Elektro-L3 HRIT Array

Elektro-L3 sends its high-resolution HRIT images at 1691 MHz every three hours, at xx:12 UTC. HRIT runs at 1.15 MS/s, far faster than the LRIT images we decode live on the Elektro-L3 ground station page, so each symbol carries much less energy. Our 85 cm dish gets an Es/N0 of about −1.5 dB on HRIT, and the decoder needs about +2 dB before the first frames come out.

Instead of a bigger dish, the HRIT array uses several small ones. Every station records the same transmission as raw I/Q and uploads it. A server aligns the recordings and adds them coherently: the satellite signal is the same at every station, the noise is not. N equal stations give 10·log10(N) dB more than one, so two stations gain 3 dB and four stations 6 dB.

Several small ground stations record the Elektro-L3 HRIT downlink; an array server aligns, combines and decodes the recordings
Each station records the same HRIT transmission as raw I/Q. The array server aligns the recordings in time, frequency and phase, adds them coherently and decodes the result.

Latest HRIT images

What the array decoded in the last 24 hours, combined from all stations that delivered a recording. The decodes are still partial, so the images have gaps; every additional station closes more of them.

Elektro-L3 full disk from HRIT, Channel 5, 5.7–7.0 µm water vapour, 2026-11-19 12:12 UTC, 21 % of the disk received
Latest HRIT slot: 19 Nov 2026, 12:12 UTC · Channel 5, 5.7–7.0 µm water vapour · 21 % of the disk received · 3 stations (DC8PAT, OE3HWC, SQ9GOL), Es/N0 +1.6 dB, 28,568 frames · full resolution
Slot (UTC)Ch 10.50–0.65 µm visibleCh 20.65–0.80 µm visibleCh 30.80–0.90 µm near infraredCh 43.5–4.0 µm shortwave infraredCh 55.7–7.0 µm water vapourCh 67.5–8.5 µm infraredCh 78.2–9.2 µm infraredCh 89.2–10.2 µm ozoneCh 910.2–11.2 µm thermal infrared
19 Nov
12:12
3 stations (DC8PAT, OE3HWC, SQ9GOL), Es/N0 +1.6 dB, 28,568 frames
– – – Elektro-L3 full disk from HRIT, Channel 4, 3.5–4.0 µm shortwave infrared, 2026-11-19 12:12 UTC, 7 % of the disk received7 % Elektro-L3 full disk from HRIT, Channel 5, 5.7–7.0 µm water vapour, 2026-11-19 12:12 UTC, 21 % of the disk received21 % – Elektro-L3 full disk from HRIT, Channel 7, 8.2–9.2 µm infrared, 2026-11-19 12:12 UTC, 6 % of the disk received6 % Elektro-L3 full disk from HRIT, Channel 8, 9.2–10.2 µm ozone, 2026-11-19 12:12 UTC, 6 % of the disk received6 % Elektro-L3 full disk from HRIT, Channel 9, 10.2–11.2 µm thermal infrared, 2026-11-19 12:12 UTC, 5 % of the disk received5 %
19 Nov
09:12
3 stations (DC8PAT, OE3HWC, SQ9GOL), Es/N0 +1.8 dB, 22,877 frames
– Elektro-L3 full disk from HRIT, Channel 2, 0.65–0.80 µm visible, 2026-11-19 09:12 UTC, 5 % of the disk received5 % – Elektro-L3 full disk from HRIT, Channel 4, 3.5–4.0 µm shortwave infrared, 2026-11-19 09:12 UTC, 17 % of the disk received17 % Elektro-L3 full disk from HRIT, Channel 5, 5.7–7.0 µm water vapour, 2026-11-19 09:12 UTC, 7 % of the disk received7 % Elektro-L3 full disk from HRIT, Channel 6, 7.5–8.5 µm infrared, 2026-11-19 09:12 UTC, 8 % of the disk received8 % Elektro-L3 full disk from HRIT, Channel 7, 8.2–9.2 µm infrared, 2026-11-19 09:12 UTC, 7 % of the disk received7 % – –
19 Nov
06:12
3 stations (DC8PAT, OE3HWC, SQ9GOL), Es/N0 +2.0 dB, 20,787 frames
– Elektro-L3 full disk from HRIT, Channel 2, 0.65–0.80 µm visible, 2026-11-19 06:12 UTC, 11 % of the disk received11 % Elektro-L3 full disk from HRIT, Channel 3, 0.80–0.90 µm near infrared, 2026-11-19 06:12 UTC, 26 % of the disk received26 % – Elektro-L3 full disk from HRIT, Channel 5, 5.7–7.0 µm water vapour, 2026-11-19 06:12 UTC, 17 % of the disk received17 % Elektro-L3 full disk from HRIT, Channel 6, 7.5–8.5 µm infrared, 2026-11-19 06:12 UTC, 38 % of the disk received38 % Elektro-L3 full disk from HRIT, Channel 7, 8.2–9.2 µm infrared, 2026-11-19 06:12 UTC, 9 % of the disk received9 % – –
19 Nov
03:12
3 stations (DC8PAT, OE3HWC, SQ9GOL), Es/N0 +1.9 dB, 18,506 frames
– – – Elektro-L3 full disk from HRIT, Channel 4, 3.8 µm infrared, 2026-11-19 03:12 UTC, 28 % of the disk received28 % Elektro-L3 full disk from HRIT, Channel 5, 6.4 µm water vapour, 2026-11-19 03:12 UTC, 14 % of the disk received14 % Elektro-L3 full disk from HRIT, Channel 6, 8.0 µm infrared, 2026-11-19 03:12 UTC, 25 % of the disk received25 % Elektro-L3 full disk from HRIT, Channel 7, 8.7 µm infrared, 2026-11-19 03:12 UTC, 10 % of the disk received10 % – Elektro-L3 full disk from HRIT, Channel 9, 10.7 µm infrared, 2026-11-19 03:12 UTC, 10 % of the disk received10 %
19 Nov
00:12
3 stations (DC8PAT, OE3HWC, SQ9GOL), Es/N0 +2.2 dB, 17,903 frames
– – – – Elektro-L3 full disk from HRIT, Channel 5, 6.4 µm water vapour, 2026-11-19 00:12 UTC, 34 % of the disk received34 % Elektro-L3 full disk from HRIT, Channel 6, 8.0 µm infrared, 2026-11-19 00:12 UTC, 11 % of the disk received11 % – Elektro-L3 full disk from HRIT, Channel 8, 9.7 µm infrared, 2026-11-19 00:12 UTC, 8 % of the disk received8 % –
18 Nov
21:12
3 stations (DC8PAT, OE3HWC, SQ9GOL), Es/N0 +2.2 dB, 13,307 frames
Elektro-L3 full disk from HRIT, Channel 1, 0.6 µm visible, 2026-11-18 21:12 UTC, 8 % of the disk received8 % – – – – – – – –

Rolling gallery of the last 24 hours, newest slot first; older slots drop out automatically. The percentage is the share of the Earth disk received in that channel. These are partial decodes: black bands are image segments the array could not yet receive with enough signal. Every HRIT transmission is combined from the recordings of all stations that delivered one, and published a few minutes after processing. Click any image for the full 2784 × 2784 pixel resolution. Elektro-L No.3 data: Roscosmos / SRC Planeta. Combined and decoded with SatDump.

How it works

  • Record. Each station records six minutes at 1691.0 MHz, 2.4 MS/s, starting at xx:11:30 UTC, and uploads the file (1.7 GB) to the array server, where it is kept for 36 hours.
  • Align. The recordings differ in start time, sample clock, carrier frequency and phase. The combiner measures all four along the transmission and corrects them, also mirrored spectra and lost USB buffers.
  • Combine. The aligned recordings are added with maximum-ratio weights: the stronger a station, the more it counts.
  • Decode. SatDump decodes the combination like a single recording.
Block diagram of the DC8PAT station: dish, LNA, RTL-SDR, Raspberry Pi with nmux, LRIT decoder and HRIT recorder with upload
Reference setup at DC8PAT (JN58ua). One 2.4 MS/s stream at 1691.0 MHz feeds the live LRIT decoder and the HRIT recorder.

At DC8PAT the HRIT recorder shares the receiver with the live LRIT decoder: one stream from the RTL-SDR feeds both, nothing is retuned. One HRIT recording looks like this:

In-band power over time and signal-to-noise density over frequency of one HRIT recording
One HRIT recording (15 November 2026, 12:12 UTC, DC8PAT). Top: in-band power over the six minutes, with the transmission in the middle. Bottom: signal-to-noise density across the signal, which is about 2 MHz wide.

Live dashboard

The dashboard updates every few minutes. Uploads shows which station delivered a full recording for which slot, how much data and how long after the transmission it arrived. Combiner shows the Es/N0 of each station and of the combination, how much each station contributed, how close the combination came to the ideal sum, and how many frames SatDump decoded. The dashed lines mark the decoder threshold: about +2 dB for the first frames, +3 dB for most of them. A slot is combined as soon as two stations have uploaded, and again when a further station arrives.

Coverage and stations

Elektro-L No.3 stands over the equator at 14.5° W and covers Europe, Africa, the Atlantic and eastern South America. The map shows its coverage next to Elektro-L No.4 and No.5, and the stations of the array so far, each at the centre of its Maidenhead locator. From central Europe the dish points south-southwest, at about 215° azimuth and 25–30° elevation.

World map with the coverage of Elektro-L No.3 (14.5 degrees West), No.4 and No.5, and the HRIT array stations DC8PAT, OE3HWC and SQ9GOL in Europe
Elektro-L coverage and the HRIT array stations. Map data: Natural Earth.

Status

Three stations take part so far: DC8PAT (JN58ua), OE3HWC and SQ9GOL. On synthetic test recordings with real Elektro frames the combiner loses less than 0.1 dB, and four stations at −3 dB each decode together although none decodes alone. On 17 November 2026, 21:12 UTC, the array decoded real HRIT frames for the first time:

StationEs/N0Decoded frames
OE3HWC−0.1 dBnone
SQ9GOL−2.2 dBnone
Combined+1.7 dB9948 frames

Every frame has a valid header and frame counters in sequence. The combination loses 0.3 dB against the ideal sum. It still misses about half of the transmitted frames, so the image files are not complete yet; that needs about +3 dB, one more station of this size.

DC8PAT is off the air at the moment while we repair its receive chain. OE3HWC alone reaches about 0 dB, two to three dB short of the decoder threshold; two or three healthy stations together should get there, and every further station adds margin.

Join the array

Anyone with an L-band dish that receives Elektro-L3 can take part. The recorder, upload tools, combiner and documentation are on GitLab. Each station gets its own upload token; get in touch if you want one.