LED not that efficient.
 

After reading around on the Internet I was finding that LEDs are supposedly good gor around 100 lumend per watt. I thought that's pretty good. So I went around checked one of each kind of LED I have, most of which were bought last year. They are only doing 65 to 80 lumen per watt, or 60 to 76 per volt-amp.
Volt-amp is a better measure of somethings power supply inefficiency, plus I have found the power company charges for watts but penalizes for low power factor when your volt amps are higher than watts.
For example CFL with their tiny cheap ballast will use 23 watts and 40 volt-amps on a 100w replacement, so they are really only about double the efficiency of a regular 100w light bulb which also uses 100va.
A 100w replacement LED will use less than 20va so it's a little more than 5x efficient than an incandescent, and at least double the efficiency of a CFL. Which is pretty good.

But I don't know where any of these unicorn 100 lumen per watt LEDs can be purchased.
I got looking into this because I was told that my T5 lights were inefficient, outdated florescents and that I should replace them with LED. Well it turns out that my T5 lights are doing 90 to 105 lumen per volt amp depending on if it's a high output or standard.
Way more efficient than any LED I can currently buy.
All the ones that I can find do well under 100 lumen per watt, then I see some cheap Chinese crap claiming to do something like 160 lumen per watt I'm suspicious that they just wrote numbers on the box to move product.
I have found lots of complaints about no name Chinese LEDs being comically over rated.
So for now I'm going to keep using my T5 fixtures.
I'm glad I didn't just assume LED were more efficient and go waste money replacing my T5 lights with them.

Yes, but a factor you're missing in your T5 fixture calculations is that florescents radiate light omnidirectionally. This means a shocking amount of their light just bounces around the fixture and never leaves it. To the tune of 30-40%.

Likewise, the rated lumens on a fluorescent tube are for its initial light output (light output after 100 hours of "seasoning"). The actual mean light output will be approximately 10% lower.

Finally, and this is a biggie, buy dimmable LED tubes. You'll be surprised just how little time you need the full blast output... most of mine have the dinner set to around 50% and that'll go to 80-100% maybe once a week.

EDIT: LEDs also don't leak nearly as much UV as fluorescents. My work still uses T8 tubes, and anything you print in ink will be faded garbage after only a year or two :-/

I could get dimmable T5 ballasts for my dead fixtures, but I don't want them.
I don't know how the light would bounce around in the fixture half of its open.
Cheap Chinese tubes fade after a little use, I use mostly 30,000hr German tubes.
T8 tubes suck.

Another concern is the upper operating environment temperature limit on my LED tube is only 113°F.
My shed is easily go to hit that this summer. My garage, hopefully not.

T5 run in a 120°F environment no problem.

Maybe it is the type of bulb? Are you trying to find LED tube bulbs to retrofit old fluorescent fixtures?

My Cree A21 (traditional incandescents looking) bulbs are 1700 lumens / 15 watts = 113 lumens per watt

EDIT: It might be color too. The "Soft White 2700K" version of my bulbs are only 100 lumens per watt.

I found this out a while ago. LEDs are not rated the same as Incan or flour because they don't "burn out" but instead just get dimmer and dimmer. Caveat emptor. I bought LEDs from a big box store and their output dropped 20% in a month.

But how many volt-amps are the 100+ lumen per watt lights using?
Cheap power supply inefficiency is a big part of the problem.

I bought some Sylvania "250 watt replacements" about 4 years ago back when they were pushing $44 each after sales tax.
Then I moved, took the old LEDs with me and installed them at my new place, then went back and bought more of the exact same led floods when I installed more lights, now $33 each.
As far as I can tell the new and old ones don't look any different from each other.
But the old ones were also on a timer, so they were rarely ever on more than 30min. So they may still be like new.

The Cree 15 watt bulbs are in the ceiling so we will have to be content with the smaller bulbs from the lamps.

CREE: 860 lumens / 9 watts / 120V / 75 mA / 5000K (95.5 lumen / VA)
FEIT: 1600 lumens / 15 watts / 120V / 140 mA / 5000K (95.2 lumen / VA)

Where do I get these and how much are they?

The power company sells by the watt, not by the volt-amp...So, unless they're lying as to the actual power consumption of each bulb, the stated wattage is what you will be paying.

As far as I understand it, the power difference between V-A and watts isn't lost, so presumably the power company gets it back (minus losses), and just sells it to someone else. If it doesn't get dumped in to some other part of your circuit, like some purely resistive load that doesn't care.

Now, for someone who isn't connected to the grid, I'd assume power-factor has a HUGE impact on efficiency. Which is probably why you've taken note of it.

I would assume, like most things, that the PF (Volt-amps vs actual watts, for those reading) for LEDs is mostly determined by the tiny built in power supply that converts the AC in to whatever voltage DC needed to "Drive" the LEDs. Correct me if I'm wrong, but I would assume the inefficiencies you're pointing out come from said power supply, not from the LEDs themselves. Simple power supplies (like a transformer) tend not to be very efficient. I would assume for the sake of cost-savings, that LED bulbs use cheap - not very efficient - power supplies.

If they were powered from DC from the start, you wouldn't have any of this power-factor issue. That being said, there are ways to convert A/C to DC efficiently, with a power factor of 1 (to 1)...I would think if you used one of these to power/drive all your LEDs off one such source, it would be quite efficient over-all. You could power each bulb individually, but it would probably be cost prohibitive.

The chargers for my EV use switching power supplies and have a PF of 1, or darn close. This helps maximize how much power coming out of the outlet actually gets in to my batteries. It means I can use a 15-amp circuit instead of a 20 (or more) that would be needed for a cruder charger. Also means I don't have to use oversized wires to carry the extra amps moving back and forth.

For those reading who want to understand it...If you tried running 1500w worth of these bulbs off a 15 amp, 120v household circuit...you'd blow your breaker and overload your wiring. You wouldn't be pulling the ~12a that you would with a PF of 1 (like a purely resitive load, like an incandescent bulb or an electric heater)...you could easily be pulling twice that amperage. Half of the power would go right back up the line back to the power company...but they would only charge you for the power(watts) used.

Anyway, to most people, they will never know the difference, particularly with something like a light bulb. To the occasional person who is running thing off an inverter or generator, it can really matter.

Seems the OP knows most of this.

Regarding 100 lumens per watt...you might find the LEDs themselves are actually this efficient. The 110v LED bulbs I use are rated at 7-800 lumens and supposedly consume 8 watts. It's pretty close. And, as you point out, it's the power supply that is killing the V-A. Perhaps it would be as simple as applying ~110v DC to them and you would be free of the V-A inefficiencies. And perhaps the power supplies simply wouldn't work on DC, and the only way you'd be able to do it would be to bypass their power supply and drive them directly. (sounds like a lot of work). You'd need a good, efficient switching power supply if you're starting with AC.

(You used to be able to buy CREE bulbs...as far as I'm concerned, they were the original LEDs that allowed actual bulbs to be manufactured. Other brands have since come out with similar output LEDs, so they're no logner the only option)

How does the utility penalize for low power factor? I thought the meter measures watts, and that's what you pay for?

Concerning T5, I probably wouldn't install them in my garage these days. I wasn't happy with how quickly one of my bulbs went out, and they don't perform so well in cold temperatures, and they take a while to reach full brightness. I'd probably go LED now.