LTE: Carrier Aggregation - Activation and Deactivation of Secondary Cells


To enable reasonable UE battery consumption when CA is configured, an activation/deactivation mechanism of SCells is supported.
If the UE is configured with one or more SCells, the eNodeB may activate and deactivate the configured SCells. Activation/Deactivation does not apply to PCell.

After Configuring/Adding the SCell, as explained here, the SCell is in deactivated state. When the SCell is modified, the UE does not change the activation status.
During Handover, if the same SCell is in use in the target PCell (i.e., SCell is not released during Handover), the SCell in the target Cell in initially in the deactivated state.

There is no explicit activation/deactivation of uplink component carriers whenever a downlink component carrier is activated/deactivated, the corresponding uplink component carrier is also activated/deactivated.

In order for the UE to receive data (PDSCH/PDCCH) on the SCell, it has to be activated which is different from configuring the SCell. If the SCell is deactivated, the UE maintains the configuration provided by RRC but it is not possible to receive either PDCCH or PDSCH.

Activation/Deactivation of Secondary Cell
A typical use case of Activation of the SCell would that the network configures the UE with one or more component carriers (CCs) but deactivate all of them except the Primary CC. When there is a need of more data throughput (for example, there is a huge amount of data to delivered to the UE in the downlink), the network can activate several secondary CCs to maximize downlink throughput.

Another use case could be that if the PCell is fully loaded, the SCell can be activated and the data transfer can be scheduled only on the SCell (Load balancing).
The network can deactivate the SCell when there is no more data to be delivered to the UE or the channel quality of the SCell turning to be bad.

The UE (and the network) can deactivate the activated SCell without explicit signaling which is based on sCellDeactivationTimer. This is the amount of time (in radio frames) the UE has not received any data on the SCell.
SCell Activation is done via MAC Control Element whereas the deactivation mechanism is either by using MAC control element or by the expiry of the sCellDeactivationTimer.

The Activation/Deactivation MAC control element is identified by a MAC PDU subheader with unique LCID: 11011. It has a fixed size and consists of a single octet containing seven C-fields and one R-field (Reserved bit and is set to “0”). Each C-field represents an SCell with SCellIndex which ranges from 1-7.
The MAC control element carries a bitmap for the activation and deactivation of SCells: set to 1 denotes activation of the corresponding SCell, while a bit set to 0 denotes deactivation.

With the bitmap, SCells can be activated and deactivated individually, and a single activation/deactivation command can activate/deactivate a subset of the SCells.
The Activation/Deactivation MAC CE and its subheader are illustrated in the below Figure.


As discussed before, there is another mechanism than MAC CE to deactivate the activated SCell which is the expiry of sCellDeactivationTimer.
RRC may configure the UE with sCellDeactivationTimer which is located in mac-MainConfig-v1020. sCellDeactivationTimer can take values starting from 20ms to 1280ms (20, 40, 80, 160, 320, 640, and 1280ms)

E-UTRAN configures sCellDeactivationTimer only if the UE is configured with one or more SCells. If the field is absent, the UE shall delete any existing value for this field and assume the value to be set to infinity.
UE maintains a sCellDeactivationTimer (as configured by the RRC) timer per configured SCell and deactivates the associated SCell upon its expiry. The same initial timer value applies to each instance of the sCellDeactivationTimer.

If the UE receives an Activation/Deactivation MAC CE in subframe #n activating the SCell, the UE shall start (SCell Activation) or restart (SCell Reactivation) the sCellDeactivationTimer associated with the SCell in subframe #n+8.
If the UE receives PDCCH on the activated SCell indicating an uplink grant or a downlink assignment it shall restart the sCellDeactivationTimer associated with the SCell.

Similarly, if the UE receives PDCCH on the serving cell scheduling the activated SCell (Cross Carrier Scheduling) indicating an uplink grant or a downlink assignment, the UE shall restart the sCellDeactivationTimer associated with the SCell.
If the UE receives an Activation/Deactivation MAC CE in subframe #n deactivating the SCell or if the sCellDeactivationTimer associated with the activated SCell expires in subframe #n, the UE shall deactivate the SCell no later than in subframe #n+8.

If the UE receives an Activation/Deactivation MAC CE in subframe #n activating the SCell, the UE shall apply the below defined actions no earlier than subframe #n+8 and no later than subframe #n+24 or subframe #n+34 (as defined in section 7.7.2 of 36.133)
― Transmit SRS on the SCell in case if UL CA and SRS on the SCell is configured;
― PDCCH monitoring on the SCell;
― PDCCH monitoring for the SCell (Cross Carrier Scheduling);
― CSI reporting for the SCell. The UE should start transmitting valid CSI report no later than subframe #n+24 or #n+34 (as defined in section 7.7.2 of 36.133) but it can report out of range (CQI index = 0) values from subframe #n+8, otherwise the eNodeB have to perform blind decoding of PUSCH (if scheduled) from subframe #n+8 till subframe #n+24 or #n+34

If the SCell is deactivated, the UE shall apply the following activations:
― SRS shall not be transmitted in case if UL CA and SRS on the SCell is configured
― The UE shall not transmit UL-SCH on the SCell (UL CA)
― PDCCH on/for the SCell shall not be monitored.
― The UE shall flush all HARQ buffers associated with the SCell.
― The UE shall not transmit RACH on the SCell (Introduced in Release-11, and the RA procedure on an SCell shall only be initiated by a PDCCH order). When SCell is deactivated, the ongoing RA procedure on the SCell, if any, is aborted
― The UE shall stop reporting CSI from subframe #n+8 if the sCellDeactivationTimer associated with the SCell expires in subframe #n or if the UE receives an Activation/Deactivation MAC CE in subframe #n deactivating the SCell

I have created a tool to display timings of different actions defined above based on the activation or deactivation SFN/subframe. Try this timing calculator here

Watch the following deck for more details and a demo of the timing calculator


Reference: 3GPP TS 36.321, 36.213, 36.331, 36.133, and 36.300

30 comments:

  1. Hello Mr. Kumar, While the carrier aggregation is in process and a device is connected to 2 carriers(assume). The sub-frames received by PCC and SCC to the UE why they are always simultaneous. If they are not simultaneous then why? So please eleborate this part..

    ReplyDelete
    Replies
    1. The PDCCH received in PCC and SCC need not be simultaneous. As long as SCC is active, the data can be received in any subframe irrespective of PCC receiving data in that subframe. For that matter, only SCC could be used to send data. It all depends on how much data eNodB has for this perticular user and what is the load situation. Scheduling on SCC and PCC is completely flexible.

      Delete
  2. Hello Mr. Kumar.I am doing some job about carrier aggregation and I have some question. If I aggregated 2 CCs, both of them are 10MHz, so I get a 20Mhz carrier aggregated. What's the differences between this carrier aggregated and a normal 20MHz carrier when I use them?My English is not good, thank you !

    ReplyDelete
    Replies
    1. HI ,
      We can take from network side as wel.
      In case PCC cell having only 10MHZ bandwidth due to licence or Band limitation.But still UE wants to send more data within a short time.So this time can use CA if Network have additional spectrum.
      Compared to CA and non CA device cost also bit high.
      Here everything based on demand,Network implementation and bandwidth availaility.

      Delete
  3. It depends on where you are looking at the difference.

    For example on the UE side:
    In case of CA, the UE needs two receivers to decode data from two CCs, where as with single 20MHz carrier, one Rx is sufficient. More processing in case of CA (including decoding, measurements etc...)

    On the eNodeB side, it is same...two different Transmitters. But the Network has the flexibility in (for example) load sharing ...at the cost of signalling overhead.

    ReplyDelete
  4. This comment has been removed by the author.

    ReplyDelete
  5. Hi Kumar,
    I have 10 Mhz L18 and 20 Mhz L26 and we make CA on those 2 carrier, when we did DT by nemo, sometime the Scell THP not detected, is that the CA is not working?

    ReplyDelete
    Replies
    1. Hi,

      Sorry, I don't have any practical information on this band combination.

      Good luck!!!

      Delete
  6. what's the reason UE cannot monitor PDCCH/SCH on the newly activated carrier earlier than subframe #n+8? Cannot find it in 36133.

    Thanks!

    ReplyDelete
    Replies
    1. The timing requirements are carefully defined in order to ensure that there is a common understanding between the UE and the eNB.

      For example in case of Activation, the UE need to send HARQ Feedback after 4ms after receiving activation MAC CE ->> The eNB only knows that the UE has successfully received the MAC CE in about 4 + (1.5 to 2ms) uplink signally processing delay at the eNB...eNB would take about 1.5 to 2ms to schedule a UE before actual transmission.
      So it make sense to have at least 8 ms.

      Delete
  7. how CQI/AperiodicCQI feedback works with Carrier aggregation?
    1- ue report CQI for both Cells each Periodic CQI report?
    2- or Ue report CQI for only one cell each Periodic CQI?

    ReplyDelete
    Replies
    1. UE would send Periodic CSI for each configured cell separately based on the configuration provided by the RRC

      Delete
  8. I have CA (blind activation, not Buffer based) between 1800 15M and 800 10Mhz.
    During Ping Test (32bytes) repeated 3 time for each test,
    CA is always configured. but not activated.
    I want to enabled CA active during Ping Test. How can i do ? Thanks.

    ReplyDelete
    Replies
    1. Well, if you are talking about UE side, we can't do much about it. On the network side, depending upon the (implemented) options, you should be able to find an option for this. I have experienced only two options though, one to always activate sCell after sCell has been configured (not being used by networks in Live) and the other one is buffer based. Please try to set the buffer threshold to the minimum value and increase the size of the ping to that extend to see it works.

      Delete
  9. Hi kumar,

    Why ue release the secondary scell before handover?
    is it possible to perform handover with scells? if yes in which release its supported?
    and what are the benefit of performing handover with scells and without Scells?

    ReplyDelete
  10. HI Kumar,

    I can see the Scell configuration or addition in Layer 3 messages , but from which exact L3 or L1 message can I see details of MAC CE to identify from Ue logs exactly when Scell is activated and deactivated?

    thanks in Advance

    ReplyDelete
  11. Hi kumar

    What is the minimum power which can be maintained in scc when the device latches on PCC.


    ReplyDelete
  12. bruh pls help me i beg you,how do i stop ca? my network keeps on swithcing from 4g to 4g+ i get a better speed at 4g and terrible speed at 4g+ i am from india using airtel sim

    ReplyDelete
  13. Hi Mr. Kumar,
    When ScellDeactivatedTimer will start if Scell activated in subframe number 5 and Timer set as 20ms???

    ReplyDelete
  14. I have a query related to the frame timing requirement for different carriers in carrier aggregation scenario.



    IS IT NECESSARY AND MANDATORY FOR DIFFERENT CARRIERS IN CARRIER AGGREGATION TO BE ALIGNED IN SFN (SYSTEM FRAME NUMBER) OR ONLY SUBFRAME ALIGNMENT/FRAME BOUNDARY ALIGNMENT IS REQUIRED ?



    I referred to 36.104 and it only mandates subframe alignment or frame boundary alignment and not SFN alignment. PLease advise.



    This requirement applies to frame timing in TX diversity, MIMO transmission, carrier aggregation and their combinations.

    Frames of the LTE signals present at the BS transmitter antenna port(s) are not perfectly aligned in time. In relation to each other, the RF signals present at the BS transmitter antenna port(s) experience certain timing differences.

    For a specific set of signals/transmitter configuration/transmission mode, time alignment error (TAE) is defined as the largest timing difference between any two signals.

    6.5.3.1 Minimum Requirement For E-UTRA: -

    For MIMO or TX diversity transmissions, at each carrier frequency, TAE shall not exceed 65 ns.

    For intra-band contiguous carrier aggregation, with or without MIMO or TX diversity, TAE shall not exceed 130 ns. –

    For intra-band non-contiguous carrier aggregation, with or without MIMO or TX diversity, TAE shall not exceed 260 ns. –

    For inter-band carrier aggregation, with or without MIMO or TX diversity, TAE shall not exceed 260ns.
    For NB-IoT: - For TX diversity transmissions, at each carrier frequency, TAE shall not exceed 65 ns.


    ReplyDelete
  15. This comment has been removed by the author.

    ReplyDelete
  16. Hi,
    If Scell is deacitvated, could be because of RLF based on RLM/Poor SNR which UE triggers by itself. How would network know the SCC is deactivated? As there would be no CSI?
    If the Scell remain deactive as UE might be in poor network coverage area, does network removes the SCC/or move it to not configured state?

    Looking forward to get insights on these points.

    Thanks!

    ReplyDelete
  17. This comment has been removed by the author.

    ReplyDelete
  18. This comment has been removed by the author.

    ReplyDelete
  19. This comment has been removed by the author.

    ReplyDelete
  20. Hi kumar,
    What is the criteria for scell activation by enodeb

    ReplyDelete