Portable Current Source
I made this circuit several years ago specifically to test high voltage LED arrays and unmarked zener diodes from the junk bin. I never got to post pics of it until now.
I lost my original CAD drawings so I don't have any to post but I will try to describe how the circuit works. It is basically a two-transistor current source (Fig 6B) but with the supply voltage and current sense resistor R3 switchable with different values. Supply is boosted from 18650 cells to two micro DC-DC converter modules in series with dual supply outputs all in series.
Output compliance voltage is selectable between 12, 24, 36 and 48V. Output current limit is selectable between 100uA, 1mA, 5mA and 20mA. The lower voltages prevents damaging sensitive LEDs when reverse connected. The 100uA current is useful for testing nitride based LED damage from ESD or poor quality chips.
This is the unit. Not the most compact but battery powered and USB charged!
There are a few connections for output.
The female headers are for connecting clips or the SMD tweezer. The 2mm jacks are for plugging in multimeter probes for measuring voltages across the DUT.
The functions are on sticky labels.
Back of the unit.
I used a readily available TP4056 charger board for USB charging and a boost PCB to derive 5V from the 3.6V cells.
clips and multimeter connected. Ready to test some parts.
Maximum output voltage when open circuit.
Here is a nitride based green LED on 100uA.
Multimeter displays the voltage across the LED.
The green LED on 20mA.
Here, we test a 16V zener diode at 20mA.
An example of a damaged LED.
At 100uA, it does not light up and the Vf is low meaning there is parasitic resistance shorting the die.
At about 5mA it lights up but not very bright.
Testing a series diode array from an LCD monitor.
WIth the low currents, we can test for dead chips on COB arrays without blinding myself.
And check the matching of the various parallel strings.
At 20mA, it blends in and gets pretty bright.
Here is a high voltage COB LED. This is at 100uA.
At 20mA, it gets really bright. This LED array is rated at 300mA.
Here is a high voltage LED array.
49V is barely enough to light it up to max brightness.
A 10W LED at 100uA. This is a good quality LED.
After some use, I decided to add stainless mesh to complete the case.
USB charging port.
I need to get some rubber feet to prevent it sliding across the bench.
Most of the circuitry. Rotary switches are under the charge and boost PCBs.
I'm going to sort these LEDs I got from the scrapper and sort the working from the duds.
I don't have a suitable sensor so I took an NPN TO-3 transistor and popped the cap off.
Unfortunately, the one I chose had silicone glue on the die so I had to scrape it carefully without breaking the connections.
The emitter and collector terminals are tied together and is the negative output. The base is the positive output. Connected to a voltmeter, the output voltage is a good indicator of brightness.
Made a lightbox from blank PCBs.
The brass tube is the aperture to the transistor die. The pogo pins on top is for connecting the LED under test.
The output voltage does not correspond to a specific lux or lumen reading but it gives a rough idea on which LEDs are the brightest. Just choose the ones with the highest voltage readings!
output current is not calibrated exact but reasonably stable. Variable output current allows testing LEDs for suitable operating current to achieve the desired brightness. I even used it for testing optocouplers or checking circuit functionality that uses them. I have had it for years and it stays beside my multimeters as I find it useful at times.
24 Oct 2016:
The test jig is too light and moves around a lot. Stuck it on a round piece of SS plate to weigh it down.
Alligator clips are connections to the DMM.
Improved the LED board holding jig.
Close up of the LED board. The LED measures about 5mm square.
LED board is placed face down and slid towards the pogo pins until the tab hits the end stop. At this point, the LED is aimed right at the tube.
Having tested a few hundred now, the new jig makes the work faster and more consistent.
Page created and copyright R.Quan ©21 Oct 2016.