Specifically, I am asking about the factors that make certain fluorescent lamps more efficient than others. They say that T8 is more efficient than T12, but why? Is it the electrode : tube ratio? If that is the case, I would assume that short tubes (like F4T5) are one of the least efficient designs, and some really long T5 tube would be the most efficient. Is this true? I think gas fill also would also have something to do with it, but I don’t really know.

I don't know about the American energy saving lamp format, but usually it is the phosphor coating that determines the efficiency of fluorescent lamps. Rare-earths triphosphors converts UV to visible light better than calcium halophosphate and deluxe phosphors.

Annother factor is how close the arc is to the phosphors. That's why the PowerGroove T17s are more efficient than the non-PG T17s.

The major contributor is the electrode fall, so about 15V of the total arc drop is lost without any chance to contribute to the light output. So e.g. the F4T5 with its 27V arc is able to justify barely 12V of its drop so less than 2W to actually do something leading to light. Compare to e.g. F13T5 with about 90V arc (but otherwise the same design), the F4 has barely half of the efficacy.
Another contributor is the phosphor type and also the radiated spectrum. The rare earth phosphors with their narrow spike emission means they radiate less outside of the visible range, so more of the visible light.
The color desired color mix is the next contributor: The closer the spectral line is to the limits of the visible range, the less sensitive the eye is, so the same radiated power yields less lumens. Mainly the saturated red consumes the most radiated power for the least lumens. The problem is, if you are targetting high quality color rendering, you need to radiate significant portion also in the deep red. So a very high CRI lamp (95+) uses to lose 20..30% of lumens compare to an CRI80 standadr. The difference is mainly the power required to generate the deep (saturated) red part of the spectrum.
On top of that the fact the red photons carry less energy means convering to these means more of the original photon energy gets lost.

Then another contributor to differences is the already mentioned distance the UV has to travel through the Hg vapor (the same electron energy level difference is responsible for both radiating particular energy photon when releasing that difference, but as well to absorb such energu photon). So the desire to use as narrow tube as possible (without going into high current density area where other mechanisms cause other energy losses, or without overheating the phosphor - there the rare earths can operate at way higher loadings so allow thinner, higher intensity tube designs). The "PowerGroove" was one trick to reduce this distance without exceeding the current density, different fill mixes and rare earth phosphors allowed the operation at higher intensity, so with thinner tubes for the same power (the modern T5HO, or the thin spiral CFLs).
Another factor is the buffer gas mix:
Krypton leads to higher efficiency in the UV generation than Argon. Heavier atom steals less energy from the electron at collision, so more remains in the electron to do something useful, like ionizing extra atom or exciting some Hg one to generate the UV.
Plus if the buffer gas tends to absorb energy of slower electrons (and directly radiate its line, very unlikely able to excite the phosphor) than those needed to excite the Hg, you get another extra losses.
But nothing is black and white, the lighter gasses gave their function in high intensity tube designs to actually allow to operate at that higher intensity so with e.g. shorter distances (see above), even the lighter gasses like Neon are used, the energy loss due to the light gas just could be less than the gain from the thin tube and so on.

The general tube shape complexity and the consequent uniformity of the phosphor coating is another point where the efficacy differences may come from. The thing is, the phosphor layer needs to be not too thin, as then it would let way too much of the UV to pass and get absorbed by the glass without generating any light, too thick layer means it will block some of the visible light already generated. So to be at the sweetspot, the manufacturing process needs to deposit it uniformly around the tube. Pretty simple for straight tubes, but more complex shape means more difficult, so less successful uniformity, so further away from the sweet spot, so less light. The worst are the spirals, but on the other hand if you are limited to a small volume, you just can not do better.

But as already mentioned, some seemingly worse performing details could actually enable some other trick, which happens to bring more gain, so nothing is black and white, after all engineering is an art to find good technical compromise, to weasel out between all those various restrictions and drawbacks.

@Medved

Excellently explained, that makes so much sense.

So, the ideal efficiency tube would be super long and thin, have krypton buffer gas, and be coated with an even distribution of low-CRI phosphor. Stray away from that, and you get less than optimal efficiency.

Obviously this isn't practical, but it seems like something like this could be made with 8mm white neon tubing (probably quite a low CRI), with custom made hot cathodes and krypton-mercury fill. I might draw this idea out. Heck, if you make it long enough, you can make a negligible sacrifice and go with standard neon cold cathodes to get greater lifespan. This would have to be run on a bulky and expensive neon sign supply, so definitely not practical or economical, but yay efficiency right?

@Medved

Excellently explained, that makes so much sense.

So, the ideal efficiency tube would be super long and thin, have krypton buffer gas, and be coated with an even distribution of low-CRI phosphor. Stray away from that, and you get less than optimal efficiency.

Obviously this isn't practical, but it seems like something like this could be made with 8mm white neon tubing (probably quite a low CRI), with custom made hot cathodes and krypton-mercury fill. I might draw this idea out. Heck, if you make it long enough, you can make a negligible sacrifice and go with standard neon cold cathodes to get greater lifespan. This would have to be run on a bulky and expensive neon sign supply, so definitely not practical or economical, but yay efficiency right?

actually one of the most efficient fluorescent tubes full stop, is the British Argon filled 8ft 85W T12 fluorescent tube in Triphosphor guise, it has an initial lumen output of 8450 lumens, thats near-as-damn-it 100 lumens per watt, on *magnetic* control gear, I would dearly love to get such a tube and see what its high frequency efficiency would be, or what such a tube could achieve if it was fine-tuned for high frequency operation

buffer gas fill pressure also plays a part, higher pressures protects the electrodes better, but worsens efficiency, lower pressure to a degree increases efficiency but less electrode protection

this is one of the things as I understand it that can make certain tubes "HF only" in that they will have lower buff-gas fill pressures, and rely on the tube being driven at HF to protect the cathodes

in adverse conditions it can also affect striking ability of the tube, proper-cold temperature tubes like this example

https://www.lighting-gallery.net/gallery/displayimage.php?pos=-77915

or Philips TL-B tubes use a lower then normal buffer-gas fill to facilitate more reliable striking at lower temps at the expense of lamp life


I highly recommend reading these pages on James's website, and the links contained within


https://www.lamptech.co.uk/Spec%20Sheets/D%20FL%20Halo%20Ar%20GE%20F96PG17-CW.htm

https://www.lamptech.co.uk/Spec%20Sheets/D%20FL%20Sylvania%20F48T12-CW-VHO.htm


and since you mention Low CRI, if we are eschewing CRI, then just get a green coloured fluorescent tube centred around the peak sensitivity curve of the eye, I know some F40T12/G tubes are around 4400 lumens so thats about 110 lumens per watt from a regular-ol F40T12

@LightBulbFun
Wow, I wonder what their secret is to get such high efficiency from a T12!

While my proposed lighting system does not particularly value CRI, I think an all green phosphor would be taking it a little far lol. Still want to be able to see some colors!

85W T12 efficacy: Long tube with rather high voltage drop (so relatively small cathode fall), high efficient phosphor,...