. . . . .Sir Bill:

And here I was imagining a pulling off of the base and the diamond scribing around the periphery of the bottom of the glass envelope and a surgical crack of that envelope.

And then, the inspection of the filamental superstructure and pulling out of the "Svedish micro tweezers" and stereo microscope for a closer inspection of the filament proper to see if a rework on a spot weld was necessary.

Since the heavy "band" construction of those filaments is used, it almost precludes the development of a hot spot and erosion across the filament to end up with an open circuit.

BUT ! ! ! you would have found the problem at the onset, with your having to pull the socket and the detection of that distinctive darker grey to black appearance of that grainy solder remnant still left between the filament wire and the base pin.

That problem was readily apparent back in the hey- days of the use of the 5Y3' s and, its even hotter running, 5U4 big brother.

The usual situation for the first timer with a bright enough "bulb" to be able to realize what was actually happening, was to try to force solder within the tip of the two filamant pins. BUT if you have ever torn down a tube to see the wire leads coming from from
the envelope glass frit seal areas, we know how well that oxidized wire conditon is to even accepting any further solder at all. Even when being activated with a flux, with the exception of a VERY aggressive cleaning flux, akin to nitric or hydrochloric acids.

HOWEVER his next step on the learning curve was the finding that the use of a small mill or triangular file would permit the filing
away of the two filament pins on their outer exposed periphery, such that a slot is formed with a view of the internal wire then exposed.

That slot then lets one scrape the wire down inside, where its full length is exposed, with the micro tip of the finest gauge of jewelers screwdriver available, will then permit a lengthwise scraping of the wire inside.

The access slot will now accept flux insertion along with a full solder fill in of the slot and a final dressing down of any overflow with a mill file, to restore the proper rounded profile of the pin.

That filament wire WAS typically only having a tinning on the very end tip near its companion filament base pin, thus the failure with only a small surface area of tin/lead making connection versus the better conductor of the copper wire or the brass /tin plated brass of the tube connector pin. Hot and cold cycles along with high current in a small connective area brought on your condition.

NOW, As a matter of fact, do you realize that there actually was a tool that was made for the specific repair of that situation.

I happen to still have one of the units, punch stamped as the ADEL Tool Company .

In a referencing of the illustrations below, take [A] which is a top view of the rectangular block of brass and the insertion of a stainless steel "shear" the length( +) of the brass blank and the use of a stainless steel fulcrum pin as the blue reference dot is depicting.

The green circle is showing a hole in the end of the blank , the length of a tube pin, making it a snug fit.

To use the unit, one locates one of the filament pins and opens the shear to the [C] depiction and slips the tube pin within the cavity and presses the units together , fully exerting, and HOLDING, a full 'lebenteen billion ergs of pressure. One then swings the shear arm fully downwards.

Since the tube pin is within an enclosed cavity the forming of the pin must remain constant, so the other only possible action is for the corner (shown as the black arrow in the blow up) of the shear to penetrate into the thin wall of the brass tubing and fully pierce it, now, with the pivoting action of the shear upon the fulcrum, a slot is cut into the length of the brass pin up to the point of the corner then swinging back inwards at the end travel of the lever. Look at the [C] drawing and the [orange] tube pin materiel shows the displacement of that slice of metal the length of the tube pin and then its slowly stopping the cuting arc at the top, where it shows the full strip being retained within the tube cavity.

The tool is then slipped from the tube and the nice uniform slot is shown with its shiny edges showing, then being ready for flux and rosin cored solder to be flowed down into the heated tube pins length. Trim off the solder flash overextendings and all is done.

I certainly done my share of 5Y3's- 5U4's- 6146's- 2E24's- 2E26's on olden times.

Now-a -days one might as well take a well worn down # 409 cut off wheel in a Dremel tool, where the now effective diameter was down to ~1/4 inch or so, and then use it to permit up close access for cutting a slot into the tube pins outer side.