Daily Dose of 6G

The ten milliseconds nobody proposed changing

Part two of TR 38.760-1. Yesterday RAN1#126 closed in Maastricht. The frame it is designing for 6G is, in its bones, the frame 5G already has — and the reason has almost nothing to do with radio performance.

Transcribed from 3GPP’s published Release 21 timeline (© 3GPP 2026). Whatever TR 38.760-1 concludes on frame structure becomes specification at the highlighted row.

The frame is the interface to the past

Edition 08 found a physical layer that had declined almost every opportunity to reinvent itself. Frame structure is the same story told more sharply, because here the conservatism has a name: multi-RAT spectrum sharing.

An operator deploying 6G in 2030 will not have an empty band to put it in; the spectrum will be carrying NR, and in many markets LTE underneath that. If the 6G frame keeps NR’s 10 ms radio frame, 1 ms subframe and slot grid, the two radios can be interleaved in the same carrier. If it does not, 6G becomes a technology that requires a cleared band — which is to say a technology that deploys late.

That is the argument, and as far as the reporting shows it has not been seriously contested. Ericsson’s standardization blog described the emerging 6G frame structure in June 2026 as “very similar to the one used in 5G”. Juan Montojo, VP Technical Standards at Qualcomm, writing in RCR Wireless on 1 July 2026, put it as reuse of the 5G NR slot-based framework and scalable numerology.

The numerology numbers — and what tier they sit in

Edition 08 declined to print the per-band subcarrier-spacing baselines, because only one blog-tier source carried them, and promised to return to the question. Here is the return. A second independent secondary source has since been found carrying the same values, which is corroboration but not promotion: this remains reported material, not specification text.

The reporting attributes these to RAN1#122 — Bengaluru, 25–29 August 2025, per 3GPP’s own meeting calendar — under the frame structure agenda item, via a feature-lead summary the source identifies as R1-2506603. We have not seen that document. What is worth noticing is not the numbers, which are simply NR’s numbers, but how much of the table is still marked for further study. A minimum of 15 kHz is not a decision that 15 kHz is the answer; it is a decision that nothing narrower is on the table.

One subcarrier spacing per band, and why that is a real decision

5G NR permits mixed numerology within a carrier, and implementers have paid for it ever since: guard-band arithmetic, cross-numerology interference, scheduler complexity, and receivers that must be prepared for a spacing change. The reported 6G direction is a stated preference for a single subcarrier spacing per band or per cell, with synchronisation signals and the initial bandwidth part pinned to that same spacing, so a device never has to acquire a cell across a numerology boundary.

If that holds, it is one of the few places where 6G is plainly simpler than 5G rather than more capable — and it is the kind of decision that is cheap to take in 2025 and expensive to revisit in 2028.

What is actually new: channels that outlive their slot

The one structural departure in the reporting is subtle. The slot is retained as the logical timing unit — processing timelines, HARQ codebooks and control monitoring occasions are all still counted in slots. But physical channels are reported to be allowed to extend across slot boundaries, rather than being cut to fit inside one.

In NR, a transmission needing more symbols than the slot has left gets fragmented or deferred. Decoupling the physical transmission from the logical grid removes that constraint without discarding the grid everything else is defined against — a small change with a large blast radius.

Synchronisation signals: still an open question

Sync is where the coexistence logic gets uncomfortable. The study is reported to be examining high-level aspects impacting the 6GR sync signal structure and associated periodicity — which is study-phase language for “this is not settled.” Two coexistence shapes appear in the discussion: 6G either shares NR’s synchronisation signals outright, or it rate-matches around them, treating NR’s always-on transmissions as reserved resources.

The first is cheaper in spectrum and worse for independence — a 6G carrier that borrows NR’s sync cannot outlive NR. The second keeps 6G self-sufficient at the cost of permanently working around another generation’s overhead. Nothing visible indicates which way RAN1 leans. Shinya Kumagai of NTT DOCOMO INC. holds the pen on TR 38.760-1, and the portal shows it still at draft v0.4.0, uploaded 14 August 2026 — checked again this morning, unchanged by Maastricht.

The meeting arithmetic, after Maastricht

RAN1#126 ran 24–28 August 2026 in Maastricht and closed yesterday. The archive site whatthespec.net counts 1,744 documents across 53 agenda items — a scale worth holding in mind when reading any single reported agreement.

With #126 done, FS_6G_Radio has five RAN1 meetings left before it closes on 4 June 2027: #126-bis in South Korea, #127 in Calgary, #128 in Osaka, #128-bis in Lyon, and #129 in China, which ends eight days before the deadline. Five meetings to turn a v0.4.0 draft into conclusions Rel-21 can transcribe.