In the HDMI2.1 source test, the oscilloscope simulates the behavior of the sink and provides the termination resistance and termination voltage. The EDID emulator simulates the EDID of the sink and provides resolution/rate information. The EDID emulator of HDMI2.0 also provides SCDC information to complete the communication with the source, so that the source output requires TMDS signals. The test items are divided into single-ended signal test and differential signal test, and the corresponding connection methods are single-ended connection and differential connection respectively, which are used to collect single-ended and differential signals to complete the corresponding test items.

Test difficulties of HDMI1.4b/2.0:
1) In some solutions, the terminal voltage needs to be provided by an external power supply, or the terminal voltage is not adjustable, and the limit situation cannot be verified;
2) Single-ended test and differential test signal acquisition need to change the hardware connection, the process is cumbersome and time-consuming;
3) The test signal rate varies with the resolution, and the resolution needs to be set manually, and the test cannot be automated.

These problems have been perfectly solved in the Tektronix HDMI2.1 FRL test solution.

In pursuit of better visual effects and experience, people are not satisfied with the 4Kp60Hz display resolution, but are also pursuing the experience of 8Kp60Hz and 4Kp120Hz. But the bandwidth required for 8Kp60Hz is about 64G (RGB/YCbCr 4:4:4 format), far exceeding the supported range of HDMI2.0. Therefore, the HDMI Association has added the HDMI2.1 FRL (Fixed Rate Link) mode to increase the interface bandwidth and meet the needs of 8Kp60Hz. At the same time, it is necessary to combine the corresponding YCbCr 4:2:0 encoding and video compression technology.

How FRL Mode Achieves Increased Bandwidth

There are two common methods for increasing the bandwidth in FRL mode. Method 1: Increase the channel data rate; Method 2: Increase the number of channels when the rate remains the same. Both methods are used in FRL mode. While keeping the HDMI physical interface unchanged, the rate supported by each channel is increased to 12Gbps; in addition, the original TMDS Clock channel is redefined as FRL Lane3 (the clock is embedded in the data stream); TMDS Data 0/1/2 correspond to FRL lane 0/1/2, as shown in the figure below, has a total of 4 data lanes. This achieves a bandwidth of up to 48Gbps. The encoding method of the signal is changed from 8b/10b of TMDS to FRL 16b/18b format, and the encoding efficiency is higher.

FRL mode can be divided into two modes: in 3 lanes working mode, only supports 3 Gbps and 6 Gbps rates; unused Lane3, source and sink all need to use differential 50Ω~150Ω termination. In 4 lanes working mode, it supports four speeds of 6/8/10/12 Gbps.

HDMI2.1 source test

There are 9 test items in total, as shown in the following table, take test Lane0 as an example.

1) The test signal is a fixed code pattern, and the test defines a total of 8 code patterns Link training pattern 1~8, abbreviated as LTP1~8. Unlike HDMI1.4b/2.0, there is no requirement for code type.

2) The test signal rate is fixed and does not need to vary with resolution.

3) The interference of other lanes needs to be considered, such as the HFR1-1 project. When testing Lane0, Lane0 needs to send LTP5 patterns, and Lane1/2/3 send LTP6/7/8 patterns respectively. The test method is more complicated.

Difficulties in source testing

Realization of Termination Voltage

Tektronix oscilloscopes and probes do not require an external power supply, and can not only provide a standard 3.3V termination voltage for compliance testing required by the association. In the user-defined mode, an adjustable termination voltage is also provided, for example, a termination voltage of 3.0V is set to verify the condition of the source-side chip when the termination voltage changes.
Automatic acquisition of single-ended and differential signals

Corresponding to single-ended projects and differential projects, it is necessary to collect single-ended signals and differential signals respectively during testing; in the HDMI1.4b/2.0 test, differential signals are collected through differential probes; after manually changing the probe hardware connection, single-ended signals are collected. Signal. Changing the connection is cumbersome and cannot be automated, resulting in low test efficiency.

Tektronix Tri-mode probe (three-mode probe), under the control of the test software, alternately works in single-ended mode (A-GND and B-GND), without changing the hardware connection, it can realize the acquisition of 8 single-ended signals , and then automatically calculate the differential signal, thus realizing the automation of the entire project. In addition to the three-mode probe solution, Tektronix also provides a cascaded automation solution for two oscilloscopes, which can simultaneously acquire 8 single-ended signals through 8 channels, making the test more efficient.

Solve the problem of test complexity

As the speed increases, the HDMI specification defines new equalization technologies and cable models, which also complicates the testing process. The specification defines two Cable modes: Category 3 Worst Cable Mode (WCM3) and Category 3 Short Cable Mode (SCM3). Two kinds of equalization: CTLE 1～8 dB and DFE 1-tap d1 value 25mV.

After the signal is collected at TP1, the cable model is applied to obtain the waveform at the TP2 position, and the reference equalization is applied to obtain the waveform at the TP2_EQ position.

The eye diagram calculation method is more complicated, considering not only the insertion loss of the cable model, but also the crosstalk introduced by other data lines.

The Tektronix solution optimizes the algorithm for the above situation, and the test time is short.

Test rate and pattern automatic switching

Previous tests required manual resolution changes to achieve changes in test signal rates. Now Tektronix uses the test software and the EDID/SCDC simulator to set the test signal rate in FRL_Rate in SCDC (Status and Control Data Channel) offset 0x31, and set the pattern for each Lane in offset 0x41/42. It realizes the automatic switching of the rate and code pattern required by the test, realizes the complete automation of the test, and improves the test efficiency.

Tektronix HDMI2.1 FRL Automation Solution

Configuration 1: DPO 70000 SX oscilloscope cascade scheme. Two DPO 70000 SX oscilloscopes, synchronously cascaded using the UltraSync cable, can adjust the skew of 8 channels to within 1ps, ensuring the synchronization of all single-ended signal acquisitions. After collecting 8 single-ended signals at the same time, 4 pairs of differential signals are automatically calculated and generated. The test process does not need to change the hardware connection, the signal path attenuation is small, the test speed is fast, and the efficiency is high. With EDID emulator, automatic switching of rate and pattern is realized.

Configuration 2: DPO70000SX oscilloscope with Tri-mode probe. Using the characteristics of the Tri-mode probe, under the control of the test software, it alternately works in the single-ended mode (A-GND and B-GND), and completes the acquisition of 8 single-ended signals in stages. The testing process also does not require changes to the hardware connections. The connection diagram is as follows, the oscilloscope will automatically de-embed the probe to eliminate the influence of the probe on the signal. Taking into account cost and efficiency, automated testing is also achieved through EDID emulator.

Oscilloscope Bandwidth Considerations

In the HDMI2.1 specification, the recommended oscilloscope bandwidth is 23GHz or above. Due to cost considerations, you may ask, is a 16GHz or 20GHz bandwidth oscilloscope okay? On the one hand, from the perspective of rise time and bandwidth, the fastest rise time allowed by HDMI2.1 signals is 22.5ps (20%-80%). The rise time measured by the oscilloscope can be calculated using the following formula:

As can be seen from the above table, the higher the bandwidth, the smaller the measurement error of the rise time. From a bandwidth perspective, an oscilloscope's bandwidth is defined as the frequency at which the amplitude of the sine wave observed by the oscilloscope is attenuated by -3dB. In the actual test process, the non-sinusoidal signal needs to consider the 3rd to 5th harmonics. The HDMI2.1 signal rate is up to 12Gbps, the fundamental frequency is 6GHz, and the 3rd harmonic frequency is 18GHz. The 16GHz bandwidth oscilloscope measures the 3rd harmonic component will be attenuated by more than -3dB. On the other hand, if the maximum FRL rate of the tested HDMI2.1 DUT does not reach the upper limit of 12Gbps, the bandwidth requirements of the oscilloscope can be actually evaluated according to the above calculation method.

Simply put, in order to ensure better measurement accuracy and test compliance, the higher the bandwidth of the oscilloscope, the better.

Summarize

Tektronix oscilloscopes utilize channel adjustable termination voltage, single-ended characteristics of Tri-mode probes/oscilloscope cascade characteristics, and cooperation with EDID/SCDC simulators to achieve true automation of HDMI2.1 FRL source end testing and improve testing efficiency. Optimized algorithms specifically for FRL signals speed up testing. So as to help customers to quickly verify HDMI2.1 products and accelerate the process of customer product marketization.