LED Strip Light (Desk Lighting)

Dramatic increases in the cost of electricity together with the general desire to minimize electricity consumption, as part of the effort to decrease personal carbon footprint as a contribution to the greenhouse challenge/fighting climate change, lead to investigating the use of LED lighting. Such an objective may sound trite, but if enough individuals would take similar steps, significant change could occur.

The key strength of LED lighting is reduced power consumption. When designed properly, an LED circuit will approach 80% efficiency compared to incandescent bulbs which operate at about 20% efficiency(1). Obviously this means LED lighting is converting significantly more energy into light (which is the purpose of a “light”) rather than producing waste heat.

Cost of electricity is currently 23 cents/kWh, plus 10% GST (i.e. Goods and Services Tax) giving a total of 25.3 cents/kWh. This means operating a 100W incandescent light globe for 8 hours a day, 365 days a year would cost approximately $74/year, of which approximately $59 is for just waste heat into the room, rather than useful light. Using conversion figures from the Carbon Trust(2) this is the equivalent of about 153 tonnes of carbon dioxide, of which approximately 123 tonnes was just for the waste heat. Now, admittedly I do not sit at my work bench for 8 hours a day - although it would be ‘nice’ to have such amount of hobby time - but even an average of 2 or 3 hours per day would be producing in the range of 30-46 tonnes of carbon dioxide a year just for waste heat.

Unlike incandescent bulbs that radiate light in 360 degrees, LED light is focused and directional. This potentially means the already efficiently produced light from an LED can be used to just light the area of interest (for example, the working area on my hobby work bench). Apparently since LED’s do not emit UV rays they tend not to attract insects (bonus!) and unlike fluorescent lamps contain no toxic mercury(3). Because of the low power requirements of LED lighting, this makes them very suited to solar power (and since I had some scrounged 1W LED’s from a discarded emergency light and solar cells from an old calculator, gave rise to the separate LED Garden Light project).

However, the downside is that using LED’s requires a little more than just hooking up a power source with a switch (but, only a ‘little’ more as will be discussed and detailed below). From Wikipedia(4) “The current/voltage characteristic of an LED is similar to other diodes, in that the current is dependent exponentially on the voltage. This means that a small change in voltage can cause a large change in current. If the maximum voltage rating is exceeded by a small amount, the current rating may be exceeded by a large amount, potentially damaging or destroying the LED. The typical solution is to use constant-current power supplies”.

The section in Learning and Component Testing gives details on my various experiments/trials and tribulations with using power supplies for this. The final circuit used in this particular project is discussed below.

The circuit used for the constant current driver for the LED's was sourced from an article by Giorgos Lazaridis (5). With reference to the details in (5) and the components labelled on the diagram in the Schematic Diagram section below, when switch SW1 is closed, voltage as applied to the gate of MOSFET Q2 and current flows through the LED's and the sense resistor (formed by the combination of R2 and RV1).

As the current increases through the LED's and the sense resistor, the voltage drop across the sense resistor also increases. When this voltage reaches ~0.7V transistor Q1 starts to conduct, pulling the MOSFET gate to ground, and hence turning off the MOSFET. Therefore, current through the LED chain is regulated by the value of sense resistor.

To calculate the value of the sense resistor use the following formula:

where:

• V_be = base-emitter voltage transistor Q1 (~0.7V)
• I_led = limiting current through LED (amps from datasheet)

The white LED's used in the project have a maximum continous current of 300mA, therefore, the sense resistor value is 0.7/0.3 = 2.3 ohm. In the constructed circuit, the sense resistor is composed of the total resistance of R1 plus RV1 . The potentiometer RV1 allows varying the resistance and hence the current and therefore the brightness of the LED's.

A value of 1ohm was selected for R2 (which gives the minimum sense resistance even if the potentiometer is wound to zero resistance) which with RV1 at minimum resistance still gave a measured resistance of ~2.5ohm (due to manufacturing tolerance of RV1).

The value of R1 is selected to enable ~1mA of current to flow to the gate of the MOSFET

Power Supply

Each of the 1W 350mA white LED's have a voltage drop of ~2.8V as measured for the particular items used. From the collection of scrounged/saved 'wall warts' from various laptops and other electrical equipment, I had a 'wall wart' that produced nominally 12V DC output at 1.5A from 240V input AC.

The actual measured output of the 'wall wart' was 12.8V and the voltage drop across a string of 4 x white LED's was 11.8V. Therefore, the voltage drop across the MOSFET Vm is:

where:

• V_i = supply voltage from wall wart
• V_leds = voltage drop across the 4 x LED's
• V_rs = voltage drop across transistor Q1 base-emitter ie. 0.7V

Which for the values measured in the actual test circuit are Vm = 12.8V - 11.8V - 0.7V = 0.3V. Therefore, the power disspated in the MOSFET equals (Vm x I_led) which is 0.3V x 0.3A = 90mW. This is well within the MOSFET max power disspation of 65W from the datasheet.

Also, the MOSFET datasheet reports 0.43 W/^oC which gives an expected neglible temperature rise for the MOSFET, and hence no extra heat sink required. As reported in the Testing/Experimental Results section this was exactly what was observed (nice!)

Note: Image loading can be slow depending on server load.

• 

  Silver Membership registration gives access to full resolution schematic diagrams.

  LED-Strip-Light Schematic

LED-Strip-Light Schematic

This project did not require a PCB as such. The construction was done using copper clad prototyping board (veroboard). The red X on the diagram denotes a copper track cut.

Note: Image loading can be slow depending on server load.

• 

  Silver Membership registration gives access to full resolution schematic diagrams.

  LED-Strip-Light Veroboard

LED-Strip-Light Veroboard

• 

  Silver Membership registration gives access to full resolution schematic diagrams.

  LED Heat Sink PCB Layout

LED-Strip-Light Veroboard Layout

The LED strip light is a relatively simple project and no particular difficulties, other than the usual care and attention required when constructing any electronic circuit, should be expected.

The major item of potentially difficulty is associated with making heat sinks and mounting the LED's (see Photographs Section for the approach taken here). Commerically available heat sinks are of course available, and high power LED's can also be purchased ready mounted on heat sinks. However, purchasing the LED 'beads' seperately gives significant cost savings, for example, in quantities as low as a few dozen, in the order of $0.20/each. The downside being that the LED 'beads' are geneally surface mount devices and need a heat sink if operated at higher current.

I have some sheets of copper clad mylar (flexible copper 'PCB' board) which enabled a LED mounting solution (i.e., easy soldering of the surface mount LED's to connection wires) in conjunction with small sections of aluminium bar to act as the actual heat sink. This 'LED mount' was then screwed to a length of perspex to act as the actual overall mounting bracket. The construction method is detailed pictorially in the Photographs Section.

Making the PCB mount and heat sink for the LED's is perhaps overkill for some applications. For example, you could solder connection wires directly onto the LED beads (at least the particular items I sourced had relatively large pads) and then insulate appropriately before attaching to some sort of material to act as a heat sink (perhaps even the mounting surface itself would be sufficient as a heat sink).

However, I planned to use the strip lighting in a variety of locations (e.g. down lights for my bar lighting). Therefore, having a 'neat and tidy' solution that could be removed was essential if 'she who must be obeyed' was to allow installation!

Note: Image loading can be slow depending on server load.

Album 1: Strip Light Circuit Board and Construction

Album 2: Strip Light PCB for LED mounting production

ref001: http://www.lc-led.com/articles/ledlights.html

ref002: http://www.carbontrust.com/home

ref003: http://www.ecoefficiency.com.au/Portals/56/factsheets/foodprocess/ utilities/ecofoodutil_fsu7.pdf

ref004: http://en.wikipedia.org/wiki/Light-emitting_diode

ref005: http://www.pcbheaven.com/userpages/LED_driving_and_controlling_ methods/?topic=worklog&p=3