Down the rabbit hole, and a first discovery

 In the beginning, sometime in 2019, I wanted to do some indoor hydroponics. One thing begat the next thing, and I began to learn things.

The first attempt was to order some Samsung COBs that were designed/engineered for growing. At the time, I did not have a way to cool them properly or to supply the high current required, so they have sat unused. But an examination of them prompted me to look around at what LED devices were available, in what quantities and at what prices. If I wanted to roll my own, I did not have any intention of tossing lots of money at the problem, I just wanted to find a better way to make it happen.

The 2835 white LEDs

One day, while scrolling thru eBay, I came upon a seller who had some reels of 2835 1 watt LED chips (Cree J Series JK2835 18v). After reviewing the datasheet, I ordered a reel. Meanwhile I was looking for a power source. At a local thrift store, I found an HP printer wall-wart that supplied 15.5v and 30v. It was purchased because of the 30v output, but the 15.5v became much more interesting.

Before I had any proper breadboard for 2835 SMD parts, I was very gently soldering leads directly to the LEDs. This allowed me to determine that a 18v LED could generate light from a 15.5v source, and at a much lower current than the datasheet revealed. The light was not obviously as bright as running it at full rated voltage/current, but it was unexpected and promising. After a number of tests, involving limiting the current, I found that these devices would generate light down to as low as 1 mA (which is significantly below the 60 mA rated current from the datasheet). They also ran at 1 mA with no appreciable sign of heat.

Making a few PCBs

The next step was to generate a PCB for a full up test. This required modeling the circuit, current flows and a regulator. The initial attempt was using a single LM317LZ (TO-92 footprint) and 100x 2835 LED devices. That proved optimistic, as the regulator would go into thermal shutdown, because ~120 mA was more heat than it could handle. The second attempt, dividing the PCB into two zones each with it's own LM317LZ and powering 50x 2835 LED devices, was a smashing success. Of the 10 boards received, I have assembled eight of them. Those boards are now powered from Ryobi 18V battery packs. A typical board is drawing 109-112 mA, and will run for 19+ hours on a fully charged 18v 2AHr pack. Total power per board is ~2.2 watts.

Comments

Post a Comment

Popular posts from this blog

Characterizing some LEDs

Image
 Not every LED selling source can supply high quality datasheets. When the prices are too good to ignore, it's time to do your own characterization. What is needed ... • a voltage source - My voltage supply is a Ryobi 18v battery, with an adapter found on eBay for powering Hot Wheels go karts. The adapter has both an on/off switch and an inline fuse. The default fuse is 30A, but I've changed mine to 3A. • a mini board that holds a regulator - These are little PCBs that will hold one adjustable regulator (AMS1117-ADJ or LM317LZ, your choice, they both have a 1.25v Vadj terminal. You will be using this device as a two-terminal constant current regulator circuit. This means that There is a Vin, a Vadj, and a Vout. You have to supply a resistor to program what current value the regulator is targeting. The regulator will then home in on the desired current, while dropping any excess voltage required to do this. • a mini board with positions for 8 SMD resistors - I'm partial to ...
Read more

The basic theory behind light emitting diodes

Image
 This is a Voltage/Current curve of a typical light emitting diode. The exact values of each spot on the curve are less important, but the curve and what happens in each segment, is more important. Segment A - Voltage is increasing, but has not yet reached the point where the diode junction is emitting enough light to be visible. If you have a datasheet for the device, this will probably be the first 75%-80% of the rated forward voltage (Vfwd). Segment B - As the forward voltage enters the knee part of the curve, light may be perceived. Depending on the exact voltage and/or the current, the light may be faint. Once you have entered the curve area, forward current will begin to change as the voltage increases. Depending on the exact device ratings, the device may begin to dissipate some heat. Heat will be low to negligible, increasing as you approach the device rated forward current. Light should be obvious, and increasing as you move deeper into the knee area. Segment C - The light...
Read more

Making the LED knee region larger and more easily usable

 In an earlier blog post, I talked about the knee region. This is the area of an LED's voltage-current curve, where the current begins to increase sharply. For most 3v LEDs, the knee region is a fairly small voltage range. Controlling the knee region requires a very careful adjustment of voltage.  There is a way to expand the knee region: connect multiple LEDs in series !  All the LEDs voltage aspects are cumulative, when connected in series. So the overall vFwd increases by the number in series, but so does the voltage range of the knee region. The basic assumption here is that all the LEDs you are connecting in series have the same data sheet part description, and also they were binned together voltage wise. A single LED might have a knee region that is only 0.1v-0.2v wide. By placing five in series, the voltage rage of the knee region now becomes a range of 0.5v-1.0v wide. An additional advantage of five (or six) in series, is that you can power it with a 15-20 volt su...
Read more