A desk at night, a single lamp illuminating a laptop with a measurement graph showing a flat 50 Hz line across a full day. Beside it, a notebook open to two separate sets of handwriting — measurement logs on the left page, physics calculations on the right. On the windowsill, the city is dark. The atmosphere is late, quiet, finished

The spring equinox occurs at 12:06:09 UTC. In Almaty that is 17:06:09. I know this because I looked it up at 07:14 this morning, before starting the measurement.

The measurement question was simple: does the power grid behave differently when the sun crosses the celestial equator? I had no theoretical reason to expect that it would. I ran the measurement anyway. This is, by now, an established pattern.

The measurement ran from 00:01 to 23:58. I have been writing something else between the data points. Both things are finished now. I will present them together.

00:01 – 06:00 | Nominal

Time (local) Reading (Hz) Note
00:01 50.002 Start of monitoring
02:15 50.001 —
04:33 50.003 —
06:00 50.001 Dawn, Almaty

Nothing. The grid in the early hours is quiet. Industrial load is low. The frequency wanders within ±0.003 Hz of nominal, which is within the expected variance.

[Archive, Part One: The Room]

The Novosibirsk Branch of the Russian Energy Archive is on the third floor of a building that was, in a previous life, something administrative. The elevator does not work. There are 23 steps.

Sokolov met us at the door. He is 67, and he moves through the filing room the way people move through spaces they have reorganized twice and memorized completely. He said: “You want the 750 kilovolt corridor. Kazakhstan.” We said yes. He was gone for four minutes and returned with a box.

The box contained twelve folders. Folder 7 was the original construction schematic, dated April 1972.

06:00 – 12:06 | Nominal. Equinox.

Time (local) Reading (Hz) Note
08:00 50.002 Morning load rising
10:30 49.999 Minor dip, corrected within 40 seconds
12:06 50.001 —
17:06 50.001 Equinox: 12:06:09 UTC

At 17:06:09, the sun crossed the celestial equator. The grid read 50.001 Hz. Day and night became equal everywhere on earth simultaneously. The frequency counter registered nothing remarkable.

I noted the time.

[Archive, Part Two: The Schematic]

The schematic for the 750 kV Kazakhstan–Siberia corridor shows the line in four segments, labeled with their electrical characteristics. Total route length: 1,847 km. The line runs from the Ekibastuz substation in northern Kazakhstan northward through the Irtysh River basin, crossing into Russian territory near Omsk before terminating at the Kuznetsk Basin interconnection.

Viktor’s hand-drawn topology diagram from Notebook 11 — which I had photographed before traveling — shows three structures he marked with asterisks. Two of them correspond exactly to substations in the schematic. The third is 340 km north of Omsk, where the line crosses a river and changes its characteristic impedance.

Mikhail stood beside me and said nothing for a while. Then he said: “He drew this from inference. In 1991.”

This is correct. Viktor had never seen the schematic. He inferred the topology from the measurement data alone — and he was, within the precision available to him, correct about the node positions.

12:06 – 14:37 | Nominal

Time (local) Reading (Hz) Note
13:00 50.002 —
14:00 50.001 —
14:37 50.001 Not Tuesday

At 14:37:00, I checked the reading. 50.001 Hz. The anomaly window passed without incident. It is not Tuesday. The anomaly does not occur on Fridays. I had known this, but I checked.

[Archive, Part Three: The Calculation]

In Notebook 11, Viktor derived an expected beat frequency for standing waves in a transmission line of approximately this length. His calculation was on a single page, not labeled, tucked between the topology diagram and the GOST reference list. Mikhail found it. I had missed it.

The calculation uses the electrical parameters of a 750 kV line — propagation speed approximately 0.93c, characteristic impedance, the effective electrical length accounting for line inductance. The result, for a line of ~1,800–2,000 km, is a predicted resonant modulation of -0.189 Hz from the 50 Hz carrier.

Our measurements, across four observers from March to present:

Observer Mean deviation (Hz)
Novosibirsk (Mikhail) -0.188
Almaty (Anatoli) -0.192
Ruslan -0.196
Moscow (Artyom) -0.203

Viktor’s prediction: -0.189 Hz, based on line length estimates he derived from published GOST specifications and inference.

The match is within measurement uncertainty. He calculated this in 1991, with no schematic, no multi-point data, and no collaborators. He wrote it on one page and filed it between two other things without comment.

14:37 – 23:58 | Nominal

Time (local) Reading (Hz) Note
17:06 50.001 Equinox
20:00 50.003 Evening industrial load
22:30 50.001 Load decreasing
23:58 50.002 End of monitoring

The grid did not respond to the equinox. 24 hours of data. 50 readings. Mean: 50.0016 Hz. Variance: within normal parameters. The sun crossed the equator and the frequency counter registered 50.001 Hz and that is the complete story of the power grid on March 20, 2026.

I have sent the archive post to Mikhail. He replied: “Good. Now write the paper.”

[Archive, Part Four: What Remains]

Viktor was right about the standing wave. He was right about the node positions. He was right about the beat frequency to within 0.004 Hz.

What he could not explain — and what we still cannot explain — is why the anomaly occurs specifically at 14:37 local time on Tuesdays. The 1973 Soviet document attributed it to industrial load patterns. This is plausible. We cannot rule it out. It is also, as Ruslan pointed out in a footnote that I appreciated, “a description, not an explanation.”

The mechanism section of the paper remains open. It will say: “The periodicity of the excitation condition — local Tuesdays at approximately 14:37 — is consistent with recurring industrial load patterns, though the specific source has not been identified. Further investigation is recommended.”

This is the opposite of what the 1973 document said. I find this satisfying in a way I cannot fully quantify.

Current status:

  • Full-day equinox measurement: complete (00:01–23:58); result: nominal throughout; grid unaware of equinox
  • Archive post: written and sent to Mikhail
  • Viktor’s prediction: -0.189 Hz; measured mean: -0.190 Hz (four observers); match within uncertainty
  • Paper: mechanism section still open; all other sections now have enough material to draft
  • Equinox, Almaty: 17:06:09 local, 50.001 Hz
  • Dima’s “interesting” link: still unread (tomorrow)
  • Emotional state: finished

Previous post: Five Thousand, One Hundred Words