Add rain-scatter draft and mark future-dated posts as drafts

content/blog/how-a-transistor-amplifies.md

@@ -1,7 +1,7 @@
 +++
 title = "How a transistor actually amplifies"
 date = 2026-04-23
-draft = false
+draft = true
 +++
 
 <link rel="stylesheet" href="/css/scene.css">

content/blog/how-tcp-and-udp-work.md

@@ -1,7 +1,7 @@
 +++
 title = "How TCP actually talks (and why UDP doesn't bother)"
 date = 2026-04-27
-draft = false
+draft = true
 +++
 
 <link rel="stylesheet" href="/css/scene.css">

content/blog/rain-scatter.md

@@ -0,0 +1,37 @@
++++
+title = "Rain scatter: using a thunderstorm as a mirror"
+date = 2026-07-10
+draft = true
++++
+
+Point two microwave stations at the same rain cell from different directions. Each one shouts into the storm, and the raindrops, being roughly the right size to notice a 3 cm wave going past, scatter a little of that energy in every direction, including toward the other station. Both operators hear a watery, wind-smeared version of each other arriving from a patch of sky that neither is pointed at directly. The weather, briefly, has volunteered to be a mirror. A lousy one, but a mirror.
+
+This is rain scatter, and it is one of the few propagation modes on the microwave bands where the thing that would normally kill your link is the thing that makes the contact possible.
+
+## Why drops scatter at all
+
+A raindrop is a small sphere of water, which is to say a small lump of something with a dielectric constant roughly 80 times that of the air around it. An electromagnetic wave hitting that discontinuity induces a tiny oscillating dipole inside the drop, and that dipole reradiates. When the drop is much smaller than the wavelength ([Rayleigh scattering](https://en.wikipedia.org/wiki/Rayleigh_scattering) regime), the scattered power per drop scales as *D*<sup>6</sup>/λ<sup>4</sup>: diameter to the sixth, wavelength to the minus fourth. Big drops and short wavelengths are enormously more effective than small drops and long ones.
+
+A typical thunderstorm has drops from about 0.5 mm (drizzle) to 5 mm (the fat ones that hurt). At 10 GHz, λ is 30 mm, so even the largest drops are still comfortably Rayleigh, and the *D*<sup>6</sup> term means the rare giant drops dominate the return entirely. The tiny drizzle drops contribute essentially nothing. A storm's radar reflectivity is, to a useful approximation, just the sixth moment of its drop size distribution, and meteorologists call that sixth moment *Z* and report it in dBZ.
+
+## The operator's version of the radar equation
+
+Weather radar sends a pulse, listens for the echo from its own pulse, and uses the round-trip time to locate the rain. Rain scatter contacts use the same physics without the round trip. The transmitting station illuminates a volume of rain; the receiving station, pointed at the same volume from a different angle, collects a fraction of the reradiated energy. The geometry is bistatic, the math looks like the radar equation with separate path lengths on the outbound and return legs, and the whole thing lives or dies on two numbers: how much rain is in the common volume, and whether both antennas actually intersect it.
+
+The second part is less obvious than it sounds. Both stations have to aim at the same parcel of sky, at elevations high enough that the common volume sits above each station's horizon plus enough altitude to clear the surrounding terrain. For stations a few hundred kilometers apart the aim points sit somewhere between 5 and 15 degrees of elevation and roughly halfway between them in azimuth. Radar maps are how you find the storm; an azimuth/elevation calculator is how you find the common volume inside it.
+
+## The signal sounds wrong, and that's the point
+
+A rain scatter signal does not sound like a clean line-of-sight contact. Every raindrop in the scattering volume is falling at its terminal velocity, being shoved sideways by the wind inside the storm, and moving relative to every other drop. The echo from the whole volume is the coherent sum of millions of randomly moving scatterers, which means the received signal shows fast [Rayleigh fading](https://en.wikipedia.org/wiki/Rayleigh_fading) (tens of dB at audio-rate) and a Doppler spectrum hundreds of hertz wide. SSB voice through rain scatter has a characteristic fluttery, watery quality; CW comes through with the keying still intelligible but each element smeared. Modes designed for weak-signal work (the JT/Q65 family) handle it by expecting a broad Doppler spread and not flinching.
+
+## Why 10 GHz is the rain scatter band
+
+The *D*<sup>6</sup>/λ<sup>4</sup> scaling is why rain scatter is almost entirely a microwave phenomenon. On 2 meters the scattered power per drop is about 1/1,000,000 of what it is on 10 GHz, and the common volume isn't big enough to make that up. As you climb through 3 GHz, 5 GHz, 10 GHz, the scatter gets stronger; past about 20 GHz the Rayleigh approximation starts to break down for the big drops (they become Mie scatterers) and atmospheric absorption from water vapor and the drops themselves starts eating real signal. 10 GHz sits in the sweet spot where the drops scatter enthusiastically, the attenuation through rain is still tolerable over the short legs inside the cell, and amateurs have been allocated enough spectrum to actually operate.
+
+Which is why the ARRL 10 GHz and Up contest sees a measurable spike in logged contacts on days with convective weather. A thunderstorm sitting between two stations that otherwise couldn't reach each other is, from a contest-point perspective, a gift. Meteorologically terrifying, radio-operationally excellent.
+
+## Finding the scatter volume
+
+The practical workflow is boring and works. Pull up a radar map. Find a cell with a decent dBZ core, ideally 40 dBZ or higher, somewhere between you and the other station. Compute azimuth and elevation to the cell; tell the other operator where to point. Both stations aim at the top third of the cell, because the precipitation is densest there and the geometry usually puts less terrain in the way of a steeper elevation. Start calling. If the cell is active enough, contacts happen on signals that the path profile insists are impossible, and a half-hour later the cell moves on and the path closes again.
+
+The storm has no idea you're using it. It's just doing what storms do, which is to say making life difficult for farmers, airports, and unclosed car windows. That you can use its internals as a bistatic mirror for a radio contact is a coincidence of wavelengths, drop sizes, and geometry that nobody designed on purpose, and which nobody who hasn't tried it quite believes until they hear their first watery CQ coming back out of a thunderhead.

content/blog/stacked-noise-figure.md

@@ -1,7 +1,7 @@
 +++
 title = "Noise figure, stacked: why your LNA goes where it goes"
 date = 2026-04-29
-draft = false
+draft = true
 +++
 
 <link rel="stylesheet" href="/css/scene.css">