current measurement - LED Parallel Forward Voltage Higher Than Indvidual LEDs Vf

I measured the forward voltage of two CoB LEDs at 600mA then wired them in parallel driven by 1200mA.

Why is the parallel V_f higher than the individual forward voltages?

Forward Voltages:

Only LED1 @ 600mA      33.4V
Only LED2 @ 600mA      34.2V
Parallel @ 1200mA      34.9V 

Test Circuit

Driver is Constant Current Mean Well 60 Watt, 54 Volt, 1200mA.

---

Two measurements were made with 1200ma flowing through the paralleled pair of LEDs.

LEDs were mounted to a large copper heatsink. Measurements were made immediately before LEDs got hot.

Measurement 1. As shown the current measured through the shunt resistor is for both LEDs.

Measurement 2. When the switch is flipped, it measures only the current through LED1 yet current is still supplied to both.

The current through LED2 = measurement 1 - measurement 2

The voltage drop across the shunt (0.010V) is negligible when compared to the V_f of about 35V.

Results:
I_BOTH = 1231 mA
I_LED1 = 768mA
I_LED2 = 463mA (BOTH MINUS 1)

---

Why I ask

On a cannabis forum the stoners do not know how electricity works. They have a sub-forum just for LED grow lighting which includes building DIY fixtures.

I design LED grow lights. On this forum its the blind leading the blind. It is my position that CoBs and strings of LEDs should never be driven in parallel with a CC driver with out load balancing. When I state this I get a lot of opposition from uninformed members.

To prove my point I ran this test. I was surprised that the parallel forward voltage was higher than the forward voltage of both LEDs when they were separately driven alone.

This is not something that needs to be fixed. I just cannot explain why the parallel forward voltage would be higher than both of the individual forward voltages.

Answer

If you want to measure each individual CoB LED Vf and your CC LED driver module can be adjusted to 690 mA, then I suggest the following:

^{simulate this circuit – Schematic created using CircuitLab}

You can now measure the Vf for the CoB LEDs separately and somewhat more accurately based in current and temperature.

You could even (for a short time) remove R2 and from this get some idea of the slope of the Vf at 600 mA and then 690 mA.

If the voltage difference of your CoB modules are more than 0.5 - 2 V then I'd assume your current balance is going to be off by a large margin.

Final solution

Let's just for an example assume that your modules do come out at the numbers you proposed in the question: LED1 @ 600mA 33.4V LED2 @ 600mA 34.2V

What you need to do is to raise the voltage on the lower voltage module by 34.2 - 33.4 --> 0.8 V

If you place a 1.2 Ohm resistor (Use 1-2 W) in series with LED1 you will approximately balance the current flows through the two modules. This won't accurately track with module temperatures but it should get you within 10% or better current balance.

If your modules come out at different Vf values then you can calculate the balance resistor required.