• About EEVblog

    Check Also

    EEVblog #1054 – How an Analog PC Joystick Works

    How does a PC read an old school analog joystick? It might be more interesting ...

    • Simon_muc

      Hi Dave,

      The colour properties of the old tube are given in the number “840” (behind the wattage). The “8” stands for a colour rendering index of 80-90%, the “40” for a colour temperature of 4000K .

      Greetings from Germany and keep up the good work!

    • rghisi

      Nice analisys Dave, as always! 🙂

      As I could understand from the Dave CAD diagram, you are using magnetic ballasts, right?

      And what about eletronic (switch mode) ballasts? Do they really improve the efficiency? And in the lumens output, any gains?

      Pardon me if this does not make any sense.

      Greetings from Brazil!

      • Yes, they will give you better efficiency. Around 15 percent might be a typical figure, but it varies.

    • f4eru

      the starter is not closed at power on.
      it first is open and is in fact heating from being a neon light.
      when it closes it makes the filaments heat up but the starter is in fact cooling down then reopening.
      once open, the main mercury vapor tube is firing, and the starter cannot heat up again because the voltage on the main tube is lower than the conduction voltage level on the neon starter tube.

      Usually a two tube light has the big capacitor not as a pfc, but in series in the input circuit of one of the tubes, and this makes a phase shift for the second tube to avoid flicker.

      • Worf


        It’s also how the starter “knows” it needs to close again if the tube doesn’t fire when it opens. You can also see it that the starter glows when it’s trying to start the tube

        The filaments are two parts – first is to heat the vapor when cold, the second is to be an electrode supplying power to the vapor.

        The starter does just that – it “shorts” to get the filaments warm enough to create the vapor in the tube, and breaks to give the tube a jolt to start the ionizing process and light up.

        The first small spike is the starter heating up (tube is open-sircuit, so mains is across the starter and it glows and warms up). Then you see the starter close and the large starting current flows as the filaments heat up. Starter cools down and breaks, and the ballast is a huge inductor so it tries to apply a large voltage across the tube to kick it into conduction.

        The ballast serves two purposes – first is to limit initial inrush current, the second is to provide the starting impulse to kick the tube on (when the starter opens, the ballast’s field collapses which generates a pulse of current which gives a spike in voltage and hopefully enough of a kick to force the tube into full conduction).

        And the current waveforms show that 0-300ms, the starter is warming up, then 300ms-1.5s is the starter opening and closing to kick the tube into conduction (the spikes being the ballast energy supplying some kick). Finally when it’s on it’s steady-state.

    • Rick

      Each one of these videos contain small bits of useful info. I started watching this video just by curiosity and suddenly learnt how to connect a ballast (in series with the tube) for my future UV exposure unit.

    • JRR

      I thought that the purpose of the ballast was to provide a high voltage pulse when the starter opened. The starter closes, the filaments heat up and vaporize the mercury in the tube, then opens and the ballast dumps a high voltage pulse, ionizing the mercury vapor, and then the thing runs at 110v or 220v. It does limit inrush current also.

      In the US, I haven’t seen a fixture with an old style starter for many years. The older stuff is rapid start, but as of now electronic ballast is required by law.

      • It does, and I mentioned that later on in the video I believe.

        • There is no inductive ‘kick’ from the ballast. If, when the starter opens, the tube striking voltage is not lower than the starter breakdown voltage then the starter will simply fire again. If the tube striking voltage is lower then the tube will light instead.

          You can prove this by replacing the inductive ballast with an equivalent resistive one. All that will change is the running efficiency, the fitting will start just the same.

    • Thomas

      Hi Dave,

      Great video! Never realized how easy it would be to measure light.

      Do you know of an easy way to measure the color of a light? You mentioned using a camera but I think this will not be very accurate: I’d like to get a good measure of the color spectrum of led lights. Are there any easily available instruments that could do that?


      • A proper pure white whitebalance card and a video camera does a reasonable job. No idea about instruments to do it, lever looked into it.

    • Rafael

      I loved the video; as I get older I started to pay much more attention to the quality and amount of lighting I get in my work and living areas – this makes a huge difference to keep your attention and avoid fatigue when working…

    • I gues there is a reason why there are no windows in this office – usually the outside world, especially women in stileto shoes have a lowering effect on work efficiency 😉

    The EEVblog Store generally ships twice a week, on Tuesdays & Fridays, Sydney time. Dismiss