My LEDs have a forward voltage range of 3.0-3.8 V, as designated by the packaging they arrived in. When wiring them up, what voltage should I assume they are for calculating voltage drops? I'm wiring 10 parallel rows of 9 series LEDs, and I'm presently assuming that the typical Vf would be the lower value, 3.0V. Is this a reasonable assumption? I could test 10 and take an average, but is that necessary for series? What if I set it up as 3.1V or 3.2V instead?

 

I don't know of any other method than by experiment. That does sound a bit sloppy, having to figure on a 27 - 34.2 volt drop. What is your source voltage?

 

My LEDs have a forward voltage range of 3.0-3.8 V, as designated by the packaging they arrived in. When wiring them up, what voltage should I assume they are for calculating voltage drops? I'm wiring 10 parallel rows of 9 series LEDs, and I'm presently assuming that the typical Vf would be the lower value, 3.0V. Is this a reasonable assumption? I could test 10 and take an average, but is that necessary for series? What if I set it up as 3.1V or 3.2V instead?

Yes, it is better to take the lowest possible Vf to calculate the limiting resistors but dont connect the leds directly in parallel. Each leds has to have its own limiting resistor and then connect them in parallel.

 

Without any other information I would suggest that Vf is normally distributed with a mean of 3.4V and a variance of 0.01. This woul be equivalent to saying that 3.0 is the -3σ point and 3.8V is the +3σ point of the distribution.

The only other reasonable a priori assumption would be a uniform distribution of values in the range 3.0 to 3.8V. The mean is the same at 3.4 volts, but the variance is 4/75 = 0.0533...

With the normal distribution your chances are almost 70% that the value will be in the range [3.3-3.5], whereas they are only 25% with the uniform distribution.

 

You likely have a fair amount of money invested in those LEDs.

It would make good financial sense to test them before committing to a design that may not work as you had hoped, rather than trust to luck and sheer random chance.

Expect a maximum variance of slightly over +-5% on your Vf. If you're running your LEDs at 10% under their rating, this isn't really a problem. But, would you want to do that?

It's up to you.

Oh, you can easily create a decent constant-current source using a few batteries, an LM317 regulator and a resistor in the range of 45 to 51 Ohms (it depends upon the individual regulator; an ideal regulator would need a 48 Ohm resistor)

Connect the output terminal of the LM317 to the ADJ terminal using the 45 to 51 Ohm resistor. Supply the input terminal of the LM317 with a positive voltage (+ terminal of a battery) of 6 or more. Connect your LED in series with an ammeter to the negative voltage (- terminal of the battery). Connect a voltmeter across the LED. You will then know the Vf of the LED at a particular current. If the current is too high, increase the resistance between the OUT and ADJ terminals of the LM317 slightly.

 

My LEDs have a forward voltage range of 3.0-3.8 V, as designated by the packaging they arrived in. When wiring them up, what voltage should I assume they are for calculating voltage drops? I'm wiring 10 parallel rows of 9 series LEDs, and I'm presently assuming that the typical Vf would be the lower value, 3.0V. Is this a reasonable assumption? I could test 10 and take an average, but is that necessary for series? What if I set it up as 3.1V or 3.2V instead?

If you power the LEDs using a current source rather than a voltage source it will
not matter as long as the current source has sufficient compliance.

You could also use a current sink. Not as efficient but very easy to build (FET, opamp,
shunt).

I have a schematic of a load-cell (current-sink) near the bottom of this page ---

http://www.luciani.org/geda/util/matrix/index.html

(* jcl *)