Reliable, Practical BroadR-Reach Decoder/Analyzer Solution

High-performance automotive networks using the BroadR-Reach 100BASE-T1 Ethernet protocol are increasing with the growth of vehicle features that require advanced data-processing capabilities. Unfortunately, capturing and analyzing these signals can be costly and complex. But not anymore. Pico Technology has introduced the first PC-based oscilloscope solution that’s straightforward, accurate, cost-efficient and non-invasive. Watch our video to learn how it works.

Transcript

00:00:04 hello my name is Trevor Smith I'm the Business Development Manager here at Pico technology and I'm joined by my colleagues and Tom khaliqul and Patrick Alec and we'll be introducing in this video the the picoscope automotive Ethernet decoder analyzer so if you guys would like to take a seat for a couple of moments we'll give a little introduction so first of all a little

00:00:35 look back at what's gone on in the automotive industry over the past two or three decades so back in the back in the day when I first started driving my car then each component in a vehicle the brake pedal had a little switch on it that triggered the the rear lights to go on or trigger the brake lights to go on when I press the pedal that all changed in the 1980s when wiring looms in

00:01:00 vehicles became too complex and too heavy and we saw the first introduction of automotive networks the original networks were point-to-point networks so each node was connected to the adjacent node individually it's it gave some improvements in terms of the weight and the complexity of the networks that were involved but still still not really perfectly the wiring looms were actually

00:01:31 quite complicated back in those days and there was no good standardization from from one manufacturers network to another so as we moved into the 1990s we saw the the development of us-based networks and where each node was connected to a bus type structure and any node can could be connected or any ECU could be connected to any other ECU on the system there were various

00:02:02 different classes of networks ranging from low speed devices for convenience features such as turning on the the heater in in a wing mirror or something like that right the way through too far much faster devices for character for functionality to do with the control of an engine so looking at air speed indication rotation speeds and feeding

00:02:31 all of that into the engine control unit so those those networks developed some quite strongly throughout the 1990s and in particular there was one network that became dominant in the industry and it was the kin based network so kin was developed in the mid-1980s but became widely deployed in vehicles throughout the 1990s and and beyond most vehicles on the on the roads these days have got

00:03:03 some form of kin base networking in them standardized by the ISO organization and kin is a two wire system it's a differential bus it has a can high and a can low signal so it can bus is immune in the presence of interference which makes it a robust network for a wide variety of applications more recently has been the evolution of can FD can flexible data which gives us up to 10

00:03:36 megabits per second data transmission and with payloads in each packet of up to 64 bits so can has served us quite well for a number of years there are other standards that have commonly been used as supplements to the to the canvas architecture so Lin is something that's popular for for low speed sensor and actuator applications the the the waveform that we're seeing on the screen

00:04:09 here is a sent bus waveform sent is single ended nibble transmission it transmits four bits of information with with each edge there and it is used extensively in sensors to do with and management and position control so throttle position indicators mass airflow sensors temperature information that's being transferred typically to the engine control unit and the majority

00:04:46 of vehicles on the road these days are a mix of can and one or more of these other networks so those that those architectures that mix of architectures has served us serve the the industry rather well for 10 or 20 years but there's a fresh set of expectations and demands that that we as as users and drivers of these vehicles are expecting so the next generation of vehicles

00:05:17 really will become in effect mobile data sensors and they will process huge amounts of information so there are all sorts of sensors to detect proximity to obstacles and other vehicles and that sort of things so there's a whole range of driver assistance aids that that are being deployed in vehicles now that use cameras radar lidar and and other sorts of sensors so there's there is a big

00:05:53 move to and particularly as we go towards the adoption of autonomous vehicles there will be a huge increase in the amount of data that these vehicles will be using and processing so the the industry needs to adopt a need to find and adopt a new standard and of course it's fairly widely reported that the the standard of choice that the industry has is clustering around now is

00:06:25 the broad our reach automotive Ethernet it's a standard that's been adopted by the I Triple E as as the hundred base t1 standard commonly commonly known as and it is something that is from voted by the open of Alliance brought our reach physical layer transceiver specification the beauty of broader reach is that it is a form of Ethernet that takes

00:06:54 advantage of the wide that the widespread use of Ethernet technologies that are already around in non-automotive applications so the the idea of broad our reach is that it takes existing Ethernet technologies but puts them into an automotive environment and in particular what's needed is to to get the rather than the multi cable looms that that the car manufacturer the

00:07:26 vehicle manufacturers are trying to avoid is that it transmits data over a a single unshielded twisted-pair copper cable so in fact this is this is kind of how the the communications network looks like between two transceivers so for instance a master on the left and a slave on the right so the the section over here would be a traditional ethernet hardware on both sides in fact

00:07:59 but the the defy the the physical layer implementation of broader reach it takes us to a world where we have just a to two twisted pair wire that's transmitting the information of the ethernet so gone are the days where we had a cat5 cable with different strands of copper cable in there some transmitting some receiving that's that's not the way that the the broader

00:08:32 reach automotive Ethernet is implemented instead really what we have is just a a simple single unshielded twisted wet twisted copper pair of wires to transmit the data from from trans feed transceiver to transceiver in the Ethernet system so that gives rise to certain problems when it comes to testing and integrating a working system because the the Mac and

00:09:06 the fire device will typically typically be implemented on the same chip so if we want to if we want to observe and see what is the information that is being transmitted in both directions then we've got no way of probing in inside that section of the other device so the probing points that we have really are at the the master and the slave devices rather than inside the the the ethernet

00:09:38 device is further down the system so that's that introduces one or two problems and and some some challenges from a test engineering point of view and a design verification point of view that we needed to address with the the picoscope from broader reach decoder and analyzer so at that point I Allah I'll hand over to Anton to take us through the the decoding mechanism

00:10:09 thank you very much Trevor now as you mentioned there are a few challenges that are presented with decoding alternative Ethernet but let's first understand the signal that we're looking at automotive Ethernet uses Pam free data encoding for each of the Uni directional components from master to slave or conversely from slave master Pam 3 means that the signal is composed

00:10:42 of three level signals the symbols so each symbol can be minus 1 0 or plus 1 volt and each pair of ternary symbols is going to encode 3 bits of information so there are nine possibilities for a bit state of two ternary symbols we're going to not use one of those folks tapes and we'll be left with eight possibilities which can encode three bits of data and in turn we're going to

00:11:17 slice up the parallel data going between the Mac and the five chips into three big chunks and encode them violet pam three encoding so the you directional components of of this pantry signal from each transceiver to the other are going to interfere across the channel and you're actually going to end up with a waveform that looks a bit like this if you probe somewhere along the

00:11:49 channel this presents a challenge because if you are a transceiver on the channel you know your own contribution to the signal and you can subtract it and infer what the other party is trying to tell you however if you are a passive listener you aren't able to figure this out and subtract that contribution therefore we have to develop a technique to separate out the directional

00:12:27 components and in practice out there there are solutions to this problem that involves inserting a a Hardware directional coupler in the link but these solutions are costly and they're also very intrusive because they change the characteristics of the channel we realized that there is an alternative here by probing up to different points along the channel we can we can use

00:13:02 frequency domain analysis to work out which components of the waveform are traveling in each direction so this this directional coupler feature is presented in in the form of a typical map channel and picoscope albeit with a with a few more parameters than usual so so in order to define directional coupler which is going to extract out the signal going from probe point on

00:13:46 Channel eight probe point on Channel G we also have to specify the cable length of the cable type and the cable length because we use the velocity of propagation in the particular cable type and also the separation between the probe points in order to to do our analysis and given this information we're able to go to the frequency domain as I said work out which component is

00:14:24 going in which direction and and reconstruct the directional signals and I should also point out that getting the signal in the opposite direction so from point D to point eight all we have to do is reverse these arguments so after you've done this you will get a pan three signal and the separated waveforms going from master to save or save to master along the same channel can be

00:15:04 extracted then they can be decoded using our our brawler each decoder which my colleague Patricia is going to tell you a bit more about so as Anton said we also provided the decoder for you Prada reach decoder if you haven't used before any of our decoders from articles Paulette in order to set up a decoder which is for instance brother each one you go to

00:15:38 tools from the top ribbon and you select a serial decoding option and it should pop up a dialog which you can by pressing create from the display list you can select the one that you're interested in which for instance in now would be like broader reach and by clicking this it should also brought up a settings dialog which is required for further setup of the decoder in order to

00:16:15 decode the data correctly so the first part specifies the signal which would be the decoupled math channel signal further there are configurations which are if you don't do anything it should calculate it for you but if you want to tweak it even more there's a high threshold and low threshold that defines a high high value of the digital level which is 1 and middle which is 0 and

00:16:49 below the low level would be minus 1 so that's your pan 3 digital beats and also you need to specify whether the decoded signal is a master signal or a slave signal because they are using a slightly different scrambling scrambling technique so that would give a hint to the decoder algorithm which one to use so this is a basic a packet broader rich packet that was decoded from the data

00:17:24 behind it as you can see the very beginning is the what is called preamble which is kind of a synchronization like can't shake synchronization pattern for both master and slave or xever to transceiver synchronization route which at the very beginning when the pin which is called TX enabled goes high it inserts six consecutive zeros

00:17:56 which indicate the beginning of the packet beginning of the transmission so the broader reach decoder decodes and decodes the packet which contains information about MAC address of the source MAC address of the destination of some payload eater type which also defines VLAN information etc etc and in the end there is a CRC for byte CRC definition or information which is

00:18:31 validating the packet whether it was sent correctly or not that's basically it you have additional features like you can select which fields to present or not depending on which which information you're interested in by selecting new fields or you can adjust the font and depending on your preferences also you can filter packets that you're interested in

00:19:04 depending on the columns so depending on the type of information you know and as you filter out the information that you you have you know you record that you can also search through different kind of information like for instance in in here with that inter type that we are interested only in this value showed here and you can by clicking these arrows back and forth you can navigate

00:19:33 through your filters for filtered table so that's basically it's a back to Trevor Thank You patty so we'd just like to finish off by saying what are the hardware requirements for the for the broad are each decoder analyzer and actually they're pretty simple the the only hardware you need is the is a picoscope 6,000 any

00:20:00 model in the range from 250 megahertz up to the the 500 megahertz models so take your pick of those and you need to of the ta o 4 5 differential probes what one each for measuring each end of the the test points on the broad are each network that we looked at before so if you're lucky enough already to have a picoscope 6000 and the differential probes then there is no additional

00:20:28 investment required at all to to start looking at your your broader reach communications so we'd like to finish then just to say that the broad are each analyze the decoder now is included as standard with the latest version of picoscope 6 software that is available free of charge for download from our website and if you do have any additional questions then we've written

00:20:57 a little section on the A to Z of picoscope on our website and you can find that at www.att.com/biz to see the a little bit more information about the material that the land horn and Patty have just presented thank you