TSG RAN — the group that specifies the radio

Easy to get wrong

One from each chapter that has one. The course carries 28 in all.

• 1.3 What TSG RAN is, and why the radio gets a group of its own

  Those are counts of what this machine holds, not of what 3GPP holds. The catalogue and the meeting database are mirrors, kept up to date, but a mirror of a mirror. Read every number in this course as "the local record says", and where the difference could matter the chapter says so.

• 2.2 The six working groups, and where the line between them runs

  A transport channel is not a logical channel. The transport channel says how something is carried — with retransmission, with beamforming, with a fixed format. The logical channel says what is carried. The document is explicit that the two must be kept apart §5.5, and mixing them is the commonest way to misread a layer-2 diagram.

• 3.2 What the numbers say about the size and shape of each group

  A study report is not a draft specification. It is a finished document with its own number that stays published forever. A group whose paper is mostly study reports is not a group that has not finished — it is a group whose work needs a written argument behind every number.

• 4.2 What NR is — the shape of the 5G radio network

  The picture has no wire to the device on it, which surprises people. The radio link is called Uu and is not drawn as an interface between nodes, because it is not one — it is the whole subject of the rest of the document.

• 5.4 The radio protocol stack, sublayer by sublayer

  RLC numbers its packets and so does PDCP, and the two numbers are independent §6.3.2. Reading an RLC sequence number as if it were a PDCP one is a reliable way to conclude that packets have gone missing when they have not.

• 6.3 The three states a device sits in

  "Camped" is not "connected". A camped device has read the system information and knows where it would go if it needed to, and has told the radio network nothing. The base station does not know it is there.

• 7.2 How a connection is set up, and how it is moved

  Contention-free is not "better". It costs the network a dedicated preamble per device, so it is spent where the timing matters — handover, and re-establishing alignment on a secondary timing group — and not on ordinary access.

• 8.3 Where the radio network meets the core

  A 5QI is not a bandwidth and a QoS flow is not a bearer. The radio network is free to put a guaranteed flow and a non-guaranteed flow on the same radio bearer, and the clause says outright that how to optimise that is not standardised §12.1.

• 9.3 The specification families, and what a number tells you

  Do not read a study report's number as a statement about its owner. A number in the 38.8xx range can belong to RAN1, RAN2, RAN3, RAN4 or the plenary. Open the information page instead.

• 10.3 The documents that attract the most change, and what that says

  "Documents changing it" is not the same as "approved changes". The count includes every revision, every withdrawn attempt and every draft that named the document, because the database records the specification a document points at, not whether it got in. Read the ranking as a measure of effort, not of outcome.

• 11.2 How a radio feature travels from a question to a rule

  "Study" here is a document type in a database, not a defined 3GPP term. When the difference matters, say "a document the database records as a study". The counts in this chapter are counts of that field.

• 12.4 The meeting rhythm — plenary, working-group week and bis

  The bis meeting keeps the previous meeting's number with "bis" after it — RAN1#124-bis follows RAN1#124 and precedes RAN1#125. It is not a separate number in the sequence, so a document from a bis meeting can look out of order if you sort by meeting number alone.

• 13.4 Where RAN meets the groups that write the rest of 3GPP

  A liaison statement's status in the record tells you almost nothing about its effect. R1-2603888 is recorded as "not treated" — the meeting did not get to it. That is a fact about an agenda, not a fact about the question, and the record says nothing more.

• 14.4 Conformance testing, and why a separate group writes the exam

  That the test document changes almost as often as the requirement is not duplicated effort. A single new band adds one row to a requirement table and a whole test case, with its own configuration and its own tolerance, to the test document. The test is the larger piece of writing, not the smaller one.

• 15.7 How to follow TSG RAN yourself, from a cold start

  Only 23 RAN meeting folders are mirrored on this machine ^[1], the most recent ones. Everything older has to come from 3gpp.org. Never assume a document is here because it is listed here.

• 16.5 What the record does not tell you

  A clause number in this course always belongs to TS 38.300 and to nothing else. If you see a clause reference on any page here, that is which document it is in.