Q-Switching and the Giant Pulse

Epoxy a front-surface mirror to the shaft of a small brushless motor so the reflecting face sits on the axis, and balance it — an unbalanced mirror at 20 000 rpm is a projectile. Enclose the motor and mirror. Wear the safety goggles. Set it as one end of the cavity from the Fabry-Perot blueprint. For all but a few microseconds of each revolution the mirror faces the wrong way and the round-trip loss is total; once per revolution it sweeps through alignment and the cavity exists. Measure the open window: put the photodiode where the output coupler would be, shine the laser pointer down the axis, and read the pulse width on the oscilloscope. Measure the shaft speed with the optical tachometer. Window in microseconds equals mirror aperture divided by the rim speed. Compare it against the notebook's build-up time of a few tens of nanoseconds. It is a hundred times too slow, and that gap is precisely why the Kerr cell existed by 1962 and the Pockels cell replaced it.

WHAT McCLUNG AND HELLWARTH ACTUALLY USED. A Kerr cell: nitrobenzene between crossed polarisers, with a kilovolt across it. With the voltage on, the liquid rotates the polarisation and light passes; with it off, the crossed polarisers block the cavity. Nitrobenzene is toxic and absorbs through skin, which is one reason nobody builds these any more. WHAT REPLACED IT, LABELLED AS THE MODERN EQUIVALENT. A Pockels cell does the same job in a crystal — KD*P or lithium niobate — with no liquid, switching in about a nanosecond. A passive saturable absorber such as Cr4+:YAG does it with no electronics at all: it is opaque until the flux inside the cavity bleaches it, at which point it opens itself. You trade all timing control for a part with no wires. WHY THE PULSE LENGTH IS WHAT IT IS. The pulse cannot be shorter than a few cavity round trips, because the light has to physically traverse the rod enough times to sweep the inversion out. For a 20 cm cavity a round trip is 1.3 ns, and a giant pulse lands at ten to twenty nanoseconds. Shorter pulses need a different idea entirely — mode-locking, which fixes the PHASE relationship between the modes you plotted in the helium-neon blueprint rather than gating the loss. THE COST NOBODY MENTIONS. At 10 MW in a 5 mm beam you have 50 megawatts per square centimetre. Dielectric coatings damage, dust on a mirror becomes a crater, and focused in air the pulse ionises it and you get a spark. Every Q-switched laser is engineered around its own output. Expanding the beam before any focusing element is not a refinement, it is the difference between a working laser and a scrapped optic. THE SIBLING, MEASURED. Free-running: about 1 ms, roughly 100 mJ, about 100 W peak. Q-switched at five times threshold in a 20 cm cavity: about 5 ns and tens of megawatts, for a peak-power gain of order a hundred thousand. The energy is nearly the same — the rate equations say almost all of it comes out, and a real rod delivers 50 to 80 % once spatial hole burning and the beam profile are accounted for. What you actually pay is hardware, alignment and the damage budget. If your application does not care about peak power, do not do it.