I came across a couple of videos by the same person on YouTube that I'm kinda skeptical about.

The first video concerns a 120V F6T5 undercabinet light which the guy fitted with a custom-made ballast that runs off batteries, that is supposedly preheat. That's really awesome, except the way the lamp looks during starting makes me question if it's actually being preheated or if it's just blinking a few times before the starter decides it's running. The ends of the lamp indeed glow, but you don't see the filaments themselves glowing. it looks kinda like an instant start simply not getting enough current.

http://www.youtube.com/watch?v=TLQAmyUhk34

The second video is what made me start questioning if this guy is for real. His second video explains how to supposedly convert any instant start fixture to preheat. However, the one step in this conversion is extremely flawed. It states just to simply connect a starter across a capacitor. Eh... okay, but that doesn't connect the filaments in any way. In-fact, in the video is looks like the other two pins of the PL lamp he demonstrates this with aren't connected at all. I questioned this in the comments, and he assures me that they were soldered and it indeed works. Okay, but why isn't this soldering mentioned in the video? And, the fact that the starter is simply connected across an already-existing capacitor STILL means the filaments aren't electrically connected. Not to mention that once again, the lamp doesn't look like it is being preheated, rather just blinking a few times and then turning on. I smell something fishy... what's for dinner dear?

http://www.youtube.com/watch?v=CHwop9w5tsU


If this is legitimate, that would be totally awesome. But his great lack of explanation and the behavior the two lamps he demonstrated exhibit suggests otherwise to me. Furthermore, I tried emulating a manual preheat on one of my cheap lanterns once by connecting one pin of each end to the ballast, and then directly connecting the other two pins via an alligator wire. The filaments didn't glow, and the ballast made a noise reminiscent of a dying squirrel.

Any thoughts?

1. First video : The fixture seems to be too small to have 6V sealed lead acid battry, but I think that the circuit inside his home made battery operated fixture, is the same to the one that demostrated in the second video.
2. Second video : He tried to convert an electronic instant start ballast to preheat, but the resulting circuit actually emulates the operation of a glow starter inside a HPS lamp, and is also very similar to this home made LPS fixture , although both have magnetic and not electronic ballast. but in all the cases (First and second videos, internal glow starter HPS lamp and the video of a home made 3x18W LPS fixture with glow starters), the starter can't preheat the electrodes of any of the lamps, but just start them with cold cathodes.

The "outer" pins are there connected by wires (see 0:21 and 1:03).
The lamp glow, because the starter is not closed immediately (neither with standard preheat ballast) after power ON, but it take about a second to heat up and close.
During this time, the ballast feed the full OCV, what make the lamp slightly glowing (the current flow via the capacitance to ground). Even this is qualitatively the same as with standard ballast, but the 60Hz make the related current way too low for the effect to be visible, while the tens of kHz make it high enough, so the glow become quite strong (~500x stronger than on 50/60Hz).
The filaments do not glow, because the starter is closed for way too short time (fraction of a second; property of the starter, so common for both standard 50-60Hz, as well as these HF ballasts).

But there is one difference between standard and HF ballast operation: During the time, when the starter heat up before it's first closure, the standard ballast provide no filament heating, while the HF operation feed quite a high current through the capacitor. So at the moment the starter first close, the filament of the HF ballast are already somehow warm, while with the standard ballast they are still completely cold.

Sometimes (and this is my experience with PL-S on HF ballast, what is essentially the configuration of the second video, only the starter and cap are inside the lamp base) the lamp ignite before the starter even closes for the first time, so the starter is there only as "standby guard" to short out the lamp when the emission fail, so prevent the ballast to overheat from the operation in the high voltage/current resonance mode.


The "battery powered" could be in fact both the Royer oscillator, or the resonance inverter (same as above, only between the resonator and transistors is inserted a step-up transformer).
But it can not be the simple flyback inverter circuit, this won't work correctly with the starter closed.

That's the most unusual startup i have seen, makes me want to try this ! I bet an electronic RS ballast would give even better results

That's the most unusual startup i have seen, makes me want to try this ! I bet an electronic RS ballast would give even better results

Well, how do you define "RS" ballast?
-It provide just enough voltage across the tube, so only lamp with heated electrodes could ignite and not the one with cold electrodes? Then even most of the "magnetic RS" won't belong to the "RS" category defined in this way, because their OCV is so high, the damaging glow (cold cathode) discharge appear before the electrodes reach the emission temperature.
 Avoiding this "premature ignition" is very difficult: The voltage window, where the lamp does never ignite when cold, but does ignite when electrodes become hot is quite narrow even with argon T12 tubes for at least narrow window of room temperatures, but practically nonexistent for thinner T8 tubes. And because the magnetic ballasts were designed with just enough OCV (because higher OCV mean higher losses and cost) to strike the hot lamp, with T12 it mean they start after they heated up, but the T8 ignited before with the weaker (but still damaging) glow and transition to arc afterwards

-It provide heater supply and the elevated voltage for ignition at the same time? Then all of the simplest HF electronic, except for the "single wire per lamp end" actually are such "RS", because they do heat up the electrodes simultaneously to applying the elevated voltage.
Those simplest HF have one very good feature (coming from the use of the resonance): They apply the high heating power only when the lamp have high voltage drop (not ignited, glow discharge), while after the hot electrode arc build up, the heating power drop to negligible level (~10% of the total cathode dissipation).
And actually this make them easier on electrodes than the magnetic ones, because the electrodes heat up to emission way more quickly, so the damage from the glow discharge is very limited.

By the way normal glow bottle "preheat" are not much better, because the long (the 1..3 seconds it usually take) time it take are not the lamp electrodes, but only the starter heating up, the real preheat time is way too short (usually ~0.2s) for the electrodes to reach decent emission temperature, so mostly the lamp is ignited with electrodes not yet hot enough, so into a glow discharge and the electrodes fully heat up only afterwards by the glow discharge cathode dissipation.

So why the CFL's last so short?