Yep, thats back to my point. LED's, dispite their intensity being proportional to current, obviously have a 'minimum' voltage as well. VVV. you may well be right about the maximum rating voltage, but I've done a few little measurements, and the V-drop across LED's is pretty much fixed, depending only on the colour/material of the die.
I agree that the voltage specified in datasheets isn't spot-on, I measured several blue LED's, and they varied from 3.8 - 4.2v. That said, each individual LED gave a constant voltage drop throughout the test.
I still don't understand why everyone seems to suggest 'parallel' configuration, when the topic is quite clearly about 'series' connection. I agree that I would nearly always use parallel for powering multiple LED's, but of course they can be connected in series! After all, I'm sure we've seen a few silicone diodes connected in series for a specific voltage drop of multiples of 0.6. (ie: 4 diodes = 2.4v drop). Also, some LED arrays used in LCD backlights are configured in both series and parallel, although it must be said, those are generally all the SAME colour/material, and therefore have the same voltage drop.
Again, why not do some tests? Its all very well in theory, we could even start talking about energy bands, band gap etc.. but nothing is beter than good 'ol experimentation. I stand by my old post, make sure you have a PSU that can put out more voltage than the sum of the LEDs, then work out a series resistor.
I'm not going to talk about 'intensity' since that is subjective. Our eyes sensitivity to light depends on the wavelength (colour) and LED's are available in various viewing angles which determine its so called 'luminous intensity' (mcd). In series, multiple colours could be very hard to match brightness, try to get ones that match intensity vs current, since each LED will get the same current.
Buriedcode :
Theoretically..you could connect a 40v supply to an LED...if the max current
available was low enough, say 10ma.
Notice how I started with *theoretically*
The maximum current that could flow through the LED would be 10ma. That was my point, it would still probably send the LED to 'Gallium Aluminum Arsenide heaven'.
Post was a tad too long, sorry.
BuriedCode