rougesw,
You do indeed have two sets of electronics working at the same time. I have no idea what is happening or how, but whatever IS happening, is happening very well.
Since I am a technological moron, I will try to explain, in those terms, just what I THINK is going on, based on personal experimentation, and discussions with others. If you take a "dumb" head like the L1 and screw it on a "dumb" voltage provider like the E1e body (without the interior regulation of the L1), you are now at the mercy of the Vf (forward voltage requirement) of the "dumb" head. If the "dumb" head requires slightly more than 3v (as the L1-WH does!), you get a dim result when you turn it on. If the "dumb" head requires slightly less than 3v (as the L1-RD does!), you get a bright result when you turn it on. So the L1 RED "dumb" head will go to full bright on an E1e body/3v provider, but the L1-WH will be dim. In either case, you should not be hurting anything since the L1 heads are designed to be driven with 3v. It is just a case of the Vf sensitivity of the various colored heads. I do not know what will happen with the L1-GR or L1-BL heads and I would dearly love to know!!!!!
Flash forward now to the L2 "dumb" head similarly placed on a 6v E2e voltage provider. Same higher Vf requirement for the white LED, so you get the same dim result from the "dumb" head.
So far, all we have gained is a very small, very bright E1e-RD as a backup to my L1-RD if the L1 body or tailcap failed. This is how and why all of this started.
Now enter the regulated-voltage provider in the form of the L1 or L2 bodies to provide constant regulated "low" voltage or constant regulated "high" voltage to whatever you screw on the front of it ("dumb" head or "smart"). By using an E2C or LU-60A adapter, we can screw the KL-3 or KL-5 heads on either the L1 or L2 body to provide REGULATED throughput from REGULATED low/high PROVIDED voltage from the L1 or L2 bodies. Since the KL-3 (gen 1 ok for 3-9v, gen 2 unknown) or the KL-5 (gen 1 ok ONLY for 3-6v, but gen 2 ok for 3-9v) will accept anything from 3-9v inputs, the L1 or L2 (or an L3 if available) bodies provide a very sophisticated and regulated voltage host.
The results bear this out. With the KL-3 on the L1 host, I get a lower output result, along the order of 2/10 lumens. On the L2 host it is more like 7/15 lumens. With the KL-5 on the L1 host my result is on the order of 3/15 lumens. On the L2 15/80(+/-).
The only practical result of all this amateurish experimentation is to have a very bright E1e-RD if needed (and maybe a BL or GR?). Everything else, although "interesting", is too large and cumbersome to be usable. The L1 (old or new) is better than an L1-KL3/5 combo, and the U2 is better than an L2-KL3/5 combo.
I CLAIM NO RESPONSIBILITY FOR ANY OF THIS EXPERIMENTATION SCREWING UP YOUR FLASHLIGHTS! I AM ONLY RESPONSIBLE FOR SCREWING UP MINE! (in the name of science and Chivas Regal).
Incidentally, DON'T try the U2 head with a 9v host (via an A-19 extender). It pisses off the voltage regulator. It seems to be more voltage than the regulator can handle. It came on "low" (level 6) as expected, but when I started to rotate the selector ring, it went straight from level 5 to level 1. Apparently, the regulator would not successfully restrict the incoming voltage.
Hope some of this helps.