Very rarely when a TH-camr says something will be interesting and exciting, it actually is. I'm excited everytime I see a notification for a Signal Path upload.
What I love from your videos is that even though the concepts are advanced, I still end up learning something. In this case, I got an understanding of the doppler effect that I've heard about many times.
In case anyone is wondering: The circuitry on the second board was some bus related stuff for CAN, a high performance TI Sitara, a DSP IC, DDR3, and the power circuitry. I'm a dumbass and pretty well destroyed the 2nd board (as you can see from the connector w/ random PCB bits attached) hence it's lack on inclusion in the package.
Absolutely fascinating. I'm gutted there were parts you couldn't talk about but its understandable. I love it when people explain RF circuitry it's way above me but you explain it so well even I get to learn stuff. Thanks for the marvellous videos.
Great episode Shahriar. great explanation for the microwave structures on the PCB I have to say this is a bit old system. The new systems the chips are packaged with technique called eWLB and they are soldered on the PCB like BGA Furthermore, the new generations the frequency generation is integrated with the TX front-end
Thanks for another great video! Could you please create a separate episode explaining how those magical microwave PCB components work? I know that you mentioned them at different occasions, but it would be helpful to have one video with all the knowledge in concentrated form.
Awesome video, thank you for all your work here. One thing I missed was at 11:29 - "Infineon {something} CMOS... so they're just {acronym} process". Thanks again!
Shahriar, at around 7:49 when you’re explaining the ramp up and ramp down, isn’t the graph on the left representative of what would be received if the object is moving away from the vehicle/transmitter? The picture shows the object moving towards. Based on the doppler effect a pulse should be frequency shifted up if the object was moving towards the vehicle, but the peak of that graph shows the received pulse as a lower frequency than the transmitted pulse. Is this correct or am I missing something?
Dear Shahriar, At the unused end of the coupler and the edge of unused lines there should be a 50 ohm termination. I don't see any termination there, please explain where is the termination or how it is going to be 50 ohm?
switch the SA for a scope when looking at the output of the mixer. For those interested: get a cheap HB100 10 Ghz unit through ebay and experiment with doppler radar yourself. Oh and btw I very much appreciate your video's.
Nice tear down! That antenna is a Taconic laminate. The Taconic rep brought it by on a visit. What's odd though is that the transmission lines are oriented parallel to the weave; IIRC the sample I saw had them at 45 degrees to keep it more homogenous. Those WR12 to microstrip transitions are pretty narrow band, but the don't require a back-short. I'm using a differential microstrip to WR15 on a board I just finished. Need a back shirt but much wider band, and can couple to the TX chip without a balun. Those rat race baluns aren't really baluns if you think about it, since there is no way to fully convert a pure differential signal to single ended with that structure; still requires a good amount of ground currents for the differential side. Notice lack of termination resistor on the branch line. They must be tapering it down to such high Z0 that it just leaks off. Anyway, good stuff.
The Signal Path if it's a purely differential signal on the balanced side (i.e an infinite Zcomm) there is no way for that signal to induce a current in the ground plane for the single ended side. If you sliced the ground plane beneath the differential lines it would fail; at least that's what my EM simulations showed. The output stage of that TX chip can probably be run either single ended into 50 ohms or differential into 100 ohms, so it has a fairly low common mode impedance, which allows it to work with that rat race balun. A Marchand balun should work since it relies on resonance, and a tapered line balun, but then you'd need vias to the bottom of the board. You whizzed by me on the floor at IMS (either Phoenix or Baltimore), and I was thinking I have seen that guy on TH-cam. Didn't get a chance to say Hi. Maybe at another one. Anyway, great TH-cam channel.
Yes, I agree that the output stage is likely quasi-differential and can provide signal into 50-Ohm single-ended. I am sorry that I missed you at IMS! I hope we can meet next time. :)
It is totally possible to convert a fully differential mode to a single ended mode by tapering the differential microstrip line to two single ended microstrip lines first and then combining then with ratrace couplers as done here! I work with similar stuff at a radar research institute in Germany and we do this on a regular basis in order to reduce parasitic bondwire inductance when compared to single ended bond wires. Funny, I ran into Shahriar at IMS back i Florida a few years ago when I was stil, a master student. He was really nice to talk to. I dont know if you remember me, but it is nice to see that you are doing well! We are also working with Infineon silicon processes.
The ADC is a 6 analog front end channels multiplexed into a single ADC, so it is not a simutaneous sampling system. How to handle this in the system point of view? Thanks a lot!
Interesting test equipment with the extra sweeping tail. Thanks for the very interesting video, it's a pity that you couldn't talk about the signal coupling to the antenna array. :-)
Hi, I'm willing to buy two of these radars to do TxRx experiments (no FMCW nor Doppler involved), but I'm afraid it will be tricky to obtain the measured reflected/transmitted signal since there is lots of tightly-packaged complex electronics involved. Do you think this is possible? what are the drawbacks and difficulties you can expect to find in this project? Thanks for the great video!
Just a quick thing to point out. At 9:10, you point out that the top cover has to be plastic to be transparent, but as I understand it, you want to have the thickness such that it works as a 1/4 wave transformer. The plastic is better for reducing attenuation through the medium though, which should in theory mean a wider bandwidth, right?
Nope. In most doppler radars, they will use an additional mixer between the CW source and RX mixer with an IF clock to raise the RX IF baseband. That way, you'll be able to see both up and downshifts in freq.
hey i have just started to enter into this radar thing , found this video and to my surprise this is awesome , even though i am not able to understand any of the technical components that you were mentioning (dont even know what is a coupler ) i am enjoying it . can u mention how do i start of from the basics to fully understand these radar things so that i can be able to build my own FMCW radar
How do you have all of these equipment at home? This could be more than a million dollars... Could you explain how you got them? Sponsorship? Remote work for a company? Do you own a company?
Hi Shahriar, I love this video. I have been hoping for a while that you would do something on FMCW. Do you have any plans for more videos on this topic? As a golfer and an engineer, I am very intrigued by radar devices (such as the "Trackman") that can track a golf ball's flight through the air and measure things like the rotational velocity of the ball, distance of travel and many other things. It would be cool to be able to do some experiments with stuff like this, unfortunately a lot of the components seem out of reach cost wise. Do you have any tips, or resources you recommend, on how to experiment with radar in the 10.5 GHz range without breaking the bank?
17:35 Does any one knows more about this topic under NDA? Please refer me to some books/articles or at least keywords needed to lookup in order to understand all these goldy structures around the ICs. I am digital ASIC design engineer with FPGA background and I enjoyed undergrad Electromagnetics, I believe I'll find a way to understand this beauty. Thanks!
Next time don't use an SA to show the result, just hook up an Audio amplifier. It just beats anything else to experience what happens when you wave the pcb above it.
Yes, I agree. An audio amp hooked to a simple 10 GHz gunnplexer will teach you all you ever wanted to know about Doppler effect. Take it out and point it at some moving cars, or a rotating fan blade, or interestingly, a lit fluorescent tube. The tones you get back are fascinating.
Nice video. However, you began by saying that there is a “frequency shift” in the received ramp from a stationary object. Yet at the end, it’s clear there is not shift in frequency when the copper piece is stationary. There is however a “delay”. Good job, keep up the nice work.
There is a frequency shift in a FMCW radar system from a stationary object because the input signal is FMCW. The delay translates to frequency shift with respect to the transmitted signal. In my test the input signal is CW and therefore no observed shift in frequency, only Doppler shift.
This is a good video tutorial for anyone interested in automotive radars, to know the fundamentals of range and Doppler shifts etc. The information (not)presented is too general for an RF engineer like me, and I could easily make out the RF circuitry and stuff! The microwave ruler maybe interesting but was not required in the presentation (so was the cat!); and I bet, you won't be able to build one commercial automotive radar unless you cover lot more serious stuff in DSP, beamforming, and phased array antenna system; so one should use the information here as preliminary. Nonetheless, a good attempt and the 'flow' is nice!
Strange why you cannot talk about design that originated from ANOTHER company? By another I mean not the one you work for. AFAIK reverse engineering is protected by law in US and many other countries because without reverse engineering every piece of complicated technology turns into a "ransomware".
I have to be cautious. Although the design is not ours, the techniques are still considered confidential. I just don't want to get into a situation where I would have to take the video down.
Christopher Young the microstrip to waveguide conversion structures are published in IEEE journals; paywalled though. The ones specifically in the video is a canned example project in Microwave Office.
Well, he likely has an NDA and it is better to be safe than sorry. You may be right, but that won't get the 800lb lawyer gorilla off his back. I also believe him mentioning that he is on an H1B visa working towards a green card - that's a very precarious situation even for people *not* coming from Iran as he does. He certainly doesn't want to get fired and risk problems with immigration (the visa is tied to your job and unless another company picks up the sponsorship, you are royally screwed).
Not just NDA's for commercial work, but Bell historically has done a lot of work for the US government. SIGINT is more important than ever, so there's a non-zero chance some of either the work and/or the work of people he's managing falls into that realm.
BTW, best episode in ages. The theory/operation episodes are what make you, Alan, Paul, etc so spectacular. I appreciate the amount of work you put into these !
It's regarding Broadband E - Band WR12 to Microstrip Line Transition Using a Ridge Structure on High-Permittivity Thin-Film Material. Stuff you probably won't be able to follow.
Very rarely when a TH-camr says something will be interesting and exciting, it actually is. I'm excited everytime I see a notification for a Signal Path upload.
What I love from your videos is that even though the concepts are advanced, I still end up learning something. In this case, I got an understanding of the doppler effect that I've heard about many times.
In case anyone is wondering:
The circuitry on the second board was some bus related stuff for CAN, a high performance TI Sitara, a DSP IC, DDR3, and the power circuitry.
I'm a dumbass and pretty well destroyed the 2nd board (as you can see from the connector w/ random PCB bits attached) hence it's lack on inclusion in the package.
Thanks for sending the unit in, that was very interesting to have an expert like Shahriar to go through it!
Love the content here. You really have a knack for making complicated subjects understandable. You'd make a great teacher.
Just watched the most recent one and I found this one.
Great work, you help me concentrate on something else than what tries creeping up.
Absolutely fascinating. I'm gutted there were parts you couldn't talk about but its understandable. I love it when people explain RF circuitry it's way above me but you explain it so well even I get to learn stuff. Thanks for the marvellous videos.
Pooch is once again the star of the show
TFW Shahriar's cat is a better RF engineer then I am.
Has anyone told you you're kind of a jerk?
You are sure acting like one.
Great episode Shahriar.
great explanation for the microwave structures on the PCB
I have to say this is a bit old system.
The new systems the chips are packaged with technique called eWLB and they are soldered on the PCB like BGA
Furthermore, the new generations the frequency generation is integrated with the TX front-end
Thanks for another great video!
Could you please create a separate episode explaining how those magical microwave PCB components work? I know that you mentioned them at different occasions, but it would be helpful to have one video with all the knowledge in concentrated form.
Good to see ECU which i am working on is featured in this video. This ECU is developed by delphi indeed.. It is called electronic scanned radar
Thanks for a very good explanation of FMCW radar - I remembered it was something like this but you really made it "click" in place!
2 in one day!?!? what a treat:)
superb explanation
Shahriar, if you teach about microwave on your channel I really think it's going to be one of the best explanations.
Awesome video, thank you for all your work here. One thing I missed was at 11:29 - "Infineon {something} CMOS... so they're just {acronym} process". Thanks again!
Perfect, thank you!
It is a hole different world than what i am used to, amazing. I would like to watch more tutorials into RF stuffs. Nice video.
Shahriar, at around 7:49 when you’re explaining the ramp up and ramp down, isn’t the graph on the left representative of what would be received if the object is moving away from the vehicle/transmitter? The picture shows the object moving towards. Based on the doppler effect a pulse should be frequency shifted up if the object was moving towards the vehicle, but the peak of that graph shows the received pulse as a lower frequency than the transmitted pulse. Is this correct or am I missing something?
Dear Shahriar, At the unused end of the coupler and the edge of unused lines there should be a 50 ohm termination. I don't see any termination there, please explain where is the termination or how it is going to be 50 ohm?
Awesome video as always thank you!
switch the SA for a scope when looking at the output of the mixer. For those interested: get a cheap HB100 10 Ghz unit through ebay and experiment with doppler radar yourself. Oh and btw I very much appreciate your video's.
Nice tear down! That antenna is a Taconic laminate. The Taconic rep brought it by on a visit. What's odd though is that the transmission lines are oriented parallel to the weave; IIRC the sample I saw had them at 45 degrees to keep it more homogenous.
Those WR12 to microstrip transitions are pretty narrow band, but the don't require a back-short. I'm using a differential microstrip to WR15 on a board I just finished. Need a back shirt but much wider band, and can couple to the TX chip without a balun.
Those rat race baluns aren't really baluns if you think about it, since there is no way to fully convert a pure differential signal to single ended with that structure; still requires a good amount of ground currents for the differential side.
Notice lack of termination resistor on the branch line. They must be tapering it down to such high Z0 that it just leaks off.
Anyway, good stuff.
Thanks for your contribution! Why do you think that you can't convert to single-ended? Those interconnected appear to be microstrips.
The Signal Path if it's a purely differential signal on the balanced side (i.e an infinite Zcomm) there is no way for that signal to induce a current in the ground plane for the single ended side. If you sliced the ground plane beneath the differential lines it would fail; at least that's what my EM simulations showed.
The output stage of that TX chip can probably be run either single ended into 50 ohms or differential into 100 ohms, so it has a fairly low common mode impedance, which allows it to work with that rat race balun.
A Marchand balun should work since it relies on resonance, and a tapered line balun, but then you'd need vias to the bottom of the board.
You whizzed by me on the floor at IMS (either Phoenix or Baltimore), and I was thinking I have seen that guy on TH-cam. Didn't get a chance to say Hi. Maybe at another one. Anyway, great TH-cam channel.
Yes, I agree that the output stage is likely quasi-differential and can provide signal into 50-Ohm single-ended. I am sorry that I missed you at IMS! I hope we can meet next time. :)
The Signal Path for sure. Maybe in Philly next year. My company is too cheap for Hawaii.
It is totally possible to convert a fully differential mode to a single ended mode by tapering the differential microstrip line to two single ended microstrip lines first and then combining then with ratrace couplers as done here! I work with similar stuff at a radar research institute in Germany and we do this on a regular basis in order to reduce parasitic bondwire inductance when compared to single ended bond wires.
Funny, I ran into Shahriar at IMS back i Florida a few years ago when I was stil, a master student. He was really nice to talk to. I dont know if you remember me, but it is nice to see that you are doing well!
We are also working with Infineon silicon processes.
Great Video
Does anybody have a link to more information about the "terminated points" at 17:26 ?
Thank you for the great video. I'm very interested in the self-driving car technology.
The ADC is a 6 analog front end channels multiplexed into a single ADC, so it is not a simutaneous sampling system. How to handle this in the system point of view? Thanks a lot!
I've seen designs for acquisition boards for the Large Hadron Collider which used COB technology for the custom analog chips.
Interesting test equipment with the extra sweeping tail.
Thanks for the very interesting video, it's a pity that you couldn't talk about the signal coupling to the antenna array. :-)
That sweeping tail is an idea that Shariar should patent... Perhaps name it 'The Cat Scan'???
It's the start of the video I see a cat sitting on top of an oscilloscope that costs in the five to six digit range, and my eye just starts to twitch.
That's his favorite spot!
Love the video. I couldn’t find the link for the custom rulers.
Hi, I'm willing to buy two of these radars to do TxRx experiments (no FMCW nor Doppler involved), but I'm afraid it will be tricky to obtain the measured reflected/transmitted signal since there is lots of tightly-packaged complex electronics involved. Do you think this is possible? what are the drawbacks and difficulties you can expect to find in this project?
Thanks for the great video!
just great
Just a quick thing to point out. At 9:10, you point out that the top cover has to be plastic to be transparent, but as I understand it, you want to have the thickness such that it works as a 1/4 wave transformer. The plastic is better for reducing attenuation through the medium though, which should in theory mean a wider bandwidth, right?
Is there a way to see the negative frequencies on a spectrum analyzer in order to extract the relative motion direction?
Nope. In most doppler radars, they will use an additional mixer between the CW source and RX mixer with an IF clock to raise the RX IF baseband. That way, you'll be able to see both up and downshifts in freq.
Very informative thank you
Maybe you can negotiate with your superiors on what exactly you are allowed to discuss and go as far as any way possible?
Can't negotiate. That's conflict of interests. I sure hope he could though !
Hi Sharihar... Could you please tell the manufacturer of the radar module and also the vehicle it is used.
hey i have just started to enter into this radar thing , found this video
and to my surprise this is awesome , even though i am not able to understand any of the technical components that you were mentioning (dont even know what is a coupler ) i am enjoying it .
can u mention how do i start of from the basics to fully understand these radar things so that i can be able to build my own FMCW radar
These antennas are called comb-line antennas.
Hello, can you please give me any kind of details on where I can buy a 77 GHz radar module either online or in a retail store.
👍👍
14:05 RF Black Magic Rulers. Sweeeet!
I just bought one.
very interesting
How do you have all of these equipment at home? This could be more than a million dollars... Could you explain how you got them? Sponsorship? Remote work for a company? Do you own a company?
Any reason the vias are not tented?
Hi Shahriar, I love this video. I have been hoping for a while that you would do something on FMCW. Do you have any plans for more videos on this topic? As a golfer and an engineer, I am very intrigued by radar devices (such as the "Trackman") that can track a golf ball's flight through the air and measure things like the rotational velocity of the ball, distance of travel and many other things. It would be cool to be able to do some experiments with stuff like this, unfortunately a lot of the components seem out of reach cost wise. Do you have any tips, or resources you recommend, on how to experiment with radar in the 10.5 GHz range without breaking the bank?
Doppler radar modules at that frequency are available from some online stores that cater to the DIY electronics market, and they're quite inexpensive.
17:35 Does any one knows more about this topic under NDA?
Please refer me to some books/articles or at least keywords needed to lookup in order to understand all these goldy structures around the ICs. I am digital ASIC design engineer with FPGA background and I enjoyed undergrad Electromagnetics, I believe I'll find a way to understand this beauty. Thanks!
microstrip to WR12 transition
Love your cat!
As usual the video was Amazing sir.
Sir I would like to know how energy is being coupled to the antenna(phased array). Thank You 😊
super masss boss
Kinda bummed that he didn't do a deep dive in this module and the microwave magic that makes it tick.
Next time don't use an SA to show the result, just hook up an Audio amplifier. It just beats anything else to experience what happens when you wave the pcb above it.
Yes, I agree. An audio amp hooked to a simple 10 GHz gunnplexer will teach you all you ever wanted to know about Doppler effect. Take it out and point it at some moving cars, or a rotating fan blade, or interestingly, a lit fluorescent tube. The tones you get back are fascinating.
Nice video. However, you began by saying that there is a “frequency shift” in the received ramp from a stationary object. Yet at the end, it’s clear there is not shift in frequency when the copper piece is stationary. There is however a “delay”. Good job, keep up the nice work.
There is a frequency shift in a FMCW radar system from a stationary object because the input signal is FMCW. The delay translates to frequency shift with respect to the transmitted signal. In my test the input signal is CW and therefore no observed shift in frequency, only Doppler shift.
I think wire bonding is also used because of harsh environment in cars. Solder joints could brake after some time if not protected very well.
This is a good video tutorial for anyone interested in automotive radars, to know the fundamentals of range and Doppler shifts etc. The information (not)presented is too general for an RF engineer like me, and I could easily make out the RF circuitry and stuff! The microwave ruler maybe interesting but was not required in the presentation (so was the cat!); and I bet, you won't be able to build one commercial automotive radar unless you cover lot more serious stuff in DSP, beamforming, and phased array antenna system; so one should use the information here as preliminary. Nonetheless, a good attempt and the 'flow' is nice!
Strange why you cannot talk about design that originated from ANOTHER company? By another I mean not the one you work for. AFAIK reverse engineering is protected by law in US and many other countries because without reverse engineering every piece of complicated technology turns into a "ransomware".
I have to be cautious. Although the design is not ours, the techniques are still considered confidential. I just don't want to get into a situation where I would have to take the video down.
Christopher Young the microstrip to waveguide conversion structures are published in IEEE journals; paywalled though.
The ones specifically in the video is a canned example project in Microwave Office.
Well, he likely has an NDA and it is better to be safe than sorry. You may be right, but that won't get the 800lb lawyer gorilla off his back.
I also believe him mentioning that he is on an H1B visa working towards a green card - that's a very precarious situation even for people *not* coming from Iran as he does. He certainly doesn't want to get fired and risk problems with immigration (the visa is tied to your job and unless another company picks up the sponsorship, you are royally screwed).
Not just NDA's for commercial work, but Bell historically has done a lot of work for the US government. SIGINT is more important than ever, so there's a non-zero chance some of either the work and/or the work of people he's managing falls into that realm.
BTW, best episode in ages. The theory/operation episodes are what make you, Alan, Paul, etc so spectacular. I appreciate the amount of work you put into these !
FoMoCo=Ford Motor Company? Willing to bet that came out of a Cadillac if my initial guess is correct.
Nope, came out of a new generation F-150, Cadillac is GM not FoMoCo.
SnowIsMyTerrain Bleh. Not a car person, but at least I guessed correctly it came out of a Ford vehicle. lol
eh. Doppler actually lived to be an old man. Left expectancy around that time was 37 years.
That's because so many babies died, people would either live to 1 or 61
Watching in 2024 now entire board in single chip..
Aliens!
Try a tuning fork. You will be surprised.
the cat is angry 😾
+fat thar He is the happiest cat!
you have lab-cats :/
No cats, just one cat! :) Dr. Pooch.
even he has a PhD ;)
I love Electronics, but this is magic not Electronics :-)
this guy sounds persian
salum
now i come back
i do radar for missile
goes boom
What is the point of the video if you don't want people to learn what you know?
Why can't talko about it? Physics is not patentable.
Physics is not patentable. Implementation is however.
Knowing how the competitor is thinking is very important. "Oh, they think we did this to achieve ...? Never thought about that let's look into it".
It's regarding Broadband E - Band WR12 to Microstrip Line Transition Using a Ridge Structure on High-Permittivity Thin-Film Material.
Stuff you probably won't be able to follow.
@@PushyPawn now I am finishing my studies and can understand :P but yes, one year ago I could not