Multi-receiver ground station: N radios / N channels / one timebase — nothing measured beyond 2 TX → 1 RX

Aberta
#457 0 comentários 0 reações 0 responsáveis Ver no GitHub

Mantenedores costumam responder em até 1 dia

Ninguém assumiu esta issue ainda.

Avaliação

Dificuldade
5/5
Tempo estimado
Mais de uma semana
Facilidade para iniciantes
25/100
Tipo de issue
Funcionalidade
Clareza
Claramente especificada
Status de atividade
Ativa
Stack de tecnologia
c

Direção de pesquisa

Start with docs/scheduled-mac.md, the linked film-set preview-link document, and the bench notes for the existing two-transmitter measurements; inspect the RX-loop and rx.ring metrics, UeRxAttribution, energy/busy sensors, and the testrig PTP nodes. Measure the requested adapter and Ethernet-node matrices, adjacent-channel effects, attribution at higher N, and co-channel delivery, then document each result’s limits and recommend a ground-station shape.

Escrita pelo modelo de indexação a partir do texto da issue.

Descrição

enhancement

Every receive-side measurement in this project is one radio on one channel, and the largest multi-transmitter experiment on record is two transmitters into one receiver (per-UE RX attribution, docs/scheduled-mac.md). A ground station that has to hear many transmitters on many channels at once has no story: no test, no doc, no tool, no measured number.

Motivation

The film-set preview-link case (docs/film-set-preview-link.md, #254 comment): 2–25 cameras uplinking 1080p60 into one video village. The arithmetic makes it a 10–25-channel plan with one receiving radio per channel — a receiving USB adapter has the same ~50–60 Mbps airtime ceiling as a transmitting one, and 25 CSMA transmitters on a shared channel defer rather than add. So the receiving end is N radios, not one, and the open questions are all about N.

Questions this issue exists to answer (measure, don't assume)

  1. N adapters on one host. How many devourer RX loops can one x86/ARM host run before the USB tree or the RX rings become the ceiling? Per-adapter frame rate and pool stalls (rx.ring) vs N, on a hub topology that actually powers N adapters (the VBUS-sag traps in the bench notes apply). Where is the knee: 4, 8, 16?
  2. N nodes over Ethernet instead. The alternative shape is one small RX+decode box per monitor, each on its camera's channel, with the recorder pulling streams over the tent LAN. This needs nothing new in the radio core but does need a timebase story: N independent TSF domains vs one PTP-disciplined timebase (the testrig PTP nodes are the scaffold, M4 in #254 is the milestone).
  3. Adjacent-channel self-interference at the ground station. 10–25 radios receiving 20 MHz channels a few MHz apart, antennas within a metre of each other — does adjacent-channel leakage from the neighbouring receiver's channel cost sensitivity? This is measurable today with two adapters and the energy/busy sensors.
  4. Per-transmitter attribution at N. UeRxAttribution holds 64 by default and was tested at 2. Does per-TA windowing stay honest at 10–25 transmitters with the same cadence ratios measured at 2?
  5. What a chip does when 25 transmitters share a channel anyway (the worst case the channel planner has to prevent): frame-count and delivery vs N co-channel transmitters, so the planner's "one or two per channel" rule is a measured constraint rather than a guess.

Shape of the deliverable

A measured matrix (N adapters on one host; N nodes over Ethernet), a doc section pairing each number with what it does not prove, and a recommendation between the two ground-station shapes. No library change is expected up front; anything found that userspace cannot do gets its own issue.

Not in scope

Range (no kilometre measurement exists anywhere; needs a field day with a vendor-driver control), the HDMI/SDI encoder side (ecosystem gap, not devourer), OFDMA/UL-MU (#236).

Relates to #254 (M4/M5 are this problem in different clothes), #381 (ARM cell node), #382 (per-chain sectors at the ground station).

Linguagem predominante
C
Estrelas
56
Forks
29
Merge médio
3h 43min
PRs com merge (30d)
75

Preparar o ambiente

Este projeto não oferece contêiner de desenvolvimento, Dockerfile nem guia de contribuição, então a configuração fica por sua conta: comece pelo README e veja nosso guia da primeira contribuição para os passos gerais.

Primeiros passos

  1. Leia a issue inteira e depois o guia de contribuição do projeto.
  2. Comente na issue dizendo que vai assumir — evita que duas pessoas façam o mesmo trabalho.
  3. Faça um fork do repositório e trabalhe em uma branch.
  4. Abra um pull request que referencie o número da issue.

Mais de OpenIPC/devourer

Todas as issues de OpenIPC/devourer

Issues semelhantes

Mais issues de C

Receba novas issues na sua caixa de entrada

Um resumo curto de issues do GitHub para quem está começando.