Introduction to Using GNU Radio
Video in TIB AVPortal:
Introduction to Using GNU Radio
Formal Metadata
Title 
Introduction to Using GNU Radio

Alternative Title 
Software Defined Radio  GNU radio

Title of Series  
Author 

License 
CC Attribution 2.0 Belgium:
You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor. 
Identifiers 

Publisher 

Release Date 
2016

Language 
English

Production Year 
2015

Content Metadata
Subject Area 
00:00
Multiplication
Wave
Multiplication sign
Electromagnetic radiation
Musical ensemble
Spektrum <Mathematik>
00:52
Ocean current
Dataflow
Group action
Sine
Digital electronics
Code
Multiplication sign
Perspective (visual)
Theory
Computer programming
Impulse response
Frequency
Term (mathematics)
Videoconferencing
Physical system
Noise (electronics)
Dependent and independent variables
Graph (mathematics)
Block (periodic table)
Software developer
Interface (computing)
Graph (mathematics)
Interactive television
Cartesian coordinate system
Category of being
Graphical user interface
Process (computing)
Integrated development environment
Electromagnetic radiation
Website
Library (computing)
03:41
Complex (psychology)
Code
State of matter
Multiplication sign
Source code
Sheaf (mathematics)
1 (number)
Set (mathematics)
Parameter (computer programming)
Mathematics
Graphical user interface
Different (Kate Ryan album)
Analogy
Cuboid
Endliche Modelltheorie
Stability theory
Simulation
Touchscreen
Electric generator
Block (periodic table)
Digitizing
GUI widget
Electronic mailing list
Sampling (statistics)
Variable (mathematics)
Regulärer Ausdruck <Textverarbeitung>
Connected space
Category of being
Process (computing)
Frequency
Vector space
Gotcha <Informatik>
Normal (geometry)
Directed graph
Point (geometry)
Ocean current
Dataflow
Game controller
Digital signal processor
Observational study
Tournament (medieval)
Real number
Motion capture
Letterpress printing
Number
Element (mathematics)
Complex number
Computer hardware
Module (mathematics)
Graph (mathematics)
Cellular automaton
Interactive television
Expert system
Volume (thermodynamics)
System call
Personal digital assistant
Logic
Factory (trading post)
09:30
Domain name
Multiplication sign
Primitive (album)
Fourier transform
Mereology
19 (number)
Field (computer science)
Band matrix
Category of being
Frequency
Sign (mathematics)
Electromagnetic radiation
Representation (politics)
Physical system
10:41
Group action
State of matter
Multiplication sign
View (database)
Shape (magazine)
Mereology
Food energy
Impulse response
Mathematics
Singleprecision floatingpoint format
Core dump
Physical system
Curve
Touchscreen
Electric generator
Smoothing
Fourier series
Sound effect
Maxima and minima
19 (number)
Frequency response
Connected space
Band matrix
Category of being
Wave
Process (computing)
Summierbarkeit
Right angle
Mathematician
Directed graph
Point (geometry)
Ocean current
Slide rule
Observational study
Connectivity (graph theory)
Graph coloring
Theory
Field (computer science)
Power (physics)
Number
Frequency
Term (mathematics)
Computer hardware
Electric field
Representation (politics)
Nichtlineares Gleichungssystem
Data structure
Finite impulse response
Form (programming)
Domain name
Dependent and independent variables
Matching (graph theory)
Mathematical analysis
Voltmeter
Fourier transform
Limit (category theory)
Signal processing
Musical ensemble
Maxwell's equations
Inductive reasoning
16:29
Beat (acoustics)
Observational study
Real number
Multiplication sign
Motion capture
Canonical ensemble
Field (computer science)
Impulse response
Fluid statics
Mathematics
Response time (technology)
Band matrix
Nichtlineares Gleichungssystem
Chisquared distribution
Theory of relativity
Gradient
Voltmeter
Food energy
Computer
Fast Fourier transform
Band matrix
Personal digital assistant
Network topology
Video game
MiniDisc
Pole (complex analysis)
19:00
Impulse response
Frequency
Dataflow
Electric generator
Graph (mathematics)
Code
Musical ensemble
Sound effect
Right angle
Disk readandwrite head
2 (number)
19:57
Unitäre Gruppe
Pulse (signal processing)
Dataflow
Group action
Random number generation
Poynting vector
Multiplication sign
Source code
1 (number)
Field (computer science)
Neuroinformatik
Impulse response
Frequency
Digital signal processing
Bit rate
Synchronization
Computer hardware
Musical ensemble
Series (mathematics)
Form (programming)
Physical system
Noise (electronics)
Addition
Simulation
Dependent and independent variables
Graph (mathematics)
Electric generator
Expression
Sampling (statistics)
Electronic mailing list
Infinity
Bit
Fourier transform
Sound card
Variable (mathematics)
Band matrix
Loop (music)
Befehlsprozessor
Vector space
Summierbarkeit
Cycle (graph theory)
Quicksort
24:38
Point (geometry)
Dependent and independent variables
Electric generator
Arm
Cellular automaton
Multiplication sign
Sampling (statistics)
Food energy
Field (computer science)
Goodness of fit
Sample (statistics)
Universe (mathematics)
Right angle
Flux
Spectrum (functional analysis)
26:17
Filter <Stochastik>
Dataflow
Matter wave
Transformation (genetics)
Multiplication sign
Electric dipole moment
Shape (magazine)
Frequency
Graphical user interface
Causality
Musical ensemble
Data structure
Form (programming)
Physical system
Dependent and independent variables
Matching (graph theory)
Graph (mathematics)
Structural load
Range (statistics)
Food energy
Data transmission
Category of being
Word
Loop (music)
Frequency
Charge carrier
Right angle
28:26
Complex (psychology)
Group action
Multiplication sign
Tap (transformer)
Range (statistics)
Sheaf (mathematics)
Insertion loss
Shape (magazine)
Parameter (computer programming)
Food energy
Impulse response
Mathematics
Graphical user interface
Cuboid
Physical system
Area
Electric generator
Block (periodic table)
Closed set
Streaming media
Range (statistics)
Bit
Food energy
Variable (mathematics)
Virtual machine
Connected space
Frequency response
Band matrix
Message passing
Frequency
Right angle
Filter <Stochastik>
Point (geometry)
Slide rule
Dataflow
Statistics
Perfect group
Graph coloring
Theory
Power (physics)
Number
Frequency
Complex number
Musical ensemble
Tunis
Domain name
Noise (electronics)
Dependent and independent variables
Scaling (geometry)
Graph (mathematics)
Line (geometry)
Cartesian coordinate system
System call
Approximation
Diameter
Word
Loop (music)
Musical ensemble
Library (computing)
34:59
Standard deviation
Group action
User interface
Multiplication sign
Tap (transformer)
Source code
Food energy
Perspective (visual)
Order of magnitude
Variable (mathematics)
Neuroinformatik
Medical imaging
Different (Kate Ryan album)
Pole (complex analysis)
Physical system
Area
Bit
Time domain
Connected space
Frequency response
Message passing
Phase transition
Selforganization
Summierbarkeit
Simulation
Filter <Stochastik>
Point (geometry)
Dataflow
Digital filter
Functional (mathematics)
Computer file
Image resolution
Menu (computing)
Plot (narrative)
Number
Power (physics)
Frequency
Degrees of freedom (physics and chemistry)
Operator (mathematics)
Musical ensemble
Binary multiplier
Domain name
Noise (electronics)
Dependent and independent variables
Software configuration management
Line (geometry)
Euler angles
Signal processing
Symbol table
Graphical user interface
Pell's equation
Personal digital assistant
Normed vector space
Musical ensemble
Window
41:09
Filter <Stochastik>
Standard deviation
Slide rule
Module (mathematics)
Addition
Software developer
Multiplication sign
Execution unit
Set (mathematics)
Bit rate
Mass
Mereology
Data model
Graphical user interface
Synchronization
Binary multiplier
Fundamental theorem of algebra
Capability Maturity Model Integration
Simulation
Algorithm
Arm
Demo (music)
Block (periodic table)
Building
Software developer
Electronic program guide
Projective plane
Electronic mailing list
Range (statistics)
Streaming media
Control flow
Mechanism design
Symbol table
Category of being
Type theory
Coding theory
Frequency
Digital signal processing
Error message
Computer hardware
Block (periodic table)
42:28
Point (geometry)
Android (robot)
Pulse (signal processing)
Sine
Group action
Module (mathematics)
User interface
Line (geometry)
Multiplication sign
Real number
Execution unit
Food energy
Power (physics)
Frequency
CNN
Different (Kate Ryan album)
Singleprecision floatingpoint format
Computer hardware
Data structure
Algebra
Physical system
Code division multiple access
Standard deviation
Electric generator
Regulator gene
Block (periodic table)
Structural load
Electronic program guide
Computer file
Projective plane
Streaming media
Bit
Sturm's theorem
Measurement
Frequency response
Band matrix
Symbol table
Wave
Coding theory
Error message
Website
Energy level
Quicksort
Musical ensemble
Block (periodic table)
Software protection dongle
47:45
Googol
00:05
going over this connection so that even the idea of the with the multiple tones over here so we can we can actually transmit the signals multiple signals for audio us that's where the idea of the but his idea would actually do telephone came from so this is interesting is that she goes what we're back into the band with this you in a 2nd we get to wasn't Maxwellians but this 1 is by late 19th century permits a late 19th century when this is happening and at the same time we're all studying this idea of electromagnetic waves and was a huge controversy here but there is a controversy inside the community about what electromagnetic waves work and so 1 of the reason why this ties in a Bell's work ties into this is and we've all experienced it and I'm sure I could generate it
00:53
somehow here on but 1 of the things that they had people thinking about of the creation of electromagnetic waves are and how were actually able to demonstrate that comes from this concept here this is the telephone switchboard so hopefully a lot of are even the younger generation of of myself included has seen this at least you know and videos er TV shows some issue plugging in into a switchboard to serve to connect a circuit between 2 endpoints but when you do that it's the same thing is if you're speakers honoring microphone is unemployed into a Jack that awful noise that that that that that loudly crinkly noise that comes up on this was termed extra currents right so the harder it made the under the the they were dealing with a tries suppress that but this thing happen extra current happens so trying to figure out why this happened is the is kind of the party interesting the story of the invention or the discovery electromagnetic waves is what's happening is already for us study 1st in the 1st in radio fluoresce what's happening here is you're going from no signal to signal very quickly so you have this not impulse but impulselike response to the to the system so no signal to signal is a very sharp transition time so the 1st name last year as far as from uh from an educational perspective working with radio is that short duration time means long response in frequency so we have these 2 properties of the instrumental and how we we understand things so and by the way all
02:30
my all this flowgraph sins and examples in all this stuff are going to go into onto my website so wouldn't download them and and play with them both fairly simple but this is with the interactions for so with this flow graph this will be the term that we use in in a radio for a basically a radio applications of so called a flow graph because data flows in it a graph data flows through it and I should come from because the process graphs of chemical diversity as a theory of the so yes so that every flow graph and this is a thing called a guinea radio companion it's a graphical interface but uh development interface environment for working with the new radio as the basic variable to graphically put down your blocks that make up your graph your application connect them together and run this now this actually does is it builds a Python program and that place subprime was was actually eating executed but it you'd see code so there's like these 3 layers here in the graph of the face a place on the application itself and in the underlying C + + library that is like in a radio at so if I go to the candidates now switch over
03:43
to the world of whom at so what we do to create a flow graph is you pop open and generated companion and you have this huge list of blocks over here or a list of categories and so very useful of these of a huge number of these blocks and and so each category will have and so we we're gonna wanna do is like look at the In this work as price topic to defer waveform generators thus far below so demodulators you know grabbing so see in our modulation section we have a bunch of different modulators by modulators and
04:27
demodulators for FM and AM that them for like analog stuff we a lot of digital model the modulation what a blocks in here I would say so as you can always just search so you know control left put pops up a search box and say I wanna vector source can kind of you know almost that really reg expert you can search for anything as a vector in it right so we need to create a source we need to somehow put the the into our flow graph and this Indian simulation exercises simulation is it has real hardware attached to it our but we have a sources you generate data into the floor afterward a process is somehow how to study this this it is a fact all I've done is I've put audio since we can hear it before general like will see how how much we can hear it in this tournament volume just in case the and the reaction you in time and this is a really by think important thing for the study of signals in the study of of radio is the ability to look at single so we drop down at times t to data starts over there just flows and flow graph until it so it goes into that same know what's inside of this we just doubleclick In the case mouse the we had to put in a bunch of different parameters so we get we get to set the date no additional work off of of floating point numbers of often you'll see complex numbers used in radial we spend most of our time doing complex math and mathph workings of complex numbers here we're simplifying by just looking at a real number of floating point number now the trick here uh and there's also a repeat so where to send out a vector sorted generated vector passing a flow graph boring repeat that vector over over years the thing is in constantly repeating itself and what the value that in go into here is based on this call on so parameter so it's a variable as 1 stood to know here is that this is actually Python code to do a lot of programmatic stuff directly in the US in this set of you here so con is a variable that was set up in our manipulate specifically for this purpose of studying this sharp rise times the factorial from nothing to the connection Regina by having a DUI interaction of interactive tool that cold enable sorry that's this is actually unnecessary hassles stable that that's another study come on is this variable here for a kid to do we check box which is you know what to do we element a check box or a check it to turn on uh so if it's if it's checked its Empresa value of 1 in logic the person value 0 and that's going to update that value in the vector source so the vector source 1990 producing zeros at increasing ones pretty simple but conceptually but I just want to make sure that we understood how we can kind of use variables in these parameters to to go back and forth so and execute this graph no you can actually see it on the screen there is a tool with which tool bars of topic or you have 6 and it executes a flow graph so this is our time sink so data is being generated this is constantly updating the the data generation and printing it to us to the time sink so just seen this 0 0 along here is the time axis and here's the amplitude access the taxes the the what is the purpose yes it's it's it's scalar offset on purpose but then I'm going to suggest entering zeros and as I can get the connection here now that she captured that so if I if I turn something off here this is another thing I want to point out here there's a lot of interactive tools that we can use so what I set up here is a trigger to trigger like oscope trigger normal trigger and so if when a single rises goes from 0 and rises that triggers the scope to to to to take a capture and hold that capture and so but if I turn it off if I just got a free running now just constantly free running because no change in state so just it doesn't care so you can turn on to read off back on again but but also 1 that's happening listen to if you can hear it you just a little click right that's this times this kind of extra currents concepts and so we can have this kind of interactive of study to to see it to play with it after the tree about on here because again what were studying there were studying is this phenomenon that happens from going from 0 to a signal very very fast in this case almost instantaneously facts that the sampled signal so from 1 sample or a 0 the next sample so it's not an it's not infinitely small but it's it's a cells we can get there in time in the sampled the signal so what was causing us that's kind of the question that that people were asking and how we use this to our advantage what kind of things here we learn from this so stop bps go back here so as
09:34
I said rebelled can discover the fact that we can treat the bandwidth over wires so there was this idea that that you have multiple tones going at 1 time I we study this the possibly
09:47
1 I mean it's 1 of the most important principles in of Electrical Engineering today is a constant the Fourier transform of this ability to convert a time signal into the frequency domain representation of it so that he worked out with different field actually but it's been applied to electromagnetic waves of primitive sense so would ever take the sign thing with very short duration time signal this rise time were to try to study a from a frequency perspectives and that's something that they didn't quite understand the the late 19th century while was this low was be able to use this property here but we know how to use it so as to what what's the use of our tools here is this is another as I said the ability to look at signals and look at them in multiple domains is a really important part of the thing when studying but when studying signals and radio the so drop back into are radius system
10:43
here go into the next flowgraph which is just slightly more complicated there will be more processing and added differencing so now instead of just looking at in time really get in time and frequency that frequency sink is action performing the Fourier transform so to take the times England representing in frequency I'm really powerful from an analysis point of view of and it's also pretty much the way that use the this Fourier transforms using pre which every cell phone being sold today Ltd is fundamentally based off of the Fourier transform sum of massively incredible of a piece of mathematics into this the and doing this can adjust for just to show a 1 property of signal processing is I've also done today of FIR FIRfilter finite impulse response filter filtering is again another incredibly important concepts of our our but I tool in a signal processing capabilities not only is it a great tool it's also great formed wait study systems because pretty much everything is a filter you put a single through something that's that's something that hardware or whatever it is is going to filter because it has a bandpass response a just passed every single frequency component from here to to gamma rays is going to have some kind of a shape to it possibly not even of flapless sometimes an uneven shape to it so this is what audio people get really upset obsessed about when the with a response of the speaker system the microphone is getting the filter shape as smooth as possible so I throw a filter here because we have to be a limit saying we want the single the band limited so I put that up there idea don't or too much about the details of the alsatian filter design tools at the ends but again you can study how we do this but this is just linear lowpass filter so now a look at this signal this the same concept and triggering off of this of this rise time connection here and I wanted to make the connection the 1st of notice that instead of being the instantaneous in a rise the filter gives a little bit of of of delayed response on a shape to that so this is again if you've ever looked at singles on an oscilloscope you've probably seen the phenomenon of very well but but popular famous mono but also the same time not only were looking at this in time world so be it animal to look at this and frequency so this in the frequency domain representation of that signal what is signal what is the transition period itself so it's not just the 1 but there's a lot of 1 it's just a single that's just on has no frequency component in that always on states of DC term rights 5 volts or 1 volt whatever we want color as so that's why we have this really large spike at the DC term of the of the frequency domain and again unfortunately the the screen is we're here so all these parts are gonna show you positive and negative frequency so this is minus and they freeze over here right in the center here is the DC terms this 0 0 hertz the so huge up a DC term because we just turn on a DC signal us as we see here but again the important concept here is a short time in a short amount of time long frequency response so with this signal is also this do is also done is captured the max hold so I could turn on the max holding see because it's very little amount of energy happening very quickly over a long period of time so very well a lot of energy very little talk of power has its is happening over short period time so this curve here is kind of the of the Fourier transform over a certain number of points that we can see the effect of a very short period of time so this along response and frequency I remember I filter so the filter kind shows up here at the edges to the variable that very loose filter but it's kind of rolling off their edges so this is the idea that so that is guys had when they were trying to study electric waves OK we know these extra currents are producing this of a response and and frequency what if we could capture part the frequency response so we know how to generate tones at a very high frequency because these impulses are very very broad bands but how we capture it the so going back to our slides here
15:15
a story comes to another famous mathematician of the 19th century days work Maxwell and the famous Maxwell's equations this is where this thing concept of things come into play here a lot of the people at the time the sciences at the time again didn't understand this time band which the Fourier transform relationship that we were able to just look at in the so he comes up with his theory mathematics behind the theory of Verity's induction so the idea of inducing a current in a system so the change in bandwidth over are searching and in that the magnetic field over time in this equation is related to the generation of the electric field on so savages where magnet form of something and you get a current it's the change in the magnet that induces the current are somewhere so this concept of time of a change in time is a core structure related to frequency so the match well user people who believed natural equations so who understood the math understood this concept that this rise high rise in strippers short time period rise time based on equations is going to produce a high frequency response you know they had this theory has understood the question was how do we actually
16:32
study this now for the city's sake of time I love this so this experiment but only for the sake of having to try to shorten my my discussion of on this is a light and jar now Elidan ja it's did specific design but will we we know it today as a capacitor so they charge this capacitor up with a static electric charge and and so I could build up thousands and thousands of volts on this thing and discharges and so it's a spot into static shock i is just a large set attractive and on how much you of charges of and up when you're playing with this it starts to hurt after a while so you come up with other was doing it but this is an old 19th tree of a tool for the science is used to study electric and magnetic fields of electricity in general static electors in general so I got warm I this is a present last year and so I play with that and I want to see OK what what does this simple like you know I talk about this guy I know this impulse in this math and all this stuff but what if I had she was able to capture this and see what this the single look like so I charge this up and I use this thing this this is a beating 10 this is new star US API when the hardware systems from this research that we use that out of the 200 summary here so that we use in new radio that have to do signals of so we can receive with we can transmit with it so in this case I'm using this 19thcentury tool that to generate its own in the 21st century tool to receive and capture that so so that was really fun now well just point out here is without going into the details of why and this is perfect I mean this was so amazing amazingly perfect that I was surprised that it looked as good as it did when I capture this is a real life capture and that is a filter impulse response is this actually these I did a lot of time studying ultrawideband of economic in grad school so this is this is the of mathematically very significantly but this is the thing that they were able to do back in the days to generate these very sharp impulses and again knowing equations we understand that they had this long bandwidth of response time so this is actually captured at 2 megahertz only because I don't feel like going up my entire disk I would like us 32 mm example know 32 merits of by time which would give me a relation an even sharper poles so the I
18:55
want to 0 so this is an inspiration for all for studying the canon
19:00
followon effects and all over large but of course the was 1 is the the the the highest or most well known of the Maxwellians I love this code of here is that this is really important thing to get back to the ability to use our tools to visualize signals so the this ridiculously easy right remember I said you just plug something in and that's generating high frequency so they knew that it was ridiculously easy generated but he had actually no idea how actually had a receive how to prove that was being generated at was so that's where we're trying to
19:35
head sooner seconds now remember I said everything is bandlimited everything is a filter like almost like in you know without without that too many other things like the of off I had everything in the world is basically a filter story users were a play with this filtering concept but in this generation here this flow graph called impulse
20:00
generator i is OK let's take this light in jar experiment and try to recreate it would simulate this this experiment so again fairly easy to be able to generate an impulse because it's just a series of zeros and then over here you know 0 1 so in discretetime sampling here so just general 1 and a bunch of zeros and we discoverable repeat that so we can see this phenomenon over and over again so regenerate that by of using this form of the for the vector source so 511 zeros followed by 1 followed by 512 0 so 1024 point vector this is really a repeated over and over again and again just from really the best place on as a place on expression basically to concatenate a list but ah receiver obs our systems all have noise so I wanted to simulate the noise so I'm using this this fast noise generator and I just have approximated the amplitude to get it to look pretty similar i and it's goes into the guy was you know a source random generation dousing variables so we come up with a term additive white was noise sort take goes you noise that's pretty much white over the frequency are or bandwidth and just added to the signal become ones we don't and 0 we just we just that adds noise to it now here this is the sum the detail here they're are called the throttle so I don't have it in the the last fluoresce because they had all you saying cooked up that audio sync has a sample rate With understand samplerate really do anything with digital signal processing so sample rate of the audio card was of reveals a 32 kilohertz or something like that it's is how a program that but so that was X o'clock in the system is that audiences and can only push data out at such a rate based on its 32 kilohertz clock so radios using a push data to it until it's buffer gets filled up and there is a whole back and then as usable for clears up we push more data into here I this is pure simulation unjust generating samples as fast as my computer can generate them and emission display them in time and frequency then nothing giving me so if I just launched this without something that that's kind a holding the system back on its you know go really really fast and and if you have too much in your flow graph if you from you too much but it will try to it will eat as many resources in in CPU cycles as they can get from your computer so the throttles an arbitrary way to hold us back to can smooth out and and slow things down it is a very bad throttle muted bad heating system it just uses the system clock which is not that accurate more than like you MS if that so it's to make some sitting here to make samples per 2nd plus or minus the reality of the actual clocks but there were just using this for it as you put Hardware in the Loop pursues we had the audio system there or radio get rid of the throttle because now the throttle is acting as a competing Clark and see of a single the 2 o'clock problem that comes up which you guys are trying to independently of on a sample of your system and because it's a bad 1 it's going to perturb your your ability to the receiver transit properly so this is pure simulation here I start this knowing and the filter instead of doing a of a low pass filter and actually started and I'm using here a uh gallows unitary field the reason why this is because I want to have a 1 to narrow on near the signal a little bit so we can study or a receiver in a 2nd but also as I said I studied a lot of ultrawideband I about pulse generated ultrawideband warera doing 4 gigahertz wide pulses and but it was generated by goes Yamano cycle so goes in motorcycle and time take in the integral that the fourier transform it's a galaxy in frequency as well but hopefully this is the most Mathilde given to us today so I just want to get as a filter is just what makes it infinity if it's the story that but here we have the impulse going up there but every 5 under every 1024 samples we have this 1 and then it's getting a distorted by some noise and filtered by this guy was in filter response so this is this is my my way of playing with this idea of the impulse generation
24:38
but the thing thing else a point this cell again although over of an easy way to look at a filter generation and a few minutes about or using a tool here to just generate a guy was wasn't filter on it yeah I think that's pretty much all 1 learned from this so this 1 over
25:01
token so so in this case on on our experimenting like it into a megahertz signal spectrum of signal Aguzzi filter and now the trick is to be able to sample side we actually get this this signal back and that's where this guy comes in that he was really the motivation for this talk is to study how it hurts do his work was the original of experiments are stated and how can we use them to to study us off a radial arm and Lin research in Israel and flux this is a quote a 1st I just love the most somebody after his 1st lecture at Karlsruhe uh sparser university at the time and he demonstrates that he is able to do these electromagnetic fields and so he says that's amazing you view base you've proven natural right you know everything works is predicted what good is this and his response was nothing I guess the the 1 of the most fundamental or important but commercially for sure so you know that the discoveries in in history amusing billion dollars cellphone industry coming goes back to this thing which was just a lot but just for fun so what did he do
26:19
how did he do what he do that lodged in understand cause lodges he that didn't have any clues how to how to actually capture that single being generated and that's where that with this apparatus does so the sparkgap transmitters so you get this so this transformer here a high voltage and then there's a spark gap that gets generated at approximately 100 millions of times per 2nd so roughly love 1 sitting at a carrier frequency of 100 megahertz the generated spark around there that's a hundred times a 2nd you get this very broad band signal being generated that signal is then transmitted over this dipole like antenna structure on those used to come loads that he had to that kind of approximate 10 them of early form of antenna but but that's why i want to filter the signal because this is a filter right I mean even though this is generating a broadband signal the response the of the that electromagnetic properties all of this load system here this antenna is going to have a filter responses so in in all in a way when you're studying antennas you can study them in terms of filters and how they behave Africa's to what kind of shape they're to perform at frequency and so it's very actually very difficult to make broadband of antennas because they all have a shape to them and we figured out of doing pretty well but it's still not that easy a dipole tends to be very very structured around uh around a certain wavelength of the signal George entering the filtered signal and all its receiver was doing is just as large a loop here and you can see you know it when the signal when it when he match these he's doing matching of filters so In other words he's doing impedance match so once he couldn't be here could figure out what the impedance matching was to generates a single here you get a spark between this 2 this this 2 balls at the end of the attack so for all this work a friend of trying to receive signals how we look at signals that we know they were doing it the entire receivers fairly trivial said that he had to
28:22
do it the right way so your it is a little more interesting in a radial flow graph
28:29
or a to not only generate an impulse but what we do receive it and receive it by and all of this but of approximating is intended as a filter and a flow graph so also 1 thing went out instead of having all these of orange the ports here so these of course so we connect I now switch them in and so the scenery in to connect up a lot you just have to hit 1 port connecting and Princess connection the blue now we're now doing complex numbers and because I want to do a complex bandpass filter the door much but that if you if you don't go into too complex of numbers adds 1 to point out why the the change from the color so the visual you'll see what's happening without having to really understand why I like as early as a question of anybody wants but for just go from a from here so that top section there is very similar what we had in the previous example notes complex some generating complex of an impulse adding some noise to it when it goes you fill from such a transmitted signal this filter is a receiver that some generating is received taps so notice here just put in this variable or stats but which I'm defining up here yeah and if I can expand this we can see a little bit about what's going on here theory some using again our filter design capabilities and in radio using this API call on of the library and I'm passing in a bunch of parameters so we have something to create this receiver so we have this filter so i come this bandpass filter but with these different parameters being set now the band within the transition bands that I've got said here are variables are defined elsewhere and apparently not that here as so I define users as of a static values so I've got a bandwidth of the antenna itself is going to have a certain bandwidth based on this number here uh transition and again if you are familiar with filter terminology is just how the shape this how the role of of the filter looks but the important but you useful tool is this a range block so again we remember we had a connection that the check box to turn something on and off this is now a slider so I can put down a slider onto all of our application and don't don't worry too much much of a pass that is I'm just giving a range just so we can go with just slide this range around the frequency domain the net update the filter as the system is running so the word BCB tuning are antenna like alive tuning of the intended to be 0 the signal so I run this but it seems to turn off the imaginary parts so just go we'll look at a real signal here so the blue signal on the top there is the of the transmitted signal that or trying to capture of additional again as we can see or try to capture the green is a receiver so this was the difficulty was that I missed to my intent is now is is not 2 to the right frequency response as so I'm getting very little of signal coming through here the so we can see this in the domain but this blue signal is the transmitted signal and the red is after going through a receiver filter armies tuned receiver filter so 1st of all notice you know yes we're seeing single here is no filters perfect there's always energy leading us through filters but this is in dB like dB relative to the old 1 floating point number so think of it as a DDM if you're if you're into long but like injury you guys here will will easily understand as deviant it's a log scale so we're going from minus like 45 thousand minus like 1 20 in log scale that's a huge decrease in power so we can see the power from the the transmitter of in the blue and even though there is power coming through here so it's it's not really showing up in the time that it really is suppressed so much in in power but there's this little guy sitting over here and that's my filter that's my bandpass filter that I've created when I started it the tones down here but yeah so I've got this center frequency slider here and I'll just use the the buttons here too so I can move the center frequency that filter some of what I'm recreating the filter taps but to generate a new filter responses see Batman truck so as I as I change as I as I increase the filter on center frequency you can see how the can I keep this yeah so area as you can see how the response of the incoming signal is now changing so no more energy is coming into our our receiver and notice the green there is starting to come back up as we're getting more and more of the energy in it so if I can perfectly tune the close to it yeah close enough paper for change my receiver antenna Viking create the loop the diameter then that antenna is going be at frequency response I can place this to my receiver in the center of the transmitted signal and captures much energy as I possibly can so those are the filter the receiver has a smaller bandwidth the the transmitter I'm some just run a center is again as much of that energy as possible and to notice 2 things the green line is slightly smaller than the blue line that loss of energy loss and it also delayed in time that is that the filter it has a is a time delay there's a group delay that a filter as it takes a little bit of time for the signal to pass through the filter and just yet is same as a as an antenna so here we have the ability to us to tune our system of well it's running and play with this idea is to to see how we can capture the simple so again I think that's the last we'll need for from that
34:55
of yet is just goes over well that's what I said about tuning this tuning signal there is
35:02
that is just for us later so of about 5 minutes of it so I just wanna pop this up really fast and but again we did a lot of of work here again I showed you a lot of things that have filters being generated so again tho there are hugely important when you try to do any kind of signal processing and is a few kind of gotchas that we need to to build play with
35:26
so I can do this so if we're able to see the tool bar can underlie the resolution is doing this but there's a Tools menu bar and as a filter design tool that can you read you has becomes a larger from for my desktop here so the would if area so we can do here is play with and designer filters we can we can study the filters in different ways with this tool also so we understand how what the the consequences or filters are so on I'm not going to details for filter you Hamming window versus Blackman window things like that was just look at it from a lowpass filter perspectives or a pass so the passband discovered pass all frequencies up to 50 thousand kilohertz and organ attenuate the single afterwards or the filter that you have 10 kilohertz to stop the the passband and slowly due integrated into the into the noise for go down here to the design and there's a filter so that a filter looks like passes everything and this is the frequency response so there is multiplying the frequencies together so anything being multiplied in the past is being multiplied by 1 and then as we go into the transition band down to stop were multiplying by a very low number in DB source surprisingness of the symbol of dramatically something to keep in mind here when you're playing with these filters in again this is a it's a very common of of of misunderstanding for people just getting into signal processing adjusting into radios is a consequence of your filter so in this case a consequence of the filter is that I have to that the number of taps this filters 59 so I have to do 59 multiplies and adds to to compute the filter so the longer your filter those struggle although the more heavier it's compute power is going to be of we want we want to minimize that so how we minimize the number of taps 2 things we've covered 2 degrees of freedom here and all at play with 1 for today but if I go from 60 case I myself 10 K to go from 1 to 0 that's enough I do 70 jive 20 K and now relaxed my filter from a signal processing perspective enough I designed this 1 I want from 59 capstone 29 taps so I have to do 30 fewer what line at computations that to run this filter but notice that it's it's it takes a longer bit of time so it's a sloppier filter which can be OK right that's the thing that I have a lot of people that come to us with problems Why does my flow reverted sticking to what you know it's the same many habitations it's because a lot of people than something like this so 51 581 taps but that's a really nice filter is an right I'm not getting any other single in the but people don't you know you you have to study have to understand the response of the entire signal in the system that you're working with because it can be OK to have a long transition and with that images gonna pull an extra energy middle of extra noise on and maybe a little bit of interference which have to be lot to to balance these ideas of the cost of the computation pursues the cost from a signal processing perspective and this is the point of this tool because only should you know and knowing and exists you can use it to study your filters this go back to a cheaper filter and notice that like we can go over here we can actually see the taps so that we look at it from different domains and our moving into the timedomain here the the phase response and the group delay sum which is the phase response here but the delay of the filter 29 tap filter 27 tap filter is gonna have a delay based on the number of taps because you're kind of moving through the steps 1 at a time the latest signal out to 27 taps obviously a much shorter delay than 581 taps so again that's another tradeoff that we need to be be aware of 5 is that the way that can be produce in our system and so there's plenty of of of other things to play around with from the sourcefilter lecture than a lot more time in there but I think this is a really useful tool artists to play with to to learn about filters but also when you're when you're actually designing a system in you actually want to design a filter by your this all the time and there is actually to save this filter I just as you use go to file save and exports as a CSV file it'd be more clever but everything receives the there's not that much of the that capture here about but so you can save it by your favorite filter that you've designed here and and and pull it up later all of the design methods that are being used in this tool are also programmatically available in new radio itself so I know you I pull that operation differed as . galcians so that's basically if I change this the only moves so were there so I connection create that goes in filter here with the magnitude response so there's a guy filter so this is just a graphical interface on top of our of our called a function calls within the and they do a lot of our of of playing around from our perspective OK so
41:10
Ingersoll of that's and
41:14
OK uh primary time here so just
41:17
unit so there's there's there's some slides at the back here just point them out of kind of talking more about the project itself that there's you know this community in this this a whole bunch of these out of 4 projects of people building projects against the radio try to show you so those the list of possible block categories that we have available to the mass of list of available algorithms in the radio already plus this entire ecosystems the community of developers were building their own set of tools and blocks and algorithms we tend to be generic try to be generic inside the radio so very general stuff filters adders multipliers us synchrony of some things that can be used against many many signals the auditory projects tend to be focused on particular signals LTE Bluetooth EDSP those types of signals as we try to the very particular inside a radio arm but I want to point that out before going back to the final demo part i'm because I want to make sure that that we understood everything I was doing up here was in simulation the is what the audio portion of it but as I
42:31
said we also this ability to connect to hardware we have these hardware tools like the use and other or devices that are to us your dongles so that we can connect into it so we can go from our generation of signals here and you're generating 2 singles at different frequencies and I just want you sideways because well we're in this talk was inspired by hertz and we that's the unit that we use to our to measure sine waves of frequencies the hurts somewhat flying in our yeah multiplying 2 singles together just to create a weird singleelement transmit Over the years and I put down this block here this usurp St. FIL transmitted over the years so the runners to the global time to sell to load the the so this guy's no transmitting at 428 megahertz a fairly small amount of energy from that increase just enough so some just transmit generating a little bit of sine here where the be 200 actually need to have given a lot gain there we go some generating a single here and so I have the radio now running on Android were employing the signaling through the few so just of the Oculus your dongle of connected through the radio this is actually filtering the signal and this using filter because I want to something more than just you know Our receiving and displaying of yes that's the signal being transmitted from there um so this is our kind of a new project right now see if this a little bit more stable and ready to ship out the door so will be seeing a lot more of these tools to OK all my do we have time for questions I'm only like a couple minutes over so on and a really fun done at time OK so well and but what they term for a transition so today we turn this off and let we just pop up with
44:42
the yellow is but that up and uh so you the other and questions it was so the I wells the band with the signals to the spark gap and some of the like the hurts experiment itself I should no I don't know if we know what the the bandwidth of it is but said he was generated spark at a hundred million times per 2nd and because it's a spa because you know it's it's this instantaneous spark as my my light experiment showed it does have a very broad of frequency response but then it's being generated pasta this antenna structure these loads that he had so it's starting to bear with his you depend mostly on the load of that that structure so we have to study with the material was the size of those loads all that I'm sure we could kind of roughly calculated if we went back and saw his original experiences the even even hurts it's was a few gigahertz this OK so so if we have broadband I precisely yes so no yes exactly precisely yeah yes if the I don't know I so this thing is so so 1 of the the points was I showed in time what that looks like I can't show you the frequency response of the power because it's so quick in time the amount of power being generated is so small that there may be a lot of energy but its energy you know 1 that is actually little energy is well energy over time in power to very very little power generates a probably because most regulators only care if you're right if you rise above if I was generated is continuously that's a profit but just once every like 2 minutes of with of I and and of the what that's right at all I think it's so the problem with the generation of wideband signals a suffer radio so like you sort of specialized equipment you know highspeed FPGA is doing things over USB united that kind of bandwidth and so you can experiment with the OFDM CDMA which is a competing standard the pulse generation for like real true ultrawideband that's not happening in the sites systems they're they're just not designed for uh the hardware for POS generated by uh ultrawideband is very very specific and and kind of highly tuned I the also I there was used about 10 years ago and we just we never really figured out the right application and that the you not only true now would do it would also be able to do in a commercially viable way were receivers and transmitters we just a and we really got there needed it and sell it also maybe in 10 years will come back in fashion you going out the door out I like that 1 more