High bandwidth FM signal generation

[enemra]

I am working on 5G projects. As part of a big project, my requirement is as follows.

1. Generate FM signal with 300-400MHz bandwidth between 500MHz and 6GHz carrier frequency.
2. Generate random noise with the same bandwidth at the same center frequency.

I have come across the following.

1. Using Software Defined Radios (SDR) - SDRs do not provide large bandwidths as per my requirement
2. Using Analog Front End ICs, such as AFE7900 from Texas. This looks expensive solution and tedious

Any other solutions to think of?

Regards
enemra

[A.Z.]

are you trying to build a cell phone "blanker" ?!?

[enemra]

A hardware that can do phone jammer might work for me but phone jammers do not require a large bandwidth like I do.

I am into 5G research, application is not jamming. In my case I also send FM signals.

[David Hess]

1. Generate FM signal with 300-400MHz bandwidth between 500MHz and 6GHz carrier frequency.

Do you mean occupied bandwidth or modulation bandwidth?

[radiolistener]

are you trying to build a cell phone "blanker" ?!?

maybe he is trying to design some kind of military RF jammer to shut down GSM, WiFi, GPS and other communications 

[coppercone2]

guy with 6 posts asking about high power amplifiers and broadband jamming signals what can go wrong?

i like how the first post is basic ass questions made by someone that never used a amplifier before going at 100W, and the next one is about broad band random noise. "heating" in quote marks. I love that one lol. Perhaps its not really heating. What can it be doing? how did someone get into 5g research when something is "heating". does he think we are using double speak or something ? "heating" is such a point of contention with ratings of components in electronics study... it sounds like someone that has no idea WTF their doing. I commonly see components maximum ratings described as being dependent on something thats not related to heat so I will put "heating" like its some abstract complex we came up 

Must be someone from the CB radio community? but those guys would not have a problem understanding heating, after all you just drop the amp in a bucket of water to cool it down.

what a strange uncomfortable user


and before someone smarter then him thinks its a good idea, FM signals are FROWNED upon because it EATS bandwidth. I feel like a spy would like this technology if they think someone is snooping on them

[pdenisowski]

1. Using Software Defined Radios (SDR) - SDRs do not provide large bandwidths as per my requirement
2. Using Analog Front End ICs, such as AFE7900 from Texas. This looks expensive solution and tedious

Any other solutions to think of?

A vector signal generator can provide wide bandwidths, provide complete flexibility in terms of signal characteristics (modulation, noise, etc.), and typically have a relatively simple user interface (but can also be controlled programmatically and/or can use signals created in tools like MATLAB).

That said, they may be slightly outside of your budget

https://www.rohde-schwarz.com/us/products/test-and-measurement/vector-signal-generators/rs-smw200a-vector-signal-generator_63493-38656.html

[pdenisowski]

A hardware that can do phone jammer might work for me but phone jammers do not require a large bandwidth like I do.

As someone who has (professionally) hunted down and analyzed illegal jammers:  most phone jammers are not precision devices where the jamming energy is strictly constrained to the band(s) of interest.  Many "multi-band" jammers may have multiple antennas (one per band) but actually just put out continuous noise.

In fact, what caused most of the people operating jammers to be caught was that they thought they were only jamming one narrow part of the spectrum but ended up jamming spectrum that "belongs" to someone with the resources to find them.

With regards to jamming cellular networks (LTE, 5GNR), it's often sufficient to simply jam the resource blocks / BWPs of the control channels (e.g. PUCCH).  I've even seen this happen unintentionally on some occasions - narrowband, non-malicious signal fell directly on the control channel and the eNB could no longer "hear" the UEs' control channels.  I did a presentation on this at a conference once, but I think it was one of those conferences where the papers aren't available on the web

*Cellular network operators do have the personnel and resources to find people jamming their networks - I worked with them for years - but most of the time this jamming is unintentional

[radiolistener]

Cellular network operators do have the personnel and resources to find people jamming their networks - I worked with them for years - but most of the time this jamming is unintentional

if I understand correctly, for 5G it will be more easy than for 2G/3G/4G, because they using phased antenna array to rotate beam, so it will works like radar and allows to get more precise position of transmitter with triangulation, isn't it?

Just wonder is it possible to use some kind of synchronized array of transmitters distributed all around the base stations and transmitted signals in that way so their interference signal at base station location gives false position of the transmitter or just spamming it with a bunch of virtual transmitter positions to make it hard to find which one is real? Are such kind of jammers are used in practice?

[pdenisowski]

if I understand correctly, for 5G it will be more easy than for 2G/3G/4G, because they using phased antenna array to rotate beam, so it will works like radar and allows to get more precise position of transmitter with triangulation, isn't it?

Steering the beam might allow for a rough determination of where in the sector the interference is coming from, but I'm not sure if any network providers who are using it for interference detection.  A more likely application is using it for interference avoidance (in the same way that "tilt" is used in pre-5G).

Incidentally, "triangulation" as a DF or interference-hunting methodology is, in my experience, both very overrated and very misunderstood.  Unless you are in a more or less reflection (multipath) free environment, it's very difficult to get good bearings: calculating the interception of those bearings is trivial, but junk bearings yield junk results. 

Since interference / jamming happens around (or at least is important around) people, this means most practical DF'ing is being done in urban and suburban environments where multipath can be an issue. 

For cellular network operators, they know from the base station stats (RSSI, e.g.) which sector or sectors are being affected, and in most cases you can simply drive (or walk) the sector until you get close and have to hunt on foot (where triangulation is completely useless).

Just wonder is it possible to use some kind of synchronized array of transmitters distributed all around the base stations and transmitted signals in that way so their interference signal at base station location gives false position of the transmitter or just spamming it with a bunch of virtual transmitter positions to make it hard to find which one is real? Are such kind of jammers are used in practice?

There are a lot of "creative" approaches to jamming, but high-end direction finding systems (like the ones we make) can usually still DF people trying to use "creative" techniques.  Generally speaking, for a jammer to be effective it has to be (a) loud, (b) wide, and (c) on, and all three of those things make jammers relatively easy to DF, regardless of how they are implemented.  A weak signal with a low duty cycle that's only a few kHz wide is harder to DF, but it's also not an effective jammer.

[mawyatt]

Check out the DARPA program Computational Leverage Against Surveillance Systems or CLASS.

https://www.darpa.mil/program/computational-leverage-against-surveillance-systems

We developed the computational chip over a decade ago for CLASS.

Best,

[enemra]

Nope. If thats the case i wouldnt come here to the forum. Moreover, I should be very strong enough to understand such stuffs if it is military.

[enemra]

Guys, thank you very much for the answers but a humble request - please do not post answers in the context of jammers or do link my previous posts like the user coppercone2 being smart. I am not working on jammer.

It is all about I want to create interference and come up with resilient communication solution for space-terrestrial based broadband communication. This involves a tiny band of frequencies over 400MHz bandwidth.

@coppercone2 (smart guy): My amplifier post was for a different frequency and does not involve FM there.

[radiolistener]

Check out the DARPA program Computational Leverage Against Surveillance Systems or CLASS.

https://www.darpa.mil/program/computational-leverage-against-surveillance-systems

We developed the computational chip over a decade ago for CLASS.

is there any public available detailed info about used approaches in this program?

[pdenisowski]

It is all about I want to create interference and come up with resilient communication solution for space-terrestrial based broadband communication. This involves a tiny band of frequencies over 400MHz bandwidth.

Most people who are trying to test reliability of a system over a non-ideal radio link will use AWGN (additive white Gaussian noise) or a single CW carrier with a given carrier/interferer ratio and at a given frequency offset from the "wanted" signal's center frequency.  The next step in realism is adding fading / multipath, defined in terms of "taps" with different multipath profiles.  And finally, if motion is involved, applying Doppler shift to the signal is also desirable.

Again, a properly-equipped vector signal generator (like our SMW200A) can add all of these impairs in a user-defined way to the baseband signal in realtime before up-converting it to RF.  There are some other manufacturers who do this at RF, but this is much less flexible and much less controllable.

Hope that helps!

[mawyatt]

Check out the DARPA program Computational Leverage Against Surveillance Systems or CLASS.

https://www.darpa.mil/program/computational-leverage-against-surveillance-systems

We developed the computational chip over a decade ago for CLASS.

is there any public available detailed info about used approaches in this program?

Best check with DARPA directly, we can't disclose details on most things for obvious reasons.

One thing that we can disclose is the ability to create dynamic in real time and position elements for phased array beam forming. Think of this as a large physical area and number of moving in position and time phased array individual elements collective forming together in real time for massive beam forming capability

For example, imagine a large number of smart-phones moving around a large campus area and being elements in a vast synthetic aperture phased array able to dynamically beam form in real time while all the elements are moving around randomly!! You can imagine the enormous computational power required for this to work, yet all was condensed into a single ~1 Billion Device chip over a decade ago!!!

Best, 

[pdenisowski]

Check out the DARPA program Computational Leverage Against Surveillance Systems or CLASS.

Interesting that (to me) at least some of this seems like "security through obfuscation."

1) Waveform Complexity uses advanced communications waveforms that are difficult to recover without knowledge and understanding of the signals itself; 2) Spatial Diversity uses distributed communications devices and the communication environment to disguise and dynamically vary the apparent location of the signal; 3) Interference Exploitation makes use of the clutter in the signal environment to make it difficult for an adversary to isolate a particular signal.

As for "dynamically varying the apparent location of the signal" - I'm not sure what that's supposed to mean.  The apparent location?  I'm guessing they mean having multiple, geographically distributed nodes all transmitting simultaneously on the same frequency, but there are also ways of DF'ing those kinds of signals as well.

https://www.rohde-schwarz.com/us/applications/super-resolution-df-method-application-card_56279-199552.html

[pdenisowski]

For example, imagine a large number of smart-phones moving around a large campus area and being elements in a vast synthetic aperture phased array able to dynamically beam form in real time while all the elements are moving around randomly!! You can imagine the enormous computational power required for this to work, yet all was condensed into a single ~1 Billion Device chip over a decade ago!!!

And how are the nodes (a) precisely time and phase synchronized and (b) controlled ?

I understand the theoretical part of this, but fail to see how it could practically be implemented on  a large (kilometers) geographical scale.

[mawyatt]

Check out the DARPA program Computational Leverage Against Surveillance Systems or CLASS.

Interesting that (to me) at least some of this seems like "security through obfuscation."

1) Waveform Complexity uses advanced communications waveforms that are difficult to recover without knowledge and understanding of the signals itself; 2) Spatial Diversity uses distributed communications devices and the communication environment to disguise and dynamically vary the apparent location of the signal; 3) Interference Exploitation makes use of the clutter in the signal environment to make it difficult for an adversary to isolate a particular signal.

As for "dynamically varying the apparent location of the signal" - I'm not sure what that's supposed to mean.  The apparent location?  I'm guessing they mean having multiple, geographically distributed nodes all transmitting simultaneously on the same frequency, but there are also ways of DF'ing those kinds of signals as well.

https://www.rohde-schwarz.com/us/applications/super-resolution-df-method-application-card_56279-199552.html

Imagine the "apparent beam" location being dynamically distributed spatially, even such that it can be "positioned" on an adversary.

In the battlefield an adversary fires a short range radar tracking missile which is quickly sensed and the apparent missile tracking radar signal is dynamically realigned to the missiles origin!!!

Best,

[mawyatt]

For example, imagine a large number of smart-phones moving around a large campus area and being elements in a vast synthetic aperture phased array able to dynamically beam form in real time while all the elements are moving around randomly!! You can imagine the enormous computational power required for this to work, yet all was condensed into a single ~1 Billion Device chip over a decade ago!!!

And how are the nodes (a) precisely time and phase synchronized and (b) controlled ?

I understand the theoretical part of this, but fail to see how it could practically be implemented on  a large (kilometers) geographical scale.

Well that's just one of the many "You can't do that!!" parts of CLASS!!

The original solutions came out of MIT Lincoln Labs a couple decades ago, then DARPA developed a program to implement such with a few other "You can't do that's"

As you can imagine the signal processing is quite involved, and as the DARPA site mentions requires 1000X our processing power to unravel, and our chip was already at the edge of "Supercomputer" performance (mind this a dedicated processing chip, not GP)!!

Edit: Sometimes things are going on "behind the scenes" that seem impossible to knowledgable folks, take for example Stuxnet, or the half century old GPS position recovery from GPS denied locations, reading fiber optic cable data non-invasively, or more recently the PolyPhase Mixer (PPM), which breaks mixer theoretical noise figure limit below 3.92dB, and a bunch of other stuff that's not for open discussion!! 

Fondly remember discussing the PolyPhase Mixer (PPM) with a couple DARPA Program Managers at an IEEE Conference, both were well seasoned RF/MW/Signal Processing Expert Professors. Being from DARPA they are always looking for "You can't do that stuff", skeptical but took note, and later we had a program to develop the PPM. We were told to present initial work at MIT LL and later DARPA created a workshop at Headquarters where there was standing room only. There was quite a bit of skepticism involved since the PPM violated conventional Mixer theory wrt Noise Figure.

After answering a few questions and finishing (thought so), one of the most notorious intellectual shedder professors (Distinguished Chancellor West Coast Prof) siting in the front row hadn't said anything, then said "Mike explain to me why the NF is below theoretical?" So we had a back and forth discussion about my reasoning as to why, which he initially whole heartedly disagreed with, then finally said OK I can accept that explanation!!!!

My public intellectual execution had been "Stayed"

https://www.eevblog.com/forum/rf-microwave/polyphase-or-n-path-mixer/msg3381802/#msg3381802

Best,

[David Hess]

Incidentally, "triangulation" as a DF or interference-hunting methodology is, in my experience, both very overrated and very misunderstood.  Unless you are in a more or less reflection (multipath) free environment, it's very difficult to get good bearings: calculating the interception of those bearings is trivial, but junk bearings yield junk results.

It becomes more difficult at higher frequencies where reflections become stronger.  Roanoke style Doppler direction finding systems have a lot more trouble than systems which rely on a beam antenna, but the sidelobes on beam antennas create their own problems.  Even in challenging environments, I had excellent results with beam antennas designed to have a minimum of sidelobes, at which point it was possible to see multipath from things like trees which otherwise was concealed by stronger sources.  With a really good antenna, one can "see" the terrain from its reflections.  Once I had a good enough directional antenna, I could track flights into and out of LAX from their reflections as a form of poor man's bistatic radar.

Since interference / jamming happens around (or at least is important around) people, this means most practical DF'ing is being done in urban and suburban environments where multipath can be an issue.

That was definitely my experience.  Large buildings with flat surfaces make for a challenging environment.

For cellular network operators, they know from the base station stats (RSSI, e.g.) which sector or sectors are being affected, and in most cases you can simply drive (or walk) the sector until you get close and have to hunt on foot (where triangulation is completely useless).

Triangulation is not always necessary.  Once a direct path at close range is available, a rise in signal strength of 6dB indicates that the distance has halved, making an estimate of distance to the target possible.  This of course depends on having a beam antenna with low sidelobes to reject multipath interference.

The 6dB rule applies equally well to reflections, so at close range it may be necessary to back off and find a higher location to try and lock onto a direct path signal.

Just wonder is it possible to use some kind of synchronized array of transmitters distributed all around the base stations and transmitted signals in that way so their interference signal at base station location gives false position of the transmitter or just spamming it with a bunch of virtual transmitter positions to make it hard to find which one is real? Are such kind of jammers are used in practice?

There are a lot of "creative" approaches to jamming, but high-end direction finding systems (like the ones we make) can usually still DF people trying to use "creative" techniques.  Generally speaking, for a jammer to be effective it has to be (a) loud, (b) wide, and (c) on, and all three of those things make jammers relatively easy to DF, regardless of how they are implemented.  A weak signal with a low duty cycle that's only a few kHz wide is harder to DF, but it's also not an effective jammer.

I tried something once on one of our amateur band transmitter hunts which was somewhat effective.  We hunted standard 2 meter FM signals, but knowing how signal strength and noise meters responded, I modulated the amplitude with low frequency noise through an exponential circuit to create a linear response on various signal strength meters.  This made it more difficult for hunters who relied on signal strength to gain a bearing, but had no effect on Doppler direction finders of course.

[ejeffrey]

The xilinx RFSoC dev kits are the usual thing people use when the commercials available SDRs run out of  bandwidth.  The dev boards cost about $10-15k, but give you several channels of several GHz depending on the model.  They also have a "mix mode" where the DAC output can be chopped by the clock to put more of the power in the second Nyquist zone.  That can allow you to cover 0-8 GHz with no external mixer.

It's a fair bit of development work to get it to work compared to an SDR or off the shelf instrument but there are pretty good examples / documentation and they are supported by the Pynq project although I havent used that so I don't know how well it works.