Well I had a second CFL to go EOL in the small bathroom. They have been in there since 2007 with a fair number of hours and lots of switches I got out of the shower and noticed a varying light intensity and determined which one it was. So I removed it and put in a GE. I had originally put in 3 Commercial Electric daylight CFLs in in 2007. I missed the first EOL.  The one I caught was very black up the tube and was glowing red and purple. I put it in a lamp in the bedroom and it glowed red on that EOL end. It finally went out. If I tapped on the lamp it would come back on for awhile then go out. I repeated this a few times and it finally would not come on any more. I wonder if this has EOL protection to extinguish the lamp if one cathode is broken. I was hoping for  a more festive end.....

It sounds like a scenario: electrode wear -> too high voltage across the lamp -> resonant capacitor (the one connected parallel to the lamp) failed short circuit -> filaments kept glowing -> filament broke

But very frequently I see rather: electrode wear -> too high voltage across the lamp -> transistors overheat -> fail hard short -> either some acting fuse break the circuit, or the excessive dissipation overheat the input doubler (in 120V CFL's)

In this case the EOL end was glowing red and then blue and back to red. It got QUITE hot at that end. I think it made a loose joint on the circuit board as when I jarred the lamp it would start working again.  And it finally gave out. I was wanting a  FESTIVE grand finale. POOO !

I'd think a broken cathode since there is EOL end at all. Arcing in the board would be in series with the lamp and the lamp ends would behave the same

Also the CFL base is not sealed and the smell of overheating PCB is hard to miss

There was no smell no smoke.  With the many short off and on switches since January 2007 I think these have done well. One out of 3 is left and from the age marks, with holes burned in the tear drop shaped spot on both ends, I expect it's not long for this world either...

Usually when there is a broken track in the electronics in the lamp circuit, the resonator is interrupted, so the inverter stops without any arcing. So there would be no smell.
Of course, when some other connection break (tank capacitor,...), it would be tons of smoke.

But according to your description of orange glowing ends, I would guess for the cathode filaments giving up after they lost the emission coat, so in fact the primary EOL protection (the filament) did it's job well. When they respond to tapping, it mean the break just recovered for a while, so the inverter restarted.
You may try to connect some fluorescent to that ballast, some of the lower wattage type (F13T5, F15Txx, F14Txx,...)

It was glowing red on the bad end, and would occasionally do purple and blue, stabilze, turn back red and it finally shut off. The other one that went EOL cracked the glass at the end where the filament was. 

In normal work of the fluorescent the cathode is not all glowing but a small spot on it where the arc stabilized

Here the effect is 2-fold :

With the emitter gone, the Vdrop on the electrode (as negative electrode) is higher which mean higher losses, higher heating

With the tube being less conductive, this may cause the inverter to run at higher frequency, so more current through the starting capacitor - the lamp may be kinda stuck somewhere between preheat and run modes instead of proper run

The blue discharge is typical for discharge without emitter. It appears to happen only above some power dissipation minimum - perhaps a metal vapor color from the sputtering tungsten ? This would explain the blue color

Also it is possible (i think) that traces of emitter were "trapped" in inner places in the electrode and the lamp returns shortly to normal color when those traces are "discovered" by the arc root

Now since the filament is allready broken, there is arc not only between lamp ends but also between the ends of the broken filament, with all related effects (glow, changing colors, arc dropping out and lamp extingushing, tapping the lamp and it restrikes etc)

I wish they would of done a 2 part electronic CFL. it would certinly would be better. and you would not need
as many CFL being toss away while the circuit board is still good. just replace the lamps. i am just stick
of electronic waste and waste of resource putting into them. that why i tend to repurpose those CFL board
into another project.

I wish they would of done a 2 part electronic CFL. it would certinly would be better. and you would not need
as many CFL being toss away while the circuit board is still good. just replace the lamps. i am just stick
of electronic waste and waste of resource putting into them. that why i tend to repurpose those CFL board
into another project.

The problem is, than the electronic have life time limits as well. And due to the high temperatures involved in the CFL's, the electroniclife is aboutthe same as the lamp life, so when the lamp goes out, the electronic is already close to it's end, so with replaced bulb it would be the ballast, what would fail soon.

However when you operate it later in way cooler environment (as a separate ballast,...), the remaining life could be still quite long, but when it would be operated in it's original environment, it would fail pretty soon.

The ballasts of CFLs are just made bad and cheap for being disposable. Make them better and they won't be behind other electronic ballasts, and can last for several lamps lifetime

Things like the 400V electrolytic capacitor : i find it balooned in many CFLs and they are dimmer and flickery. This in turn shortens the life of switching transistors and takes down the lamp efficiency. Would they put in better capacitor the ballast would last longer, be more efficient and make the lamp brighter

The ballasts of CFLs are just made bad and cheap for being disposable. Make them better and they won't be behind other electronic ballasts, and can last for several lamps lifetime

Things like the 400V electrolytic capacitor : i find it balooned in many CFLs and they are dimmer and flickery. This in turn shortens the life of switching transistors and takes down the lamp efficiency. Would they put in better capacitor the ballast would last longer, be more efficient and make the lamp brighter

This is what called "planned obsolescence" .

Planned obsolescence in CFLs is in that they are made so bad and that they are 1 part (and not electronic adapter + PL)

The following are only results of this : Since electronics are not to last too long, why invest in good components lets just put in the worst possible even if it reduces performance (brightness, life, efficiency....)