In sonar we have DELTIC delay line time compression. Auto correlation you can track a moving submarine with just two hydrophones. Used by CAPTOR smart mines in the late 70s. (Band limited white nosie from moving sub is XORed and accumulated). Way before the microprocessor was around.
Very nice video, worth watching! Small typo fix: the acronym FSK at 3:53 is Frequency Shift Keying, not keyring. Pull request sent via Github to fix it in the video.
Chirps are used in pulse radars as well - because if you cross-correlate a chirp with itself, you get a pulse out. So a pulsed radar doesn't actually send a pulse, it sends a chirp over a longer duration. The received signal is then correlated with the original chirp, which returns a pulse. You can time those pulses much more accurately this way. (the chirp is not just a random chirp - it's actually a pulse convolved with a chirp, so you get the time from the convolution as well).
Yes! This video is DEFINITELY over simplified because I wanted to give a very basic intro to FMCW. But yes, probably most pulsed radars don't just transmit a constant frequency. That's a topic for another video :) Thanks for watching!
@@MarshallBrunerRF Duty cycle on these fibre lasers is well below 0.1%. They are "pumped" by continuous laser source at longer wavelength and then triggered by a low power pulse at output wavelength. This causes an avalanche effect and whole accumulated energy exits out within nanoseconds.
Nice! I would say that at 5:40 "beat" is better described as the difference in frequency between two signals than as the difference in frequency of one signal at different times.
As an explanation - sure. I did have to pause for a moment and check the calculations, because of that. But I'm unsure of how actual radars of that type do things. It might be more accurate to say they do it this (roundabout) way to eliminate some kind of possible issue that could come from just taking the difference. Though it shouldn't be that important if the chirp is of that "shape" and you accept limitations of range and measuring velocity of the target as well. I guess it could be better technique overall for different shapes and changing chirp length and downtime (I do think that Continuous Wave Radars will still have downtime between chirps - but again, for at least those types of radars, I con't know them super well and I forgot what I used to know a long time ago). Personally I knew a bit about long range Air Defense radars (both pulse and CW), but that was a long time ago. I will now check next video. While it is good to have "basics" video initially, this channel just started, so I assume person behind it will work on how deep vs. how accessible vs. how short their videos are. I prefer deep dives that sacrifice length (so they are quite long) in order to be reasonably understandable but also very in depth at the same time. But I know that YT might initially dislike longer videos.
Very good video to start with this topic. I work with radars in cars... in this setup it becomes so much more complicated. You have to measure velocity and distance. You have many objects in different locations. You need also a direction to know the real position (this isn't done with moving antennas). This is such an intersting topic. I still learn new stuff. Radars seem to be black magic. All this combinaiton of frequency shifts while constantly changing how fast the frequiency shifts. Than calculation of the direction by utilizing antenna arrays. Methods to avoid to get blinded by other radar senors (such as using pliarized signals).
Yeah it's such an insanely cool field! Hopefully you'll enjoy my next video which is going into the range-Doppler spectrum and multiple targets in range.
About peak pulsed power versus average power: all radar systems have to cope not with an inverse square law for detection of echo returns but an inverse FOURTH power law. This is known as the radar equation. RF power, regardless of modulation used, has to travel both to and from the target and thus incurs an inverse square reduction each way. Note that this is a multiplicative affect. This requires exceptional receiver sensitivity and stability, far beyond what normal radio reception requires. This is why effective radar was not invented thirty years before by someone like Marconi. Sure, various aspects of radio obstacle detection and even simple ranging were developed in those three decades, but designing receivers sensitive enough to receive inverse fourth power returns AND yet not be de-sensed by the huge transmit pulses requires very clever engineering indeed.
I have a radar the size of a pack of cigarettes on my motorcycle for the adaptive cruise control. I was trying to figure out how it works, and it was a huge help in understanding that acronym.
This was good, but some basic principles were not covered: need for 2 antennas Chirp t not only effects max doppler; but it also effects Doppler Res, Range Res and max unambiguous range. Doppler/range coupling: (sawtooth waveforms cannot measure Doppler. Doppler shifts will result in range errors with this waveform. To correct for range distortion from Doppler shifts a triangular waveform is needed. The depiction of RF traveling in 3D space as a sine wave is not correct. Most media outlets do this. But I think it misleads ppl who are trying to understand RF/EM/Light (all one one same). The sine wave is a graph of voltage over time. Not a depiction of the propagation of RF. The depiction should show plane waves traveling out. (If vertically polarized; vertical lines traveling perpendicular to their plane; with each plane having its field vector (arrow) indicating if it’s a down wave or an up wave. Each “wavelength” is physically 2 plane waves in 3D space. One up wave followed by a down wave. Intuition is critical for understanding RF theory. Otherwise ppl will always think of RF engineering as black magic.
All this is correct but, in my opinion, unnecessary for this intro video. For one, I am producing an ongoing series on this subject - the next one being released soon will talk about range Doppler coupling, Doppler res, etc. Including all the other EM concepts you mentioned would make for a long and tedious video. Thanks for watching and commenting! I would like to cover those concepts separately in the future.
It feels like you should ideally adjust the frequency as a function of time based on the recieved information to best isolate multiple targets moving at multiple speeds
Next video will go into how velocity information of multiple targets is obtained. I haven't personally seen what you're talking about, I would have to think about it some more. Thanks for watching!
We need DYI anti-drone radar ASAP. So many people are breaching the privacy barier... We should have possibility to detect and shut down those objects.
The change in frequency should follow a sine wave to reduce the amount of logic and increase the response of the system. Using any linearly variable pattern can be complicated to extract both position and velocity data from since those lines can still be shifted in such a way as to partially or fully overlap. Also, there is a distance limit based on frequency and receiver sensitivity since change in position of whichever object you are measuring must fall within one wavelength of the change in frequency.
hi may I ask why pulsed avg pwr and desired pulsed avg pwr matter? wouldn't the inverse square law apply the same to the discrete intervals as well as the continuous ones? i think the statement isn't very clearly phrased but hoping I can get an answer or a question of some sort
I probably could have explained that part better... One of the complications is how difficult it is to produce such high powers for short time intervals. The average power is important for power budgeting.
There is a lot of fun stuff out there that even as a hamradio op i'm not allowed to mess with, radars are one of those. I found out the hard way that many regions don't permit citizens to operate radar transmitters, at the time i was searching for information about it, i just wanted to build my own weather radar for fun.
Why should we care about the average power transmitted? The pulse or the continuous wave will come back at a power following the inverse square law. If your receiver needs a power of one μW (example not realistic) it needs this power from the returning pulse and the same with continuous radars. If the distance cuts the receive power to one millionth of the send power you need peak send power of at least 1 W. Regardless of if it is 1 W at 10% of time or 1 W the whole time. The benefit of continuous is you have in most cases a finer temporal resolution. And this is at small distances more important because you need to make sure that nothing can slip through on an conveyer belt. And sharp pulses with low delay between each other has the problem with its frequency spectrum. So it is really difficult to do it for this application.
From a function point of view, what (on earth) is the significance of average power in a pulsed system. From a power consumption, and power stage heating side, pulses are only beneficial. The very origin of pulsed radar was that very fact: high signal strength from low average power without toasting the power stage. Without it, radar wouldn't have worked as early as it did!
Good point, and maybe I wasn't clear enough on this. One of the difficult things with pulsed radar is generating that high amount of power in such a short time. Especially generating it in resource constrained environments like a car is difficult. Thanks for watching!
@MarshallBrunerRF Boost converters and capacitors are "dime a dozen" components. The question is how best to utilise what average power is available. DO you pulse it, or somehow fix the problems with (relatively) low power. The problem being sensitivity and noise rejection in the receiver. For reference, I had a brief, in depth look at aviation radar altimiters. They use swept frequency mainly due to the simplicity of decoding the received signal in a pre digital era.
@@RUDRARAKESHKUMARGOHIL yeah! This isn’t always the case (for example automotive radar sometimes) but you want to increase average output power because the signal you transmit degrades as it travels through the air. So if you transmit a higher power, you can see farther distances in general. Hope this helps and thanks for watching!
Yeah! This was a super basic intro to the technology so I didn't cover that, but yes there definitely is a gap many times - I included this in the caveats section of the description :) Thanks for watching!
@@imarshad thanks for the feedback, I should be posting another video tomorrow going into more depth on the implementation and I tried to slow down the complicated parts. Hopefully you get to watch and let me know if it’s more clear😊 thanks for watching!
How you cancel the doppler-Frequenz-shift? There two possible cause of a discret frequency, the Distanz or change by doppler-effect. I‘m ask, cause here in Germany on the autobahn is the doppler-shift relevant 😂
I'm actually working on a video right now about estimating range and Doppler in FMCW radars and I'll release a python notebook where you can play around with real world values for automotive radar. This is the video I'm most excited for and it should be pretty cool. Thanks for watching! Hope you stick around for the Doppler one :)
If we do diversify learning so far we may find that CWFM isn’t new. Such practice exists before our civilization by ocean wales. Try listen to their call navigating in deep ocean.
In sonar we have DELTIC delay line time compression. Auto correlation you can track a moving submarine with just two hydrophones. Used by CAPTOR smart mines in the late 70s. (Band limited white nosie from moving sub is XORed and accumulated). Way before the microprocessor was around.
@@jamesmorton7881 fascinating! I would love to learn more about sonar and its applications, but I know very little now
Very nice video, worth watching! Small typo fix: the acronym FSK at 3:53 is Frequency Shift Keying, not keyring. Pull request sent via Github to fix it in the video.
Hey! Thanks for the comment and PR! I actually have that in the errata in the description as well.
Chirps are used in pulse radars as well - because if you cross-correlate a chirp with itself, you get a pulse out. So a pulsed radar doesn't actually send a pulse, it sends a chirp over a longer duration. The received signal is then correlated with the original chirp, which returns a pulse. You can time those pulses much more accurately this way. (the chirp is not just a random chirp - it's actually a pulse convolved with a chirp, so you get the time from the convolution as well).
Yes! This video is DEFINITELY over simplified because I wanted to give a very basic intro to FMCW. But yes, probably most pulsed radars don't just transmit a constant frequency. That's a topic for another video :) Thanks for watching!
Great video on why chirps are perfect for radar. Keep it up, I'll watch anything you make.
Thank you so much!! New video tomorrow about the hardware and software implementation of FMCW radar :)
Very cool video! 1:48 That's also why pulsed lasers are able to to ablate material while being "only" 2W of average power
yeah I mean 2W at 10% duty cycle or less is quite a bit!
@@MarshallBrunerRF Duty cycle on these fibre lasers is well below 0.1%. They are "pumped" by continuous laser source at longer wavelength and then triggered by a low power pulse at output wavelength. This causes an avalanche effect and whole accumulated energy exits out within nanoseconds.
Nice! I would say that at 5:40 "beat" is better described as the difference in frequency between two signals than as the difference in frequency of one signal at different times.
Thanks for the suggestion! Definitely happy to take any and all constructive criticism as I want these to be really helpful. Thanks for watching!
As an explanation - sure. I did have to pause for a moment and check the calculations, because of that. But I'm unsure of how actual radars of that type do things. It might be more accurate to say they do it this (roundabout) way to eliminate some kind of possible issue that could come from just taking the difference. Though it shouldn't be that important if the chirp is of that "shape" and you accept limitations of range and measuring velocity of the target as well. I guess it could be better technique overall for different shapes and changing chirp length and downtime (I do think that Continuous Wave Radars will still have downtime between chirps - but again, for at least those types of radars, I con't know them super well and I forgot what I used to know a long time ago).
Personally I knew a bit about long range Air Defense radars (both pulse and CW), but that was a long time ago. I will now check next video. While it is good to have "basics" video initially, this channel just started, so I assume person behind it will work on how deep vs. how accessible vs. how short their videos are. I prefer deep dives that sacrifice length (so they are quite long) in order to be reasonably understandable but also very in depth at the same time. But I know that YT might initially dislike longer videos.
incredible video! the animations really did help at explaining the concept very clearly. This should have, at least, 100k views!|
@@miqueasgsw6818 thank you so much! Glad it helped 😊
Amazing video, thanks for putting it together! Please do more if u can!
Definitely! I'm working on the follow-up to this now! Thanks so much for watching!
Very good video to start with this topic.
I work with radars in cars... in this setup it becomes so much more complicated. You have to measure velocity and distance. You have many objects in different locations. You need also a direction to know the real position (this isn't done with moving antennas).
This is such an intersting topic. I still learn new stuff. Radars seem to be black magic. All this combinaiton of frequency shifts while constantly changing how fast the frequiency shifts. Than calculation of the direction by utilizing antenna arrays. Methods to avoid to get blinded by other radar senors (such as using pliarized signals).
Yeah it's such an insanely cool field! Hopefully you'll enjoy my next video which is going into the range-Doppler spectrum and multiple targets in range.
About peak pulsed power versus average power: all radar systems have to cope not with an inverse square law for detection of echo returns but an inverse FOURTH power law. This is known as the radar equation. RF power, regardless of modulation used, has to travel both to and from the target and thus incurs an inverse square reduction each way. Note that this is a multiplicative affect. This requires exceptional receiver sensitivity and stability, far beyond what normal radio reception requires. This is why effective radar was not invented thirty years before by someone like Marconi. Sure, various aspects of radio obstacle detection and even simple ranging were developed in those three decades, but designing receivers sensitive enough to receive inverse fourth power returns AND yet not be de-sensed by the huge transmit pulses requires very clever engineering indeed.
this video is really helpful for me understanding FMCW
@@billourou4443 that means a ton!!!
Grant Sanderson(3Blue1Brown) version of rf .Great Work👍👍
Hahaha I can dream
Very good! I didn't know the technology. Many thanks!
Glad you enjoyed it!
Good explanation! Thx for sharing.
Glad you found it useful! Thanks so much for watching
Very well done explainer
Thanks so much!
That's great. I love radars, Your explanation was excellent and that's why I subscribed to your channel. Keep it up I'm looking forward.
@@majids8198 thanks for watching! Glad you liked it
I have a radar the size of a pack of cigarettes on my motorcycle for the adaptive cruise control. I was trying to figure out how it works, and it was a huge help in understanding that acronym.
Excellent! Crazy how small they are
Wow this is great! Please continue to make more! Subscribed.
@@reslofbeats thanks so much!
This was good, but some basic principles were not covered:
need for 2 antennas
Chirp t not only effects max doppler; but it also effects Doppler Res, Range Res and max unambiguous range.
Doppler/range coupling: (sawtooth waveforms cannot measure Doppler. Doppler shifts will result in range errors with this waveform. To correct for range distortion from Doppler shifts a triangular waveform is needed.
The depiction of RF traveling in 3D space as a sine wave is not correct. Most media outlets do this. But I think it misleads ppl who are trying to understand RF/EM/Light (all one one same).
The sine wave is a graph of voltage over time. Not a depiction of the propagation of RF.
The depiction should show plane waves traveling out. (If vertically polarized; vertical lines traveling perpendicular to their plane; with each plane having its field vector (arrow) indicating if it’s a down wave or an up wave.
Each “wavelength” is physically 2 plane waves in 3D space. One up wave followed by a down wave.
Intuition is critical for understanding RF theory. Otherwise ppl will always think of RF engineering as black magic.
All this is correct but, in my opinion, unnecessary for this intro video. For one, I am producing an ongoing series on this subject - the next one being released soon will talk about range Doppler coupling, Doppler res, etc. Including all the other EM concepts you mentioned would make for a long and tedious video. Thanks for watching and commenting! I would like to cover those concepts separately in the future.
one of the better #some4 videos ive seen for sure!
@@bean_TM thanks so much for your comment and watching!
An amazing video. I really appreciate it.
Thanks so much for watching!
It feels like you should ideally adjust the frequency as a function of time based on the recieved information to best isolate multiple targets moving at multiple speeds
Next video will go into how velocity information of multiple targets is obtained. I haven't personally seen what you're talking about, I would have to think about it some more. Thanks for watching!
This is awsome. Nice work dude
@@TannerMageeYT heyyyy! Thanks so much! Excited for when you start making manim animation 😉
We need DYI anti-drone radar ASAP. So many people are breaching the privacy barier... We should have possibility to detect and shut down those objects.
Check out Jon Kraft's series on drone tracking radar
So interesting! Thank you!
So glad you liked it! More coming soon :)
3:52 FSK is Frequency Shift Keying, not Keyring.
I included an errata in that video with this in it. Thanks for pointing it out!
Great video !
@@LucaSpezzani-gz7no thanks!!
The change in frequency should follow a sine wave to reduce the amount of logic and increase the response of the system. Using any linearly variable pattern can be complicated to extract both position and velocity data from since those lines can still be shifted in such a way as to partially or fully overlap. Also, there is a distance limit based on frequency and receiver sensitivity since change in position of whichever object you are measuring must fall within one wavelength of the change in frequency.
hi may I ask why pulsed avg pwr and desired pulsed avg pwr matter? wouldn't the inverse square law apply the same to the discrete intervals as well as the continuous ones? i think the statement isn't very clearly phrased but hoping I can get an answer or a question of some sort
I probably could have explained that part better... One of the complications is how difficult it is to produce such high powers for short time intervals. The average power is important for power budgeting.
nice work
There is a lot of fun stuff out there that even as a hamradio op i'm not allowed to mess with, radars are one of those. I found out the hard way that many regions don't permit citizens to operate radar transmitters, at the time i was searching for information about it, i just wanted to build my own weather radar for fun.
Building your own weather radar would be so awesome! I'm definitely going to be making videos covering the topic. Thanks for watching!
Super interesting. I'd love to see a real hardware implementation one could replicate and play with.
Check out analog devices' Phaser radar. Jon Kraft has some great videos on it using that hardware.
Why should we care about the average power transmitted? The pulse or the continuous wave will come back at a power following the inverse square law. If your receiver needs a power of one μW (example not realistic) it needs this power from the returning pulse and the same with continuous radars. If the distance cuts the receive power to one millionth of the send power you need peak send power of at least 1 W. Regardless of if it is 1 W at 10% of time or 1 W the whole time. The benefit of continuous is you have in most cases a finer temporal resolution. And this is at small distances more important because you need to make sure that nothing can slip through on an conveyer belt. And sharp pulses with low delay between each other has the problem with its frequency spectrum. So it is really difficult to do it for this application.
lol FREQUENCY SHIFT KEYING not keyring 😂
Crystal-clear explanation ! Thank you !
From a function point of view, what (on earth) is the significance of average power in a pulsed system.
From a power consumption, and power stage heating side, pulses are only beneficial.
The very origin of pulsed radar was that very fact: high signal strength from low average power without toasting the power stage. Without it, radar wouldn't have worked as early as it did!
Good point, and maybe I wasn't clear enough on this. One of the difficult things with pulsed radar is generating that high amount of power in such a short time. Especially generating it in resource constrained environments like a car is difficult. Thanks for watching!
@MarshallBrunerRF Boost converters and capacitors are "dime a dozen" components.
The question is how best to utilise what average power is available.
DO you pulse it, or somehow fix the problems with (relatively) low power.
The problem being sensitivity and noise rejection in the receiver.
For reference, I had a brief, in depth look at aviation radar altimiters. They use swept frequency mainly due to the simplicity of decoding the received signal in a pre digital era.
beautiful ! . you earned a sub.
Glad to have you! Thanks for watching!
Frequency shift keying, not key ring.
Added to the errata :)
can you explain at 1:53 why we want to have a greater desired power ?
@@RUDRARAKESHKUMARGOHIL yeah! This isn’t always the case (for example automotive radar sometimes) but you want to increase average output power because the signal you transmit degrades as it travels through the air. So if you transmit a higher power, you can see farther distances in general. Hope this helps and thanks for watching!
Ya it helped ! I left the video till the doubt get solved 😅 but will see it completely now...
Isn't chirp ramp length determining maximum distance? Around 4:36
Awesome, but what about TX and RX polarization, which one is the most used, or the best suited so to speak.
Dual-pol radar is definitely a topic I will cover - especially in the context of weather radar
@@MarshallBrunerRF Nice.
> FMCW radars often also have some "off" time between chirps
That would not be FMCW, but FMOP.
Yeah! This was a super basic intro to the technology so I didn't cover that, but yes there definitely is a gap many times - I included this in the caveats section of the description :) Thanks for watching!
👍👍👍
I think the “CW” in “FMCW” Is kinda funny. By this reasoning, any modulation is CW! 😂
Yeah but it's CW compared to what most radars use which is a pulsed signal. Thanks!
It was super easy toll 5:30. I lost there... Maybe i will listen to it 3-4 times to get this portion.
@@imarshad thanks for the feedback, I should be posting another video tomorrow going into more depth on the implementation and I tried to slow down the complicated parts. Hopefully you get to watch and let me know if it’s more clear😊 thanks for watching!
Awesome
I thought it said FWMC at first and wondered what Fuwamoco had to do with wave modulation lol
lol no not the same!
How you cancel the doppler-Frequenz-shift? There two possible cause of a discret frequency, the Distanz or change by doppler-effect.
I‘m ask, cause here in Germany on the autobahn is the doppler-shift relevant 😂
I'm actually working on a video right now about estimating range and Doppler in FMCW radars and I'll release a python notebook where you can play around with real world values for automotive radar. This is the video I'm most excited for and it should be pretty cool. Thanks for watching! Hope you stick around for the Doppler one :)
This is the Golden Earring of You Tube videos.😂
TIP: Increase and normalize your volume so you're not so quiet
Thanks for the tip. Forgot to do it on this video but the most recent video should have it
the bau bau radar
Not enough people use the word initialism. :)
Haha, I knew people would call it out if I incorrectly said acronym
Look! An RGB radar!
I don't get why do we care about average power. Wouldn't consuming less power be good?
Hardware please
If we do diversify learning so far we may find that CWFM isn’t new. Such practice exists before our civilization by ocean wales. Try listen to their call navigating in deep ocean.
Fascinating... will have to look into this
Great Explanation Sir. But try to explain in little bit slow....
Thanks for watching and providing feedback! Just to be clear, you thought it should be slower? What part did you think was too fast?
@@MarshallBrunerRFsir, I mean you are explaining very fast sonit will be better to us if you explains slowly
@@ankursvideos1210 thanks for the feedback!
When I have issues with speed I go into the settings and reduce the speed to 75% of normal. This works for me with out being too slow.
@Subgunman thanks
amazing video, can you design FMCW radar?
@@TuralMontin-w9k I’m actually currently working on a video going into many of the blocks used when designing an FMCW system!
@@MarshallBrunerRF thanks , I'm looking forward to it. I subscribed to you.