Obviously both incandescent lamps and filament LED lamps are sealed glass bulbs filled with a gas and containing a light source inside. I would think their construction would be very similar, and it is, but there are some differences that I have noticed and I would like to know why they exist.

1) Envelope Shape:
I have noticed that A19 envelopes for filament LEDs seem to be a slightly different shape compared to those of incandescent lamps. In many incandescent lamps, the neck of the lamp transitions very smoothly from roughly tubular at the base to spherical in the bulb area, whereas in LEDs the transition is often more abrupt. Why would they change the envelopes when transitioning to LEDs? I am assuming much of the same equipment is used.

2) Supports:
In most lamps that I have seen, there is a glass support stem coming up from the pinch seal. This stem then ends and is connected to support wires that grab the filament, whether it be tungsten of LED. In incandescent lamps, this glass support is most of the time made of glass tubing of the same size as the exhaust tube. Presumably one single tube is pinched into the seal, with one side being used as a support and the other side blown open to be an exhaust tube. In LEDs, this inner support is most often a solid glass rod, meaning instead of using one piece of glass for the exhaust and support, they used two. I can’t even begin to imagine why this is the case cause I can’t see how making this change would make any difference to anything at all.

3) Base Mounting:
Both filament LEDs and incandescent lamps seem to be mounted to their bases with what looks like plain old basing cement. But of course LED envelopes have less surface area to adhere to since a lot of the base space is taken up by the driving circuitry. And the basing cement in LEDs looks less brown. I believe the basing cement starts green-ish and turns brown when it is fired, and many of these LEDs seem to be on the greener side. Is this to avoid cooking the circuitry when firing it? Also I haven’t seen any detachment problems but I would assume they would be more common when there is less cementing surface area, is this true?

4) Base Material:
Practically never seen an LED that didn’t have a nickel plated brass base. Why not aluminum? They were used plenty for incandescents, is it an aesthetics thing? Bare brass would be nice too but those seemed to sort of lose popularity at some point.

Thanks!

1) The bulbs are the European PS60 style, vs the old American A19.  The reason is simply that China always favoured the European bulb style even for its incandescents.  And today that is mandatory for LEDs under the international safety standards.  The old narrow neck American A19 bulb was dangerous in some lampholders, because the customers’ fingers can get within less than 3mm from the live cap shell.  That is not permitted under present creepage & clearance requirements.  Old technology legacy incandescents could still be made that way since they had always been like that in USA.  But the LED standards outlawed it for those newer technology products and required manufacturers to finally correct that safety risk.  By making the bulb neck wider, they pass the IEC electrical safety finger test in many more kinds of lampholders.

2) Supports.  Even the older incandescents used glass cane for the arbor.  Around the 1960s that changed to tube to make a cost saving, but it is slightly weaker.  For the LEDs the cost of the inner glass stem is insignificantly low compared to other components, and also the weight of the LEDs is much greater than tungsten filaments.  As a cautionary measure to avoid failures, some companies went back to cane arbors.  But more recently as designs have been optimised they are mainly returned to tubing.  Some companies have even gone a step further and switches to a stiff metal wire.

3)Torsion strength is actually no worse for the LEDs despite the reduced contact area.  With incandescents there was a big degradation in bond strength due to the high temperature.  Manufacturers had to demonstrate that they still achieve 1.15Nm after 1000 hours at 210C.  Since the LEDs run cooler there is no significant degradation and a reduced contact area is sufficient.  Many Chinese factories use the old Osram recipe for basing cement, which contains a more intense green dye.  The curing, as far as I know, is still the same.

4) Aluminium bases are not so much used because they are softer and deform more easily.  This could compromise creepage and clearance distances to conductive parts of the driver inside the base.  This is easiest to avoid by using a harder metal such as brass.  Aluminium is also harder to solder and requires fluxes that are today very expensive or banned.  So usually they were usually welded.  Welding to aluminium also a tricky process that could be done with very thin monel fuse wires of incandescent lamps, but very difficult with the thicker wire materials used for LED drivers. 

@James
1) I had no idea there were different shapes that looked like that, very interesting. I have been in contact with the live shell of a lamp that was in my dim bulb tester (actually live, not neutral), that definitely makes a lot of sense.

2) Very interesting, I haven't seen or wire supports in LEDs yet but now that you say that I'm gonna be on the lookout

3) That makes a lot of sense, I didn't think the heat would affect it much but obviously it is cured by heat so heat does something.

4) Interesting, I have seen incandescents with both types of connections. I have been looking into aluminum soldering recently and was very surprised at how difficult it is! Even with zinc chloride acid flux, a zinc-bearing solder alloy that I made, and vigorous cleaning and scraping while hot and fluxed I still couldn't get adhesion! No doubt whatever they put in there to get the solder to stick has to be pretty nasty.