WEBVTT NOTE This file was exported by MacCaption version 7.0.13 to comply with the WebVTT specification dated March 27, 2017. 00:00:00.400 --> 00:00:06.240 align:center line:-1 position:50% size:49% Let's move on to the next topic and that's eye-diagram mask testing. 00:00:06.240 --> 00:00:12.813 align:center line:-1 position:50% size:63% Eye-diagram mask testing is something that's used in high-speed digital communication 00:00:12.813 --> 00:00:19.620 align:center line:-1 position:50% size:50% all over the world, especially Gbps type communication. 00:00:19.620 --> 00:00:21.955 align:center line:-1 position:50% size:50% It could also be used on the CANbus. 00:00:21.955 --> 00:00:25.025 align:center line:-1 position:50% size:58% What an eye-diagram is, what it shows you, 00:00:25.025 --> 00:00:30.230 align:center line:-1 position:50% size:34% is it's an overlay of all bits on top of one another 00:00:30.230 --> 00:00:36.870 align:center line:-1 position:50% size:56% to show you worst-case amplitude, noise, jitter, or other waveform anomalies. 00:00:36.870 --> 00:00:44.912 align:center line:-1 position:50% size:72% Years ago, a conventional eye-diagram measurement required an explicit clock. 00:00:44.912 --> 00:00:49.650 align:center line:-1 position:50% size:64% You had a data signal and a clock signal, and you just simply triggered on the clock signal, 00:00:49.650 --> 00:00:53.353 align:center line:-1 position:50% size:64% and the data signal would automatically overlay. 00:00:53.353 --> 00:00:58.225 align:center line:-1 position:50% size:59% Today, clocks are embedded within the data. 00:00:58.225 --> 00:01:02.930 align:center line:-1 position:50% size:50% It requires a clock recovery capability. 00:01:02.930 --> 00:01:06.533 align:center line:-1 position:50% size:38% Let me show you a diagram. 00:01:06.533 --> 00:01:15.842 align:center line:-1 position:50% size:58% Here, we show a few bits of one CAN frame in that upper diagram. 00:01:15.842 --> 00:01:26.153 align:center line:-1 position:50% size:63% The CANbus synchronizes and resynchronizes on every recessive-to-dominant bit transition. 00:01:26.153 --> 00:01:32.759 align:center line:-1 position:50% size:60% All the signals or waveforms I'm showing during this presentation are dominant-bit low. 00:01:32.759 --> 00:01:36.930 align:center line:-1 position:50% size:53% They synchronize on every falling edge. 00:01:36.930 --> 00:01:41.868 align:center line:-1 position:50% size:63% Here, you can see all the sink and resink points. 00:01:41.868 --> 00:01:49.142 align:center line:-1 position:50% size:52% After capturing the waveform, you identify those specific resink points, 00:01:49.142 --> 00:01:59.219 align:center line:-1 position:50% size:58% and then you divide up everything between those resink points into ideal bit times. 00:01:59.219 --> 00:02:04.624 align:center line:-1 position:50% size:43% Then, the scope steps down and takes two bit times at a time 00:02:04.624 --> 00:02:09.296 align:center line:-1 position:50% size:53% and overlays them on top of each other, as you see here. 00:02:14.735 --> 00:02:16.169 align:center line:-1 position:50% size:29% There is one bit time. 00:02:16.169 --> 00:02:20.173 align:center line:-1 position:50% size:55% We can see here, there's jitter at each of the bit boundaries 00:02:20.173 --> 00:02:23.710 align:center line:-1 position:50% size:51% as well as some amplitude deviations. 00:02:23.710 --> 00:02:27.447 align:center line:-1 position:50% size:58% Then, you can apply the pass/fail mask to it. 00:02:27.447 --> 00:02:29.983 align:center line:-1 position:50% size:40% This is an eye-diagram mask. 00:02:29.983 --> 00:02:34.254 align:center line:-1 position:50% size:57% Notice that this mask is shifted to the right. 00:02:34.254 --> 00:02:41.995 align:center line:-1 position:50% size:65% That's because the CANbus samples later during the bit time, as shown here. 00:02:41.995 --> 00:02:46.967 align:center line:-1 position:50% size:61% The CAN FD bus, which we'll get to a bit later, 00:02:46.967 --> 00:02:49.603 align:center line:-1 position:50% size:53% samples closer to the middle of the bus, 00:02:49.603 --> 00:02:56.676 align:center line:-1 position:50% size:53% and so that pass/fail mask will be closer to the middle of the eye. 00:02:56.676 --> 00:02:58.211 align:center line:-1 position:50% size:32% Why is it called an eye? 00:02:58.211 --> 00:03:01.314 align:center line:-1 position:50% size:65% Especially on some very, very high-speed buses, 00:03:01.314 --> 00:03:06.486 align:center line:-1 position:50% size:53% that portion in the middle of that bit time begins to look like an eye. 00:03:08.622 --> 00:03:11.925 align:center line:-1 position:50% size:57% On the vertical axis, what the eye-diagram is going to show you 00:03:11.925 --> 00:03:17.097 align:center line:-1 position:50% size:41% is worst-case differential noise or interference, 00:03:17.097 --> 00:03:21.268 align:center line:-1 position:50% size:53% transmitter-dominant bit-level variations. 00:03:21.268 --> 00:03:28.241 align:center line:-1 position:50% size:49% Different nodes can actually transmit different dominant-bit amplitudes 00:03:28.241 --> 00:03:31.711 align:center line:-1 position:50% size:47% as well as what we showed earlier, the dominant bits 00:03:31.711 --> 00:03:37.250 align:center line:-1 position:50% size:55% that jumped to a much lower level during arbitration and acknowledgement. 00:03:37.250 --> 00:03:41.922 align:center line:-1 position:50% size:51% You can easily see ringing, reflections, non-monotonic edges. 00:03:41.922 --> 00:03:45.592 align:center line:-1 position:50% size:52% On the horizontal axis, it's going to show you worst-case jitter, 00:03:45.592 --> 00:03:50.497 align:center line:-1 position:50% size:45% worst-case clock tolerance errors, bit-time quantization errors, 00:03:50.497 --> 00:03:52.632 align:center line:-1 position:50% size:41% as well as slew rate symmetry. 00:03:52.632 --> 00:03:57.037 align:center line:-1 position:50% size:31% I left one important one off of this list of bullets. 00:03:57.037 --> 00:04:01.875 align:center line:-1 position:50% size:50% The other thing it's going to show you is that propagation delay 00:04:01.875 --> 00:04:05.312 align:center line:-1 position:50% size:54% when bits travel from one end of the bus and back again. 00:04:07.581 --> 00:04:12.919 align:center line:-1 position:50% size:57% This is what a CAN eye-diagram looks like on the InfiniiVision oscilloscope. 00:04:16.690 --> 00:04:22.162 align:center line:-1 position:50% size:50% Between those two red markers there that I've annotated on this display, 00:04:22.162 --> 00:04:25.265 align:center line:-1 position:50% size:40% that is that propagation delay. 00:04:25.265 --> 00:04:32.973 align:center line:-1 position:50% size:66% You can see where bit time is, and I've highlighted the approximate 75% sample point.