DL PHY Loopback Test

From YateBTS
Revision as of 15:10, 3 January 2018 by Iulian.Comanescu (Talk | contribs) (Page created from the interal Legba wiki)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

This page documents the yateenb PHY radiotest mode.

The DL PHY loopback test puts the yateenb module into a special test mode where it sends downlink data on DLSCH/PDSCH, receives it with a special PDSCH/DLSCH test receiver, and generates a performance report.

The purpose of this test is to check the performance of the PHY and radio HW.

  • This test should be used by developers after any changes in the PHY layer.
    • The QAM unmapper and turbo decoder used in this test are the same ones used for PUSCH/ULSCH, so it also validates the upper PHY part of the uplink side of the ENB.
  • This test can also be used to check some radio hardware in production equipment.
    • Full antenna-to-antenna loopback mode (radio loopback mode "none") will not work in equipment that uses FDD duplexers or TDD switches, like SatSites. Other loopback modes, internal to the radio, will still work.
    • Lab Kit filters partially block antenna-to-antenna loopback, degrading measured performance slightly.
  • In this test mode, normal LTE functions are not available. It is a test-only mode.

Starting and Configuring the Test

  • Before starting: Be sure debug is off. Otherwise a test failure can flood the terminal with messages. If you want to analyze a test failure, you can enable debug selectively at that time.
  • To start the test mode: enb phy radiotest action=start
  • To configure the test:
    • MCS, Resource Blocks, and Transport Block Size:
    • The test runs a a fixed MCS. The default is MCS=0 To change the MCS: enb phy radiotest mcs=<MCS>
    • The test uses a fixed number of resource blocks (RBs) to encode a dummy transport block. The default is 4 resource blocks. To control the number of resource blocks: enb phy radiotest rbs=<#RBs>
    • The transport block size (TBS) is a function of MCS and RBs calculated according to the algorithms and tables of 3GPP 36.213 Section 7, just like in normal ENB operation.
  • Loopback Mode:
    • The test can do loopback through the antennas (loopback "none") or internally in the radio in several possible modes:
      • lpf-to-lpf
      • pa-to-lna1
      • pa-to-lna2
      • vga1-to-lpf
    • See the LMS6002 data sheet for a figure that shows where these points are inside the radio.
    • The default mode is "none", which will not work on systems with strong hardware duplexing, like SatSites.
    • To control loopback: enb phy radiotest loopback=<mode>.

The Performance Report

To get a performance report:

enb phy radiotest action=report

The report is organized top-down:

  1. What the MAC scheduler sees
    • TBS
    • overall efficiciency in bits per resource element
    • available data rate at this MCS, taking error rates into account
  2. DLSH transport layer performance
    • TB frame error rate (FER) after CRC check
    • TB bit error rate (BER) after turbo decoding
    • channel soft bit error rate before turbo decoding
    • rate-matching parameters
    • overall coding rate
  3. PDSCH modem performance
    • QAM unmapping performance
    • SNR pre- and post-equalizer
    • equalizer statistics
  4. Radio performance
    • noise level
    • radio signal SNR

The report includes expected data rates and BER rates based on measured performance at each layer.

enb phy radiotest action= report
*** radio performance test report ***
test configuration:
using RNTI 1234
MCS=2 TBS=176 RBS=4
efficiencies (bits/elem): measured: 0.349206 from MCS table: 0.38
15187 TBs processed
TB FER 5.60519e-45
implied data total BW @ MAC layer 1070.31 kBit/sec
implied data BW for this TBS @ MAC layer 171.875 kBit/sec
DLSCH performance report:
TB BER 5.60519e-45
soft bits mean error (against ref) -11.4288 dB (expected BER 6.75669e-08)
soft bits mean error (against self) -11.4902 dB (expected BER 5.50639e-08)
soft bits channel BER 0.00218741 (capacity 0.977518)
secondary parameters:  C=1; Km=0; Cm=0; Kp=200; Cp=1; F=0; Bout=200; G=1008; RV=0;
coding rate 0.174603
PDSCH performance report:
soft bit MSE against self -11.4902 dB (expected BER 5.50639e-08)
remodulated QAM SNR 7.93211 dB 
reference QAM SNR 8.61388 dB 
equalized signal SNR 8.10173 dB (capacity 0.6016 bits/elem/layer, max MCS 5)
equalized signal power + noise 0.713305 dB
equalized signal noise power -7.98344 dB
channel model gain 12.8402 dB
channel model deviation 3.9191 dB
input SNR (whole subframe) 15.5071 dB (capacity 1.4766 bits/elem/layer, max MCS 11)
total subframe input power 12.6209 dB
input subframe noise -2.99408 dB

Normal performance is an "input SNR" of 25 dB or better and a "max MCS" of 20 or better. Note that the example above performs poorly and below the standard.

Stopping the Test

To stop the test mode and return to normal operation:

enb phy radiotest action=stop