mmWave & 5G

DCI Format 0_0

Q: How does DCI format 0_0 differ from DCI format 0_1?

Format 0_0 is the fallback uplink grant with a minimal, mostly fixed field set: it always assumes single-antenna-port transmission, resource allocation type 1 (contiguous RBs), no carrier or bandwidth-part indicator, no CSI request, and no codebook or sounding fields. Format 0_1 is the non-fallback grant and is configurable through RRC: it can add a carrier indicator, bandwidth-part indicator, up to two transport blocks, antenna ports and precoding fields, SRS resource indication, CSI requests, and flexible resource allocation type 0 or 1. As a result 0_0 is typically 30 to 45 bits while 0_1 can exceed 60 bits. Networks use 0_0 during initial access, in the common search space, and as a robust fallback when channel conditions are poor or RRC configuration is incomplete.

Q: Why is DCI format 0_0 size-aligned with format 1_0?

A UE does not know in advance which DCI format the gNB has transmitted, so it performs blind decoding over candidate PDCCH locations and a limited set of payload sizes. 3GPP TS 38.212 requires that format 0_0 and format 1_0 monitored in the same search space have identical payload size; if the raw sizes differ, zero-padding bits are appended to the smaller format until they match. The single most-significant bit of the payload is an identifier flag (0 for 0_0 uplink, 1 for 1_0 downlink) so the UE can distinguish the two after a successful CRC check. Keeping the sizes equal roughly halves the number of distinct payload sizes the UE must test, reducing decoding complexity and the false-alarm rate.

Q: How is the PDCCH carrying DCI format 0_0 scrambled and validated?

The 24-bit CRC appended to the DCI payload is scrambled (XOR-masked) with a 16-bit RNTI, and the UE confirms the format by descrambling with the RNTI it expects. For a normal scheduled grant the C-RNTI is used; for contention-based random access the TC-RNTI is used; and for configured-grant or power-control variants the CS-RNTI or MCS-C-RNTI may apply. The payload is then channel-coded with polar coding, rate-matched, QPSK-modulated, and mapped to the control-channel elements of the configured CORESET. Because the RNTI is folded into the CRC rather than sent explicitly, a UE only decodes grants addressed to it, and a CRC mismatch after descrambling simply means that candidate is not a valid DCI for that UE.

/dee-see-eye fawr-mat zee-roh zee-roh/

Within 5G NR, the fallback uplink scheduling grant carried as Downlink Control Information on the PDCCH, inside a configured CORESET. Defined in 3GPP TS 38.212, format 0_0 uses a compact, mostly fixed set of fields to schedule a single PUSCH transmission: a frequency-domain resource assignment (allocation type 1 only), a time-domain assignment index, frequency hopping flag, modulation and coding scheme, new-data indicator, redundancy version, HARQ process number, TPC command, and a UL/SUL indicator. It is deliberately minimal and is size-aligned with downlink format 1_0 so the UE can blind-decode both with one payload-size hypothesis. Because it carries no carrier or bandwidth-part indicator and assumes a single antenna port, 0_0 is the robust grant the gNB falls back to during initial access, in the common search space, and whenever RRC configuration is incomplete.

Category: mmWave & 5G

Schedules: PUSCH (uplink)

Spec: 3GPP TS 38.212

The Fallback Uplink Grant in 5G NR

Every uplink data transmission in 5G NR is scheduled by a Downlink Control Information message, and format 0_0 is the simplest of the two uplink formats. It exists so that the network always has a compact, predictable grant available even before the UE has a full RRC configuration. During the random-access procedure, the message-3 retransmission grant and many initial-access grants use format 0_0 monitored in the Type1 common search space (the RAR window) or the Type0/0A/2 common search space, scrambled with the TC-RNTI or C-RNTI. Once the connection is established and an RRC-configured uplink BWP is active, the network can switch to the richer format 0_1, but 0_0 remains the dependable fallback whenever robustness matters more than scheduling flexibility.

The defining trait of 0_0 is that almost nothing about it is configurable. Resource allocation is always type 1 (a contiguous block of resource blocks described by a resource indication value, or RIV), there is always exactly one transport block, transmission is always single-port, and there are no fields for CSI requests, SRS resource indication, antenna ports, or a second codeword. This rigidity is the source of its value: a fixed field list means a fixed payload size, and a fixed size is what makes blind decoding tractable for the UE. The few bits that do vary in length, the frequency-domain resource assignment and the time-domain assignment index, depend only on the size of the active bandwidth part and the configured PUSCH time-domain allocation table.

RF Essentials front-end modules and frequency converters are typically deployed alongside the baseband units that generate these grants, so the relevant interaction is at the PHY: the gNB encodes format 0_0, polar-codes it, QPSK-maps it onto the PDCCH control-channel elements, and the UE then transmits its PUSCH on exactly the resource blocks and symbols the grant specifies. Getting the payload size right is critical, because a sizing error breaks blind decoding for every DCI in that search space.

Payload Sizing and Format 1_0 Alignment

A UE cannot be told in advance which DCI format occupies a given PDCCH candidate, so it tries a small set of payload sizes and checks the descrambled CRC. To keep that set small, TS 38.212 forces format 0_0 and format 1_0 to share an identical payload size within the same search space by zero-padding the shorter one. A single identifier bit at the front of the payload (0 for the 0_0 uplink grant, 1 for the 1_0 downlink assignment) lets the UE tell them apart only after a CRC pass. The base payload of 0_0 is on the order of 38 to 44 bits depending on BWP size before alignment padding.

Governing Field-Size Relations

Frequency-domain resource assignment bits (allocation type 1):
N_FDRA = ⌈ log₂( N_RB^BWP × (N_RB^BWP + 1) / 2 ) ⌉ bits

Resource indication value (start RB, length L):
if (L − 1) ≤ ⌊ N_RB^BWP / 2 ⌋: RIV = N_RB^BWP × (L − 1) + RB_start

Payload-size alignment rule:
size(0_0) = size(1_0) = max( raw_{0_0}, raw_{1_0} ) → pad shorter with zeros

CRC scrambling:
CRC_scrambled = CRC₂₄ ⊕ RNTI₁₆ (aligned to the 16 LSBs)

Where N_RB^BWP = resource blocks in the active uplink BWP, RB_start = starting RB, L = number of contiguous RBs (L ≥ 1). Example: a 51-RB BWP gives N_FDRA = ⌈log₂(1326)⌉ = 11 bits.

Format 0_0 vs Other 5G NR DCI Formats

Property	Format 0_0	Format 0_1	Format 1_0	Format 1_1
Direction	Uplink (PUSCH)	Uplink (PUSCH)	Downlink (PDSCH)	Downlink (PDSCH)
Role	Fallback grant	Non-fallback	Fallback assign.	Non-fallback
Typical size	~38 to 44 bits	40 to 60+ bits	~38 to 44 bits	40 to 60+ bits
Resource alloc.	Type 1 only	Type 0 or 1	Type 1 only	Type 0 or 1
Transport blocks	1	1 to 2	1	1 to 2
BWP indicator	No	Yes	No	Yes
Antenna/precoding	Single port	Yes	Single port	Yes
Search space	Common + UE	UE-specific	Common + UE	UE-specific

Common Questions

Frequently Asked Questions

How does DCI format 0_0 differ from DCI format 0_1?

Format 0_0 is the fallback grant with a minimal, mostly fixed field set: single antenna port, resource allocation type 1, one transport block, and no carrier indicator, BWP indicator, CSI request, or SRS fields. Format 0_1 is RRC-configurable and can add a carrier indicator, BWP indicator, up to two transport blocks, antenna ports, precoding, SRS resource indication, and type 0 or 1 allocation. This makes 0_0 roughly 38 to 44 bits versus 60+ bits for 0_1. Networks rely on 0_0 during initial access, in the common search space, and as a robust fallback when conditions are poor.

Why is DCI format 0_0 size-aligned with format 1_0?

The UE does not know which format occupies a PDCCH candidate, so it blind-decodes a limited set of payload sizes. TS 38.212 forces 0_0 and 1_0 in the same search space to share an identical size by zero-padding the shorter one. The leading identifier bit (0 for uplink 0_0, 1 for downlink 1_0) separates them after a CRC pass. Equal sizing roughly halves the number of size hypotheses the UE must test, cutting decoding complexity and false-alarm rate.

How is the PDCCH carrying DCI format 0_0 scrambled and validated?

A 24-bit CRC is appended and XOR-masked with a 16-bit RNTI; the UE descrambles with the RNTI it expects. C-RNTI is used for a normal grant, TC-RNTI during contention-based random access, and CS-RNTI or MCS-C-RNTI for configured-grant and power-control variants. The payload is polar-coded, rate-matched, QPSK-modulated, and mapped to the CORESET control-channel elements. A CRC mismatch after descrambling simply means that candidate is not a valid DCI for that UE.

Related Terms

← DC Wiring (Cryo) DCI 0_1 →