Over the years, I have found that there is one circuit fragment above all others that I go back to repeatedly, and that is the PWM controller in all its guises. I have made more of these than any other electronic sub-circuit over the last 50 years and they come in handy for all manner of different applications. I'm just curious to know what others here find themselves going back to repeatedly over the years and would welcome responses from sed contributors north of the Tropic of Capricorn in this regard.
A PNP simulated inductor with a very long tail and a low noise op amp wrapped round it. The key is to drop lots of voltage in the emitter resistor and bypass the base to the positive supply rather than to ground. Two poles work better than one. With 5V drop across the resistor, you can get down to 20 dB below full shot noise, which has interesting consequences with some lasers.
Good plan; the transistor has more bandwidth than the op amp.
Two questions: would a MOSFET give equivalent performance to a PNP, and is a small inductor in series with the emitter resistor any benefit (I'm thinking of Rbb/Miller-effect cutoff of the transistor base bypass effectiveness).
"No" to #1, and "possibly but generally not" to #2.
MOSFETs are noisy and have crappy g_M for the same current.
The way I build laser drivers, the emitter resistor is pretty large, so as to get below the shot noise on the output. The total noise goes down by 3 dB when the emitter resistor drops 2kT/e (51 mV @ 300K), and I try to run it at least a few volts.
Thus in order to make a difference over a wide bandwidth, the inductor it would have to be pretty massive. On the other hand, once its effect dominated, it would reduce the shot noise characteristic further, without needing extra headroom. Being (ideally) noiseless, it would also win linearly rather than as the square root.
Keeping the transistor stable despite a low-Z base bypass might take a bit of care as well.
Common-cathode laser. The 'simulated inductor' thing is just a cap multiplier with the base bypassed to the far end of the emitter resistor rather than to ground. Like this:
with maybe another pole (also bypassed to V+) for good measure, and a nice quiet op amp looking at the same resistor, doing the DC regulation.
Amplitude squeezing, sometimes. Thirty-odd years ago, all sorts of quantum optics folk were working on "squeezed states", in which the Heisenberg uncertainty product <delta A><delta phi> was modified to give less <delta A> and more <delta phi>.
It was all Ti:sapphire lasers and optical parametric oscillators and four-wave mixing and stuff--tables full of expensive parts on expensive mounts, conferences with big banquets, yada yada, all to get about a decibel worth.
Then these Japanese guys (whose paper I can't seem to locate) took a LN2-cooled diode laser, biased it with a stack of batteries and a big resistor, and got 3 dB, because the bias current was way sub-Poissonian and that constrained the pumping.
They had the world record for some years, iirc. (I love things like that--doing something amazing with almost zero apparatus.)
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.