Learning the VIP (Internals)

InsipidMonkey

Essential Monkey
Security
Corgi Lovers
Advanced User
Premium Member
Joined
Mar 22, 2017
Messages
7,700
Reaction score
17,514
Location
New England
The fact that SDRs dont operate at high enough frequencies (yet...) for radar basically ended the whole thought process. I don't have the RF expertise to really solve the hardware puzzle.
Have you caught up on the Radenso Theia development yet?
 

jon5

PSL + (O / U)
Advanced User
Joined
Oct 9, 2016
Messages
735
Reaction score
508
Have you caught up on the Radenso Theia development yet?
No, though AI filter was talked about earlier as a possible filtering technique. A basic neural net based on the frequency differentials between spectrum peaks, would (in my completely uneducated world) a fantastically simple way to create a BSM filter.

I've really thought about just creating an open framework for doing this, and let someone with the hardware experience figure out how to perform a good spectrum analysis to feed the framework.
 

Jon at Radenso

Manufacturer
Corgi Lovers
Advanced User
Premium Member
Manufacturer
Joined
Oct 9, 2012
Messages
2,376
Reaction score
18,145
Age
31
No, though AI filter was talked about earlier as a possible filtering technique. A basic neural net based on the frequency differentials between spectrum peaks, would (in my completely uneducated world) a fantastically simple way to create a BSM filter.

I've really thought about just creating an open framework for doing this, and let someone with the hardware experience figure out how to perform a good spectrum analysis to feed the framework.
@jon5 Like this?
DSC00762.jpg


DSC00764.jpg


Rockchip 3328 quad core 1.5ghz ARM processor (for neural network and Ux)
8GB flash storage
Spartan 7 50k logic cell FPGA (purely for DSP)
Analog Devices 9248 60msps, 14 bit ADC
 
Last edited:

RedRocket

Lidar Shootist
Corgi Lovers
Advanced User
Premium Member
Joined
Apr 10, 2012
Messages
4,665
Reaction score
5,676
Location
USA
How do you even remember this?
Lets just say, some things never change, Stinger is one of them.

I believe Stinger already showed their hand with an irreversible firmware update which patched the vulnerability, shortly after this thread was made. (I could be wrong, maybe it was reversible, point is - its likely that future installs wouldn't be supportable)

At some point, i decided that if i was going to put any effort into programming software for a countermeasure device, it would be best to design something more generalized to be run on commodity hardware.
While i think that would have been popular among enthusiasts, i don't think it would end up being a worthwhile endeavor.

Originally i had though we could use an SDR to perform spectrum analysis, feed it to a filtering framework, and then feed the results to an HMI [whether that be a BT device, or a touch panel, or whatever]. The fact that SDRs dont operate at high enough frequencies (yet...) for radar basically ended the whole thought process. I don't have the RF expertise to really solve the hardware puzzle.
Yes- it's true that SDR has at present an upper limit of ~9GHz...
The easiest solution for the "workaround" to this upper limit dilemma at present time is the "heterodyne" principle, which is exactly the wise choice that Radenso chose. ;)
 

Alfabert

Learning to Fly
Beginner User
Joined
Jun 5, 2020
Messages
54
Reaction score
80
Location
Kansas, near the Wakarusas
@jon5 Like this?
<pics snipped>
Rockchip 3328 quad core 1.5ghz ARM processor (for neural network and Ux)
8GB flash storage
Spartan 7 50k logic cell FPGA (purely for DSP)
Analog Devices 9248 60msps, 14 bit ADC
yeah exactly. Would you kindly give me a pointer to where I could buy one of those ASAP?
 

jon5

PSL + (O / U)
Advanced User
Joined
Oct 9, 2016
Messages
735
Reaction score
508
Yes- it's true that SDR has at present an upper limit of ~9GHz...
The easiest solution for the "workaround" to this upper limit dilemma at present time is the "heterodyne" principle, which is exactly the wise choice that Radenso chose. ;)
There are a few SDRs that could run in the X and Ku Bands, but K and especially Ka seems way out of reach for the time being, understandably so, it's already pretty amazing they can get them to operate at such high frequencies.

Of course heterodyne is the classic radar detector technique. I tend to dislike not having a perfect snapshot in-time. Notably i'm not an RF expert, my original intent was to piggy back on the work already done by those experts and clean up the software side of the house, and introduce a way to really engage the enthusiast community.

@jon5 Like this?

Rockchip 3328 quad core 1.5ghz ARM processor (for neural network and Ux)
8GB flash storage
Spartan 7 50k logic cell FPGA (purely for DSP)
Analog Devices 9248 60msps, 14 bit ADC
Yep, though my original thought was more along the line of Stinger's architecture where the work is offloaded to a separate device (their Antennae). By the time it were to reach the CPU it's just a digitized spectrum analysis, fed to software for filtering and presentation.
Obviously every engineer has their own vision of the best way. I really respect stingers [largely] decentralized design.
 

Jon at Radenso

Manufacturer
Corgi Lovers
Advanced User
Premium Member
Manufacturer
Joined
Oct 9, 2012
Messages
2,376
Reaction score
18,145
Age
31
Yep, though my original thought was more along the line of Stinger's architecture where the work is offloaded to a separate device (their Antennae). By the time it were to reach the CPU it's just a digitized spectrum analysis, fed to software for filtering and presentation.
Obviously every engineer has their own vision of the best way. I really respect stingers [largely] decentralized design.
Isn't what we are doing more decentralized though? We have a separate antenna, separate digitizer, separate DSP processor (FPGA), and separate neural network SoM. Each of the components can be run independently with any other RF system as long as you know how to send spi commands. It seems to me they just located their digitizer in the antenna because they are a remote mount design instead of windshield mount, but it's not actually a different signal path design. I may be missing something about the Stinger though.
 

jon5

PSL + (O / U)
Advanced User
Joined
Oct 9, 2016
Messages
735
Reaction score
508
Isn't what we are doing more decentralized though? We have a separate antenna, separate digitizer, separate DSP processor (FPGA), and separate neural network SoM. Each of the components can be run independently with any other RF system as long as you know how to send spi commands. It seems to me they just located their digitizer in the antenna because they are a remote mount design instead of windshield mount, but it's not actually a different signal path design. I may be missing something about the Stinger though.
Yeah, i'm thinking strictly in remote-unit terms. For a dashmount, it makes much less sense, although having a distributed system would pave the way to sharing the core technology between dashmount and remote units. (Which IMO would be a nice bonus)

I don't know all the details about your project, i've just now heard about it. It does look like you're doing that with the modular boards there. In a remote system, wouldn't you put those modules on the peripherals instead of in the main CPU? That, with the peripherals connected via something like USB or Ethernet is what i was thinking of. Sounds like your doing that with SPI, which is great, but not as flexible. (Note that my knowledge this close to the hardware becomes very blurry, so i may be talking nonsense)

As for me, what i care about is being able to modify the software and not rely on the vendor. I think that was a lot of the excitement about this thread in the beginning, take a solid hardware and build an open RD/LJ software platform on top. There are a lot of clearly talented individuals here, as evidenced by some of these phone apps here, allowing that to flourish closer to the hardware should result in some very creative ideas.
 
Last edited:

Jon at Radenso

Manufacturer
Corgi Lovers
Advanced User
Premium Member
Manufacturer
Joined
Oct 9, 2012
Messages
2,376
Reaction score
18,145
Age
31
Yeah, i'm thinking strictly in remote-unit terms. For a dashmount, it makes much less sense, although having a distributed system would pave the way to sharing the core technology between dashmount and remote units. (Which IMO would be a nice bonus)

I don't know all the details about your project, i've just now heard about it. It does look like you're doing that with the modular boards there. In a remote system, wouldn't you put those modules on the peripherals instead of in the main CPU? That, with the peripherals connected via something like USB or Ethernet is what i was thinking of. Sounds like your doing that with SPI, which is great, but not as flexible. (Note that my knowledge this close to the hardware becomes very blurry, so i may be talking nonsense)
We are actually using USB as the internal bus for data communication, it's what that FTDI chip on the digitizer board is for. SPI is just what is used for tuning commands and such. For example, you can take apart the horn from the rest of the detector, provide power, and tune it with SPI even independent of the digitizer or other control unit.

As for me, what i care about is being able to modify the software and not rely on the vendor. I think that was a lot of the excitement about this thread in the beginning, take a solid hardware and build an open RD/LJ software platform on top. There are a lot of clearly talented individuals here, as evidenced by some of these phone apps here, allowing that to flourish closer to the hardware should result in some very creative ideas.
That is exactly why all of the code in our project will be open source. Our goal was to build a ridiculous hardware platform and then release it into the wild. Developers will not just have an open API, but direct access to samples, filter libraries, the FPGA, etc, etc.
 

jon5

PSL + (O / U)
Advanced User
Joined
Oct 9, 2016
Messages
735
Reaction score
508
We are actually using USB as the internal bus for data communication, it's what that FTDI chip on the digitizer board is for. SPI is just what is used for tuning commands and such. For example, you can take apart the horn from the rest of the detector, provide power, and tune it with SPI even independent of the digitizer or other control unit.



That is exactly why all of the code in our project will be open source. Our goal was to build a ridiculous hardware platform and then release it into the wild. Developers will not just have an open API, but direct access to samples, filter libraries, the FPGA, etc, etc.
Well count me as super excited! I really have missed a lot.
 

Jaguar

PSL +15
Advanced User
Premium Member
Joined
Nov 17, 2013
Messages
3,009
Reaction score
2,849
Location
Moderation (It's where I drink!)
Well count me as super excited! I really have missed a lot.
 

Discord Server

Latest threads

Forum statistics

Threads
86,472
Messages
1,315,645
Members
21,781
Latest member
Beemerchef
Top