Interesting test with EOL lamps...

This is a test I have been wanting to try - (ever since someone on LG did similar with an electronic ballast a while back)

Here you see a standard magnetic 2xF40 LPF rapid-start ballast, (lamps on these run in series so if one goes out they both do. LPF ballasts will run the them at half brightness and flicker if one is EOL(as long as its only bad on one end).

But note here that I've got it running 3 lamps! The 2 lamps at the top of the pic are wired in parallel! (If you've never tried getting a pair of good lamps to go in parallel It won't work, only one of the 2 will fire up)...
The key here is both are EOL(1 bad filament on each), the bad ends are flipped opposite eachother. They're not full brightness (maybe around 3/4 brightness) and have a 60hz flicker - not the standard EOL flicker. The bottom lamp looks to be running normal -no flicker or anything.
(You can somewhat see the 'mercury migration' effect that slowly appeared after running for 2 full days )

Took quite a bit of fiddleing with different lamps and flipping around, plus a plasma-globe to finally get it to fire up like this. But I found it very interesting that it actually did work!!

The ballast doesn't over-heat or anything (something I wasn't sure what would happen)

-----------
I also tried something similar with 2 F15T8's in parallel on a switch-start choke, but couldn't get both to stay lit (that's going to require further testing sometime)