I then used some twin core Automotive cable to connect them to Anderson Powerpole connectors. My preference has been to use Powerpole connectors for the lower current devices, rather than bulkier SB50s.
I’ve found it has often been cheaper to purchase a mains lamp, cut the plug off and put an Anderson connector on.
Conclusion
Despite having a power rating less than advertised, the bulbs have been in use for 15 months and are performing satisfactory.
I don’t have the required equipment (integrating sphere) to accurately measure the luminous flux (lumens) to verify the “9W” bulb is actually emitting 720 lumens. I have tried to compare with a bulb of known output, however the comparison is a bit subjective and different colour temperatures doesn’t help.
I would suggest if you have a certain wattage bulb in mind, maybe purchase the next higher wattage one to compensate.
Be the first to comment