TSP #236 - A 77GHz Automotive Radar Module Measurement, Reverse Engineering & RFIC/Antenna Analysis

[Music] hi welcome to the signal path in this episode we're going to do another patreon request we'll be taking a look at some 77 GHz Automotive radar devices and modules this one is faulty so it's intended to be taken apart so we can take a look inside but I also have bought a blind spot detection module for motorcycle this one is supposed to be working so we can actually measure the 77 GHz fmcw waveform that comes out of the radar not an easy measurement to make you need a lot of equipment and it's difficult to actually capture the fmcw waveform but I want to go deeper than just looking at these modules I want to go all the way from the antenna down to the architecture as well as looking at the integrated circuits that operate at 77 GHz responsible for generating and detecting the radar signals themselves so it should be really interesting and really end to end looking at it from top down see exactly how this is built as always thanks to the patreon supporters and PayPal supporters who make this possible as you can imagine these things are not cheap and the instrument to measure them are also not cheap but nonetheless let's go ahead and dig into these now let's take a look at this other radar module which very much looks like is intended for automotive applications as everything is abstracted away into a single can interface where power and communication is managed through the rest of the car we do have a fully sealed structure here with a radom in the front you can kind of see the outline of where most likely the RF elements would be and these things are typically made of stacks of PCB one handling all the RF which would be in the front and then one would be doing doing all the computation and the radar dsv as well as of course the interface to the rest of the car it would be a little bit destructive to take it apart but it should be quite interesting as I want to analyze every aspect of it including the rfic these things are typically glued or Ultras sonically welded to prevent moisture Ingress and they're never really intended to be open so here's the cover which also acts as the radom and if I flip it you can see part of it is thicker than the rest and this is what's in front of the antennas has to be taken into account in the design of the antenna as well as the radiation pattern let's put that aside next we have have this little cover here which covers all the ic's as well as the traces you don't want to radiate I wonder if this thing is RF absorptive or not but interesting nonetheless it covers that also has some structural Integrity then we have the RF board which is the antennas and the ic's we will take a really close look at this later and then after that we do have the connection to the board underneath and in between them is this metal plate here which acts both as a heat SN and also isolates everything from the RF board and on the other side we do have a very thick thermopad here which is going to have very bad thermal resistivity but having this very thick allows you to relax all the mechanical requirements of this and this actually connects on this side of the board so some of the thermal is coming through here but if we remove this further we will see that it does have a whole bunch of ic's on it we will talk about this board also in detail and somebody loves their thermop paths look at that that is that's a lot of thermal material and it's really quite thick as well I wonder why not make this design a little bit more friendly and maybe tighten it up a bit so you don't have to have so much Thal material but nonetheless that's it this looks like some vent here in case pressure builds up as the temperature inside of this changes so that you can have some flow with the filter in there and then that's the connector yeah so let's go take a really close look at all the RF properties of this board now and here's the front and the back of the radar module now there's a lot we can talk about this and if you want to dig deep it will take hours of discussions on all the Minor Details that go into the design of this but even without knowing anything about the chips we can very quick extract some useful information about how this radar system is actually put together so first thing we notice is that we have a cluster of set of antennas these are linearly fed antenna arrays we have four of them that are near each other and three of them that are somewhere near each other so how do we know which is TX and which is rx now first of all we notice that there is four lines here and three lines here so whatever is going on there are seven RF channels in this radar module now how are they distributed between the TX and the RX well we can look on the other side of the board and if you look this is already flipped image so we have the footprint here of the three ic's from left to right in exactly the same way and if you notice on the leftmost IC we have four differential lens that are coming out of the Chip And they're going into this connector and this connector is what connects to The Other Board which does all the DSP as well as the digitization of the if signals so right away we have four differential if ports that are coming out of the chip on the left and therefore this has to be a receivers therefore we have a four Channel receiver here on the left where we simultane L digitizing each of the linearly FED arrays at the same time so in a sense we're doing some digital beam forming digital processing in the back end now on the other side therefore these must be our TXS so we have three TX element so this is a 4 RX 3tx synthetic aperture radar architecture now what do I mean by synthetic aperture radar well we can read a lot about how synthetic aperture Radars work but to put it very simply every time a synthetic aperture radar activates one of his transmitters it listens to all the four receivers at the same time it's not doing beam forming among them it's digitizing them all individually now think of a Target that is reflecting a wave towards this because of the distances between the antennas each of these antenna arrays sees a different delay of the same signal coming back and from that the radar can estimate where the angle of arrival of those signals are now every time one of these TXS is activated all the four rxs listen and then he goes to the next one and then they listen again and he goes to the next one and they listen again and therefore each time one of these TXS is activated you get a synthetic aperture of the receiver and that's repeated over and over again depending on the distances between the transmit arrays so in a way you get 4 * 3 12 different apertures that are coming back so it allows you to grow the radar now the disadvantage of that is that it's slower because you're running each of them independently you have to run one and then the next and then the next and it takes three transmit Cycles Plus three receive Cycles in order to be able to figure out what's going on around in the radar that's pretty interesting it means that you can multiply the TX and RX channels if you spend more time and get a much better angular resolution from the radar that's a lot of information already just by looking at these antennas so then you may ask what's going on with the actual antenna arrays themselves so these are actually beam formers they're not quite phase arrays because the phase difference between all these antennas is the same because it's fixed by the delay between them but it is a phase and array in a sense it allows you to compress the beam and narrow it in the vertical Direction so the beam coming out is going to be pretty narrow because you have so many antennas the delay between this antenna and this antenna and this antenna is designed in such a way that there's a 360° face shift between them and therefore the signal arriving at each antenna is at the same phase 0° and therefore you get a beam at bore side now you also may ask why are these antennas all different sizes there's a few reasons for that but the main reason is that by making the apertures towards the ends smaller the gain of the elements at the end of the linearly FED array is less than the ones in the middle and that gives you some effect of tapering which reduces the side L of the radiation pattern coming out so even these tiny little differences make a really big difference in the radiation pattern of both the TX and the RX and of course the delays between them is 360° at only exactly one phase but because the fmcw radar modulation bandwidth is so much less than the carrier it is essentially almost okay to have the have the delay being fixed among the different channels so then you may also ask what about the fmcw then where does that come into play well every time one of these TX transmits it transmits an fmcw waveform and that gives you the range resolution of the radar so by playing with the bandwidth of the fmcw we play with the range resolution and by playing with the number of elements we have and their distances between them we play with the angular resolution and by increasing the numbers we get more synthetic aperture radar field of view which is a huge Advantage itself so there's a lot of information just from the antennas themselves but it doesn't end there if you look a bit more carefully you'll see that there is something coming out of what we know now this is an RX Channel something's going from here either in this direction or in this direction we can think about that for a second and there's a TX chip in here because one of his channels is used this way but then there is also another line going from this chip to this chip so none of these chips are actually the same they're all different and used in a slightly different way on the other side of this port we have this chip if you look at this chip this is an fmcw generator but it's not generating the fmcw at the RF carrier frequency it's generating it at a much lower frequency it's essentially a PLL that's wrapped around whatever is inside the chip in the middle therefore this is our Master fmcw generator and the PLL inside and the vco inside generates the ramps that we need with the support of this guy creating the FM modulation required to do that that also then and therefore right away tells us that the signal must be going out of this transmitter chip to the left and the right now why would you do that well if you send a signal to the right then you can use this chip as kind of an expander and create and take this signal and allow it to go this way and this way as well so now you get two additional channels but you're using one of the channels of this one to accomplish that and the remaining Channel goes this way and then it also has another channel that you can use to send to the RX chip and now the RX chip receives a coherent phase locked low signal as well coming from this one so it can downon convert the signals from this so we are expecting to find three different Ric's one of them has the PLL and the PA for this one this one is just an expander and this one has the down convert mixers as well as of course the lone noise amplifiers and everything else that receives the signal from this and again I'm doing all of this but just looking at it I know nothing about these chips that are there just by visual inspection now on the other side this is an fmcw generator this is our plore so it needs a crystal and and indeed it does have one this is a 50 MHz Crystal the rest are just basic support circuitry I don't think there's anything unusual interesting about it just the if signals all going out into the connector so we have to look at the other board now we can also take an x-ray of this but there isn't much new information we would find here because we're seeing all the cool things in the top and the bottom layer but I do have the X-ray here which is now matched to the board on the right so you can see them one to one if you look carefully here you can see very very thin wire bonds ins side of this chip that's because this chip is running at low frequency so wire binding it is perfectly fine whereas these chips here which are now on the other side of the board they have a BGA footprint because you're dealing with 77 GHz that are coming in and out of them if you look carefully you can see the antennas here on the other side of the board and there's the receive antennas on the other side of the board you can see through the crystal which is kind of cool the connector is visible all the capacitors are visible but nothing really unusual and it's important to note that the reason they have to do this is because this board board actually only has one layer of very high frequency high performance dialectric only this one it's probably some Roger material is the one that has to see the 77 GHz signal all the high frequencies rather than one layer this side of the board is a much much cheaper material you don't want to make this sport from the same Roger material otherwise this cost will just blow up so by by playing with where they put the chips how they route the PLL where they put the antennas and how they interact with all of them they can reduce the cost of this just from that simple design all these other separations that you see for example here and here these are just isolation to make sure that these elements don't talk to each other and some nice grounding between these layers over here a lot of work has been done and this area is essentially supposed to be completely covered because in the radar as you saw you don't want this area to radiate this would just interfere with the antennas you want the aperture that interacts with the physical world to only be contained in here that's where the radom is sitting in front of it as you saw during the thir so there is so much information just by looking at this board without knowing anything about the chips but of course I do want to know something about the chips so we're going to have to take them off the board and take a closer look at them so let's also take a closer look here live because I think there's some cool stuff I don't want to miss so for instance if you look over here these are our receive elements as we saw and they're single-ended right there's a single-ended signal coming over here connected here but interestingly enough all the other TX channels are differential and what you see here is a balance is a differential to single ended inverter I think they have made all of the PAs to be differential in order just to help the isolation from inside the chip so everything is differential on the TX sites but everything else looks pretty normal as to be expected now we said that we want the phase from here to here to undergo 360° now this material over here is dialectric constant is probably around 3.4 or so which would mean that the distance between two of these patch antennas would have to be around 2.1 mm in order for it to represent 360 ° and we can make that measurement and see if that's really true so I'm going to measure from here to here roughly which is a center from Center so if I adjust right there what look at that 2.13 mm which is exactly what it's supposed to be so it lines up with that being there if you look at the distance from here to here which is one of the antenna centers to the other antenna Center that's going to give ourselves some angular resolution information about the radar and that's around 2 mm also so it's roughly Lambda Lambda in both directions uh which is interesting to see distance from here to here I think is four Lambda and from here to here I think it's five or six Lambda I forget exactly but it's pretty obvious then from there that everything we said should be mostly correct now one other thing I wanted to mention is that I said that this is a synthetic apure radar you could make this into a myo radar where all the transmitters are on at the same time but in order to do that you'd have to code them if you code those signals then you can listen to the receivers at the same time and you will get the signal from all three with a different coding and then you can undo that in DSP later on and myors are also of course been studied quite extensively so nothing unusual in new there but I I would be shocked if this has anything like that I'm just thinking most likely just a synthetic aperture radar architecture and here's the other side here's a crystal definitely 50 MHz like I said this chip is our PLL part we talked about differential lines you can see here some pads for testing and most likely some Factory testing yeah looks nice you can see the material is different on this side and on this side this is the you look at it over here if I could focus on it which might be difficult but if you look at across of the board you will see that the material is not the same I can't focus that far up but you can see the top layer is a different material than the rest of the layers so LR lines up with what we were anticipating very cool and while we're here we should talk about the digital board as well and see how the signal flows through that one so here's the main connector this connector ctor connects to our RF board and we expect four if signals to come out of it and in fact it does here's 1 2 34 four pairs of differential if signals which now need to be conditioned and digitized in order for the radar to be processed and the very first thing we see is a TI AF 541 which is a quad Channel analog front end specifically for automotive radar essentially has four low noise amplifier and programmable gain amplifiers to condition the if signals independently from each other and then feed them into adcs now the adcs need to also run in parallel you cannot time inter leave or sequence them because you need to digitize them at the same time so we have 25 Mega sample per second adcs one for each of these channels and they're all running at 12 bits which means you have 100 Mega samples per second at 12 bits that has to come out of this and be processed in somewhat real time in order to get real time radar information and therefore the data coming out of this needs a lot of processing and on the other side we have another TI chip which I actually couldn't find much information about but this is responsible for getting the data from those data converters and there's some Dam next to it to store it in real time because you can't lose anything and then process it and interpret the result basically doing all of these synthetic aperture calculations and radar to extract the scene information that the radar actually sees then this guy needs to talk to the car but he doesn't do it directly he goes through this guy which is a risk processor specifically for automotive application and it has a canb and variety of interfaces built into it and this guy talks to disconnector which has the canbas on it so then that completely basically wraps up the signal flow coming to this we have if signals coming in digitized simultaneously fit into the radar processor stored calculated interpreted the result sent to the risk processor the risk processor then talks to the canvas of the car in the language it understands and the entire radar gets abstracted away from the car's point of view this is somewhat different than some of the other kind of Radars that that may be used where they have a central processing this guy does not send the raw radar data at least I don't think so it just sends the interpreted data all of it is calculated already in here and that's why this is the highest power consumption component on this board that's why it has the biggest heat sink yeah that's about it very cool everything else on this board is pretty boring dcdc converters and so on I think unusual I think there's some flash memory in here perhaps for some firmware and yeah pretty cool all right I have successfully extracted all the three dies they're sitting next to each other here under the microscope I extracted them at the same time so we have to kind of ReDiscover which one is what although I can see the part numbers are actually on the D too so let's start from the main transmitter the one that has to have the pl up we said that this one has to have a vco in it and it has to create the chirp in conjunction with the other IC we saw at the board and here is our vcl core you can clearly see it and it has a tiny little tunable area we'll zoom into this a little bit later but from the top here's our vco we do have some supported circuitry here probably dividers and perhaps even multipliers in order to get the signal to where it needs to be and as well as providing the Divide down signal to the external CHP generator so the signal comes out of here and if you notice over here it splits into three and once it splits into three the signals are all identical and they all carry the fmcw waveform leaving the chip one goes down this way and one goes down that way and the other one goes down that way and they're all length matched now you may or may not have some kind of gain control as well as face shifters and so on in each of these those are Minor Details depends on how they want to use the chip but the architecture is pretty straightforward at least at this level vco and the supporting circuitry and the Pas that drive the the paths and if you look carefully you can see that the outputs are differential just like we saw on the board and there was a Balon on the PCB that converts it to single-ended now we also saw that this master transmitter uses one of its outputs to go to the antenna but the other two outputs are used up one is sent to the receiver and the other one is sent to the the expander chip and the expander chip allows you to gain additional channels so the expander job is simpler it doesn't need a PLL it doesn't need a vco anymore all it needs is to receive the signal at RF redistribute it amplify it and create multiple outputs out of it and I think that's the chip here on the left side there it is and if you notice in this one we do have an input here it comes over here it splits into two these are Wilkinson dividers and it goes left and right and that's how we got our additional two channels so we had one channel of transmitter from here two other channels of the transmitter from the other circuit and that's the three transmitter channels so the combination of these two chips make up a transmit section now interestingly enough this portion here is the same as this portion It's Curious that they have included this now there could be a few reasons this might have something to do with being able to receive the divid down signal from any of these chips or it could be something to do with power detections or testing it's not clear I have to dig deeper into this to really understand why why they have this there interestingly enough this is also present in the receiver whatever function it has it helps them combine chips and create more different kinds of radar modules now if you go all the way to the other side we will discover the receiver which arguably is the most complicated in the sense that it also has the basan if sections in there now in this one we get our loo signal arriving from here if you notice this is also differential goes over here gets split into two into these big amplifiers there and then it gets split again again now these could also be face shifters it is possible that they have pH shifters embedded in there and it splits into interestingly into eight different paths two here two here two here and two here and then so this is the front end here's the input number one for the receiver channel one receiver channel two receiver channel 3 and receiver Channel 4 and they're all single-ended now why is there two allo arriving at two different places I'm not sure I wouldn't render an opinion until I've looked at it closer but certainly these are the mixers as as well it could be that there's two different modes of down conversion perhaps one with Filter one without or maybe they have two different gains or maybe there is one with mixer first and one without I'm not sure we'll have to take a closer look to find out but nonetheless you can see the four identical paths there that creates four if signals that we saw goes into the data converters and gets digitized outside of these chips and here's that section again repeated once again over here and it's fed directly from the L input so whatever is happening is acting a on Theo just like it is in the other circuits so this is replicated so it's cool to see that from the top down at least your architecture is fairly obvious this is a silicon germanian process infinia own internal process we've seen this actually in their 24 GHz radar where I did a chip analysis of that one too and there we saw very similar circuitry of course that's at you know one3 of the frequency or so but nonetheless we can zoom in at least and take a look at a few of these blocks let's try the vco and here's the zoomed in view of what I believe is the V vco here's the core of the vco the tank inductor is tunable via these fuses which I think are burnt out by using lasers and then the signal out of that has to leave this and then get either multiplied or divided depending on what frequency this is we can actually measure the dimensions of this tank and get a rough estimate of what frequency it is operating at whether it's at half the frequency or at the fundamental 77 GHz frequency nonetheless the architecture of a very basic radar like this is not so complicated and here's a close look at one of the trans Pas now this is in diic microscopy and I'm going to go through the prism and take a look and see the details that gets revealed there's some interesting aspects of this so right over here you can clearly see the indents on these pads which are the vas going from the pad down to the lower layers if you go through some more you can see all the imperfections as well as height variations across this so you can see that there is a dip here that dip is because the pad opening has been there this probably most of it is also affected by the acid but these distances we're talking angstroms here of course we can resolve it with the DIC microscopy because of the fact that the wavelength of the light is proportional to the distances we're here trying to measure so a lot of really neat details come out of this by using this technique now because our Automotive radar module is faulty I still wanted to be able to measure something at 77 GHz so we could potentially see the fmcw modulation and look at the bandwidth and everything so I went ahead I bought one of these this is a blonde spot detector for motorcycles you mount this in the back of the motorcycle if there's somebody approaching you from left and right behind you where you cannot see them it will illuminate one of these two lights which you can mount in front of you so you know who's in your blind spot many cars already have this mountain on mirrors and in this case this is supposed to also be 77 G it's a very very small aperture we're going to try and measure the signal coming from this and of course also take this one apart I have the module over here with the aperture pointing towards the horn antenna it doesn't really matter whether we're in the far field or near field I just want to get the signal we're not trying to characterize the actual radiation pattern or the power coming out the Erp coming out now I've aligned the horn to be the same polarization as the module now you may ask well how did I figure that out it's really simple you just rotate this by 90° and you keep the one that has the highest power that's the correct polarization then we have a smart mixer over here M1 1970e which is going to work between 60 to 90 GHz and this is in fact a 77 GHz which fits perfectly in that region and then we have that connected here all the way to the key side mxa to analyze everything and then we have all the power cables and everything just going to a simple power supply here giving us 12 volts which is what you would get from say a motorcycle or a vehicle where this could be mounted on now this thing did not have an FCC ID or anything like that so we don't even know if this is a 77 GHz module or what kind of radiation it has but at least we should be able to capture something and you'd be surprised it's very difficult to capture a radar signal like this especially with mixers as we will see why but at least we should be able to get some ideas and here's the display of the spectrum analyzer I'm using a Wi-Fi to just look at the display remotely so the update rate is quite a bit less than what I see on the unit itself but it should be good enough for our purposes so right now we're looking at a center frequency of 77 GHz as you can see and a span of 2 GHz this is around the automotive radar band let's see if anything comes out I'm going to turn on the power to our radar module and let's see and check it out there's definitely something going on now this is a very difficult signal to capture as I said for a few reason first of all we're looking at quite a bit of loss even at that small distance although we have a lot of signal to noise ratio left here the signal is also moving very rapidly as as the chirp is being generated so we don't exactly know how that chirp looks like we're going to try and see at least what we can extract from the setup now in this particular case I have the mixer connected if I go over here the input and output you can see that the external mixer is right over here but the issue is that the signal ID is turned off meaning that we don't know what's image and what's real although you could play with the yellow such that there is no images if I turn the signal ID on most of the time the signal will not be captured occasionally you see some tones here and there and that's because when you turn a signal idid on the instrument has to capture multiple waveforms in order to figure out what was real and what was the image and a signal is rapidly moving around which makes the detection quite a bit harder there is no pre-selector remember in the smart mixer so we're going to turn it back off so that we can at least see something now this instrument does have a lot of realtime capabilities and quite Advanced SPS actually but it's digitizing bandwidth is only 160 MHz and you can see that this radar chirp is much wider than 160 MHz so there's no way to real time digitize it using this instrument there are Spectrum analyzers that can do this that have two four even 8 GHz of real time bandwidth we're talking about the instruments that exceed $250 300,000 and this is not one of them although this is still a fairly expensive unit so let's go over here and change the mode and then look at the real time Spectrum analyzer this unit has all the options options so there's a lot of cool measurements you can do but in the realtime Spectrum analyzer we can look at the density of the signal coming out if I click okay over here it would switch to a density view right now it's looking at 160 MHz and within 160 MHz at 77 you see that there's a whole bunch of activity there but the signal is wider we can do some stitching this unit can actually Stitch multiple signals at60 next to each other and there it is you can see the individual stitching sections so there's a lot of stuff coming out of this radar it does have some unusual Peaks that are weird they shouldn't really be there there's a gap here which I don't quite understand it might be a measurement artifact because it's so sharp around one of the Bands but it tells us that the signal is jumping all over the place and it spends most of its time being roughly constant in amplitude the fact that you see some deviation here is because we're measuring over the air and I don't exactly know what the channel looks like but you know it's reasonably flat so it should be transmitting a constant power it still doesn't tell us anything about the fmcw W from itself and how good that CHP looks like it just tells us the frequency that it is occupying now one other thing we can do is we can plot the power versus time and the advantage of that is we can see how constant the power is at least within a short bandwidth so let's put this back to 77 GHz at the center here and look at that it this is now versus time not versus frequency so it tells us that within the60 MHz of bandwidth centered around 77 GHz the signal is present for this duration and and it's gone it's present again and it's gone and it's likely that in this dead zone over here we're basically in the frequencies outside of that Center frequency all of this makes sense and you can see how hard it is to make a proper measurement I cannot tell you anything about how good that fmcw is with the setup we have we're going to have to do something different but at least we know that this instrument is definitely doing what it's supposed to do it is looks like it is doing an fmcw it looks like it is operating in the automotive band and whether it meets the FCC regulations and Emissions that's a whole other problem now you can in here also do something you can go over here and you can go over the trace and enable Max hold and let this waveform build up there is something weird going on here it looks like there's two sections maybe perhaps two different plls that have to do the chirp interesting I think the tear down will certainly tell us more but there are some places where it puts out more power than other places like I said it's not obvious whether this thing meets the FCC regulations or not nonetheless very cool that we can do this measurement now I'm reasonably convinced that these addition of Peaks are probably PLL startups where they try to lock into the signal and then try to stabilize and perhaps the amplitude is not settled yet that's my guess you have to look inside and see how it is put together but this could be one reason if I zoom in a bit more here and look at within the digitization bit of the instrument to get a fast measurement you can kind of see this Ripple and it settles down and it goes flat I feel like that's what's going on it's a difficult thing to measure as you can see we need something to show us versus time but even within 160 megaherz we're getting some interesting information so we looked at the signal coming out and looked at on the Spectrum analyzer we saw that there's a certain amount of signal around 77 G as present but we didn't really verify if it's really a chirp we saw that it's moving around but we couldn't make sure that it is following some kind of a triangular shape well we can verify that in a different setup we can use a tectronic 6 series or the Five series that have the digital down conversion with frequency phase and amplitude tracking built into the firmware of the oscilloscope these oscilloscopes are really quite unique and if you look at my reviews you'll see what you can do with them it opens up your eyes in new ways of doing measurement so I've changed the setup a little bit now we're going to use an external mixer this is no longer a smart mixer it's just a classic diode mixer and we have to give it an Lo signal which is coming from this tectronics box here the loo is at 114th of the carrier frequency so the noise figure of this mixer is really really high the signal is not going to be great and his if doesn't have enough bandwidth for us to see the entire chirp but it's enough for us to make sure that at least part of the chirp is present if is Amplified and then fed into channel five of the oscilloscope and over here we're going to take a look at the frequency as well as the frequency versus time to see if you see any resemblance of any chirp so here at the top we have the Spectrum view from DC to 500 MHz the radar is right now turned off so what you're seeing is the shape of the noise coming from the amplifier itself and this is our time domain signal which is also noise and at the bottom is frequency versus time which is of course also noisy because there's nothing present there's no dominant tone I'm going to go ahead and enable the radar turn the radar on wait for it to settle and look at that we're getting our fmcw chirps now we're only capturing a very small percentage of it because of the noise of the system but I'm going to take a single capture look at that look here's a part of it an up ramp of the chirp and here's a little bit of leftover of the down ramp of the chirp the reason you see both is because the mixer doesn't have image rejection so sometimes you're catching the signal below 77 and sometimes above 77 so it moves towards it or away from it but look you can clearly see that this is definitely creating some kind of a chirp and you could in theory capture the entire chirp with this oscilloscope because it has 2 GHz of analysis bandwidth but unfortunately the system has too much noise you need a much better mixer and front end amplifier to do this but look it's isn't amazing that you can use an oscilloscope to do this kind of measurement a really unique instrument from tectronics so this tear down is also somewhat destructive the module is filled with an interesting putty this blue stuff over here that has been injected from a hole in the back once everything's been put together to prevent moisture from getting in there's also an O-ring that was around the cover that was all glued in place now I'm going to keep this radar a little bit more secret because this is a very cool design and it's a very different kind of module design and I think we're going to cover it in a different video so if you want to see it you definitely have to subscribe to the channel we will do the same kind of reverse engineering on this one and there you have it I hope you enjoy this extensive look at these modules make sure you subscribe so you see the tear down and Analysis of the other module we did take a look at in this video that one's totally different and it has some really cool engineering aspects to it too thanks again to the patreon and PayPal supporters as always none of my content is ever behind the pay wall and I only put a fraction of the videos actually on patreon charge them to patreon but they're always available for anybody to watch them your support is the support of the entire Community as a whole I'll see you next [Music] time

Channel: The Signal Path

Views: 90,718

Keywords:

Id: 7ideoJlr9Ko

Length: 33min 25sec (2005 seconds)

Published: Fri Dec 01 2023