As some of you know I work in a library, and we have the UV tube covers on all lamps.  We mostly have the Sylvania Ecologic tubes, but there have been 3 old Philips tubes as well.  All are 4 foot T-8.  Anyway, when all 3 of the Philips lamps failed, they melted the end of the tube cover pretty badly, turned burned brown, cracked etc.  However when the Sylvania lamps have failed they haven't done this.  Any reason why? What makes the Philips lamps different. The Philips are Alto's, but all of the Sylvania are also their Ecologic low mercery tubes so that shouldn't make a difference.  Just wondered.

Do they look like this one? That one I linked is the worst I ever seen. If the ones in your library is worse, can you take a pic of the ends and post them here? Thanks.

Even when both are "low mercury", each manufacturer use different approach, yielding different behavior.

This may be little oftopic, but I have to still ask. What exactly are those Philips alto lamps? I know they are low mercury models, but do they have any other difference? I have never seen one in here Finland. All Philips FLs on sale in Finland are philips master TLD. And I'v never had any problems with those lamps. ecxept those are little slow to brightening when temperature is even little under normal room temperature. Are those alto lamps only U.S. sale or do they sell them in Europe too?

Another difference between the Standard mercury lamps and Altos (Lo-mercury) is the endcaps of Altos are colored green while the standard lamps are silvered. The other difference is the quality. Standard lamps last longer than the altos which goes merc-starved or red-burning as some of you called the effect. I dunno about other differences. Others might know them.

Since philips alto's are low mercury, they might have a slightly different filament design that generate more heat.....

Do they look like this one? That one I linked is the worst I ever seen. If the ones in your library is worse, can you take a pic of the ends and post them here? Thanks.

It looks similar to this one, however the black endcap is still intact.  It is hard to see it fully because it is in one of those parabolic troffer lights and one can't see the ends as well as in the fixture you discribed.  It was a funny process because the lamp end blackened rather rapidly and the opposite end of the tube started swirling, then the swirling moved towards the center of the tube.  It did this for a couple of days before finally going out.
Does anyone know if a fire has ever started when one of these tube guards has melted like that? Doesn't seem like the safest thing to me.

Since philips alto's are low mercury, they might have a slightly different filament design that generate more heat.....

The overheating of tube ends happen mainly on electronic ballast, when the cathode emission material is gone - the cathode fall (potential difference between the cathode and discharge plasma) increases and cause large power dissipation on the remaining metal parts, what overheat the tube end region. And the electronic, because this type is capable to reliably keep the arc even when the lamp voltage is really high. Good ballast sense this and shut OFF the lamp, this protection is reset usually by power OFF-ON cycling. This feature is obligatory (by the code) on all new ballast, but some older installation might not have it yet.

On preheat the excessive voltage drop causes the starter to activate, effectively shorting the lamp (glowing filament dissipate much less heat then large cathode fall).

The RS ballast with the capacitor in series with the lamp might continue to burn such defective lamp, causing the overheat as well (there is 2*(OCV - normal lamp voltage) headroom for the excessive cathode fall, before the arc extinguish, due to lamp rectification - one electrode always fail at least a bit sooner then the other). But such condition will not last to so high cathode fall as on the electronic, because the 60Hz is slow enough to extinguish the arc during zero cross, when the fall becomes really large.

If the ballast has a winding directly connected in parallel with the lamp, the ballast would be overheated due to lamp rectification (rectification causes DC current, this causes DC magnetic bias, what reduces the magnetic flux headroom for the AC operation, so the core saturates, causing high current in the primary). The only reliable defense is the resettable thermal protection (or automatic by the lower temperature, or manual by push-button or by power OFF-ON cycling)

Let me point out that Philips have cathode shield guards....which could be one reason why the Philips seems to be the worst....