Multiwavelength LED Light Source

Needed a light source for a fluorescence microscope that I am working on so decided to build one from high power LEDs. The final device has 16 different selectable wavelengths, from UV (395nm) to InfraRed B (~2000 nm). LED power being adjustable from 0.1W to 3W. It is simple to make and can be used for many applications where you need different wavelengths.

The last diagram shows the general concept. The LEDs would be held in a rotatable wheel so that a specific LED could be selected by rotating an LED of the required wavelength across a positive and negative constant current power supply. The rotating LED wheel would be held in place by a spring loaded clamp or ratchet. The bases of the LEDs would be attached to a heatsink and the whole assembly would rotate around a bolt.

High power LEDS x 16 different wavelengths

LM317 power supply chip to create a constant current power supply

Thin plywood

6 way switch

Power switch

DC-In power socket

Various resistors

Long brass bolt and nut

Metal spring

Bronze strip or other springl metal strip

The first part to make is to make the LED wheel. This will hold the 16 power LEDs in place.

The LEDs were purchased from ebay. It was not possible to buy these as an assortment so purchased 10 of each color or wavelength. I managed to get the following wavelengths: 395 nm, 400 nm, 405 nm, 410 nm, 425 nm, 440 nm, 460 nm, 490 nm, 500 nm, 520 nm, 590 nm, 595 nm, 620 nm, 650 nm, 850 nm, 940 nm.

A diagram for the placement of the LEDs was prepared, printed on paper and glued onto a MDF disc from Pergo laminate flooring. This disc was leftover after cutting a hole in a piece of PERGO flooring with a hole saw for another project. Holes were drilled into the marked region with a Dremel and then the holes were expanded to the correct size with a step drill bit. I applied a rubbery glue to the rim of the LEDs and than attached each LED onto the hole with one set of leads, the positive leads, pointing inwards. The inner leads were soldered together with bare copper wires. It tested each LED with a constant current power supply to make sure that they were still working.

The base of the LEDs on the LED wheel were glued to a circular heatsink with epoxy.

The project languished for a bit till I bought a K40 Laser Cutter and brought that online. The laser cutter allowed me to cut the parts out of plywood much more easily and precisely. I cut out a thumbwheel that would attach to the LED wheel and help to rotate the LED wheel to a desired position. The thumbwheel had semicircular slots around the rim that would engage with a spring loaded ratchet and would lock the wheel in place. I had also cut out a front and back panel and the parts for the ratchet from the 5 mm thick plywood. Applied a coat of black paint to the wooden pieces.

You can see the different parts before assembly: The back panel to which I had test fitted the power switch, a rotary switch and its knob. The label for the rotary switch had to cover the original label as I had messed up the values and also cut a larger hole than needed in the back panel. A long brass bolt, the LM317 chip, the assembled ratchet after gluing the separate ratchet components together, the LED wheel attached to the heatsink, and a bronze strip that I was initially planning on using but ended up not using.

The final image is for the label that will be glued onto the thumbwheel switch.

I attached the swiveling spring-loaded ratchet to the front panel. Then placed the brass bolt through the hole in the front panel and put a washer.

I had screwed the thumbwheel, with its glued on label, to the LED wheel + heatsink assembly. The label is displaced about 90 degrees to the LED.

The negative leads for each LED were soldered to small brass screws that were screwed into the LED wheel. The brass screw heads would make contact with a little bronze metal strip on the front panel as the LED wheel was rotated into place.

The spring-loaded ratchet would now hold the wheel in place. To rotate to another LED the ratchet would be released by pulling the lever down allowing the thumbwheel to rotate and then be locked in place again with the spring loaded ratchet.

The LED wheel assembly was slid onto the brass bolt. The metal contact for the screw heads was attached onto a piece of plywood which was then attached with screws to the front panel. A washer was placed on the heat sink and the wire that connected the positive leads was soldered to this washer. The brass bolt was held in place with a brass nut.

I was constantly checking for continuity with a multimeter while doing this assembly.

To light up the LEDs I used a constant current circuit that I had used for my 4-channel adjustable current source. The diagram from bristolwatch.com explains this circuit well. I calculated the resistor values for the different current outputs, from 10 mA to 300 mA, in excel. I also calculated the required power rating the resistors, and the amount of waste heat generated in the LM317 at different voltage inputs. The LM317 could generate about 5W of waste heat so a heatsink was needed.

The circuit diagram is shown with DC power coming through a DC socket. A diode is placed in series to prevent reverse voltage and to drop the input voltage a bit. A power switch controls the power that is fed to the LM317 input pin. The common positive LED leads are connected to the adjust pin of the LM317. The current adjusting resistors are selected by the 2 x 6 way rotary switch connected between the adjust and output pin of the LM317.

The last image shows the resistors soldered to the rotary switch.

The DC input socket, the power switch, the LM317 and heatsink and the rotary switch are attached to the back panel. An 1N4001 diode is soldered to the positive lead of the DC input socket and then to the LM317 input pin.

The LM317 output pin is connected to a brass clip that clips around the brass bolt forming an electrical connection while allowing the brass bolt to rotate without losing the electrical connection. This clip was taken from a British AC plug. The output pin is also connected to the 1st common terminal of the rotary switch. The central adjust pin of the LM317 is connected to the 2nd common terminal of the rotary switch with its bank of resistors.

A wire from the power switch with negative voltage is connected to a ring terminal (in red). This wire will be connected to the wire from the metal strip that contacts the brass screw heads on the LED thumbwheel.

The front panel is slid into place with the brass bolt going through the friction clip on the back panel and resting in the hole in the back panel.

The black wire from the springy strip that contact the small brass screw heads in connected to the blue wire from the power switch that carries the negative voltage.

Kept testing for unexpected shorts and to make sure that the connection were correct with a multimeter.

I had cut and engraved the top and bottom panels and the two side panels from 3mm plywood on the laser cutter. These panels were used to support the front and back panels and to assemble the whole box by edge gluing the side panels to the back panel and to the top and bottom panels.

The front panel was kept as a removable panel and was held in place with the same brass screws that were used in the LED thumbwheel,

Turned out to be beautiful piece of equipment with a retro look and a high level of ease of use. Just flip the wheel for a different wavelength and use the rotary switch to a different power setting.

I still have to make adapters that will attach to the front panel which connected the LED light output to fiberoptic light guides, light filters or to microscopes.

Check out the video for a bit more detail.

Thanks