I am looking into persona devices as described in Appendix G of the Redhawk manual.
Is there a detailed "how to" for this anywhere?
In my scenario my 'Programmable Device' would be a Redhawk FEI device that interfaces with a kernel API that controls tuners, fans, gps, buttons and LCD displays. I would like to break this out into three or four persona devices that interface with the main FEI Device.
Thought I'd ask.
If you head to Geon's github and look at the RFNoC_ProgrammableDevice and RFNoC_DefaultPersona, you can get an idea of how these Devices interact with one another. It should be noted that these Devices are still under development. Unfortunately, the manual appendix you mentioned and these examples are really the closest thing to a "how to" there is right now.
That being said, this pattern is generally reserved for FPGAs, with the programmable Device controlling access to the programmable hardware (and FEI functionality, if present) and the persona(s) controlling access to specific bit file capabilities. If you're not interacting with an FPGA, then the pattern will most likely be more trouble than it's worth to obtain modularity.
Related
Good day!
Problem definition:
Current implementations of Bluetooth does not allow to simply support good quality of Audio(Earphones mode) and 2-way audio transition (Headset mode).
Also, even if one would manage to set this configuration up, which have huge limitations on the hardware/software used, there is no way to handle sound input from 2 different audio devices simultaneously.
So, technically - one cannot just play the Game, communicate on the Discord, and optionally listen to some music, unless he is bound to some USB-bundled earphones. Which are usually really crappy, or really expensive. Or both.
Solution sketch:
So, I came up with an idea that one can actually build such device, using Raspberry Pi, Arduino, or even barebone-component-based stacks.
Theoretical layout of connections per-se would look somehow like that:
Idea is to create 2 "simple" devices
One, not-so-portable, that would handle several analog inputs, and one analog output
One, portable, that would handle single analog Input and Output, and could be used with any analog earphones.
"Requirements" to such system would be quite simple:
This bundle have to handle Data Transition on some distance, preferably up to 10 meters, or more.
The "Inlet" device should be portable enough to keep it in the pocket, or in an arm band, or something
Sound Quality should be at the very least on the level of Bluetooth headphones profile, or if possible - even better
If possible - it would be nice to keep the price of the Solution under 500 Euros, but I'm so tired of current state of things that I might consider raising the budget...
Don't mind the yellow buttons on the Outlet device. Those are optional, and will depend on the implementation stack :)
Question:
Can anyone advice me which component-base would be a better solution to making such a tool, and why?
And maybe someone actually knows of similar systems already existing?
Personally I would prefer anything but the barebone-components-based solution, just because I'm really rusty with that area, and it requires quite the amount of tools, to handle it properly.
While using pre-built modules can save me from buying most of the hardware tools, minifying my "hardware customization" part of this solution, leaving only software part to handle (which is my main area of expertise).
But then again, if there are some experts here, that would consider other stacks non-viable - I would really appreciate to see their reasonings.
P.S. Just to be clear: If this project will prove viable - I will implement it, and share the implementation details with the communities. I am not the first one who needs such system, and unfortunately it seems that Hardware/Software vendors are not really interested in designing similar solutions...
I happen to find a "temporary" solution.
I've came across a wireless headset, that allows to simultaneously support Wireless USB Bundle connection, and Bluetooth connection to different devices, and provide nice way of controlling sound input/output with both connections.
This was almost a pure luck, as this "feature" was not described anywhere in the specs...
Actual headset name is:
JBL Quantum 800
This does not closes the question per-se, as I still plan to implement this "Summer Project" at some point, but I believe this information might be useful to those searching for similar solutions.
In terms of working on microcontrollers and microprocessors, I have been said to work with SPI Interface instead of using USB. When I was going to deeper on my hardware selection, I have noticed that there are many other Interface like MIPI Dsi, Csi or so on. So what are differences? Can I choose something with mipi dsi interface and use it on my system which will be a sensor system?
Your question is rather broad for Stack Overflow and has a really wide scope; but fundamentally, all the interfaces you've lists, SPI, USB, MIPI DSI, MIPI CSI, etc, are all just communication interfaces, or ways for external components like sensors, cameras, displays, input devices, storage units, etc to talk to a processor; each usually being designed with specific goals in mind. USB for example was designed to be generic and used to connect peripherals to desktop/laptop consumer systems from keyboads and mice to webcams to other devices; while MIPI DSI was specifically designed to interface mobile/embedded displays to the host processor. Due to these design goals for each interface, they all have quite different physical and link-layer implementations (ie, the actual number of wires and the way the signals on those wires are sent and received).
When looking to which interface is right for your system, you need to look at what the processor supports and what interfaces the device you wish to hook up to it supports. If say, you have a simple accelerometer and a simple microprocessors, its likely both with use either a SPI or I2C interface. If its a larger processor and a camera, then either USB or MIPI CSI might be good options. Also recognize that depending on software support, writing code to interface with a USB sensor can be quite a bit more complex than one for a SPI interface (thus the widespread use of SPI and I2C in embedded systems). Without more details, I cant say whats a good way to hook things up (and that type of question might better fit on electronics stackexchange: https://electronics.stackexchange.com/).
I need to do some NON-STANDARD signal processing operations with an RFID-reader, so I'd like to know if it is possible to extract antenna's individual analog (actually digital samples right after ADC) input signal samples with Motorola FX7500 (if you know how this works on FX7400 or FX9500, please do tell, could be helpful). Samples would be processed in a JAVA-based host computer program.
What I've already tried:
Investigating Motorola's own RFID3 API's possibilities, it doesn't go deep enough to actually get in touch with input analog signal samples.
Using LLRP to its full extent, it doesn't allow analog signal sample access either. RFsurvey-functionality would have been helpful to some extent, but FX7500 doesn't support it either.
Accessing RFID-reader's linux terminal, trying to find the driver function(s), that could listen the input sample stream. If current input sample(s) could be extracted from the input stream, I could (in theory) make a script, that would save a few of those sample values in a txt-file in the host computer during a tag inventory round. My linux skills are kinda bad, hence I ask this question.
The only realistic way to solution seems to be via linux terminal, so if you folks have any ideas about that (where to look and what to do), please advise!
Contents of reader:
rfidadm#FX7500abcdef:/$ ls -1
apps
bin
dev
etc
home
include
lib
linuxrc
media
mnt
platform
proc
readerconfig
run
sbin
sys
tmp
usr
var
I cannot completely rule that out, but it's highly unlikely you can get the raw signal digitized; the devices you're looking at aren't really software defined radio devices, typically.
"speaking" RFID physically is a bit different from "usual" wireless communication: The reader doesn't only observe the energy transmitted from the tag, but more importantly the fluctuations of energy extracted from the near field of the reader's antenna coil. Hence, you don't actually have a baseband of RF bandpass signal, but hardware-specific modulations of transmitted (and inversely, antenna-reflected) energy. Demodulation is hence usually done in specialized hardware.
However, do not fret: It's totally possible to build a software defined RFID reader. There have been several approaches to that, but personally, I trust these based on Ettus USRPs and/or GNU Radio best. Look through the results IEEExplore gives you, eg. this search.
Most probably this is not possible with the Motorola readers. What you can do, is use one of the RFID chipsets available on the market: either the AMS RFID IC's, or the Impinj RFID IC's. As far as I know, both IC's support retrieving the digital samples that are received. They also have a development kit to test-drive the IC's.
I was wondering if it is possible to modify the contents of a BLE beacon to include extra information. If you insert an extra bit at the end you could potentially broadcast a boolean in one direction. Theoretically, if you modified your device to read the extra bit of information this would work. Given existing protocols though it sounds like this would be a lot of work. Is there something out there like this already?
For info, I'm working on the mbed platform where you can modify your own bluetooth beacon payload.
Yes, you can do this with the new AltBeacon specification. There is a one byte manufacturer reserved field which you can use for whatever you want (tied to your manufacturer ID).
There are reference implementations of the specification available for Linux to show you how it works, and there is no reason you cannot implement it on the mbed platform.
We are considering using the vendor and product ID of a USB device (obtained via IOKit) to unlock certain features of an application. I'm aware that these values can be spoofed, but I'm not sure how easily it can be done on OSX. What is involved in spoofing the vendor and product ID? Is it something that a non-programmer can do fairly easily?
I suspect it would be fairly easy (for a programmer) to create a kext that creates a reasonably realistic IOUSBDevice object in kernel space. However, once done anyone would be able to load it.
It may also be possible to build a codeless kext using AppleUSBMergeNub to masquerade a device as another (though I haven't tested this, it looks like it might work, assuming you only try to match the IOUSBDevice without actually using it).
It would probably be easier to just hijack your API calls in userspace and give your software the answers it expects.
A non-programmer cannot do that. What you can do is writing a kernel extension (an IOKit driver) for the device, that has a matching dictionary matching the real vendor-/product-ID of the device, which causes the system to automatically load the kernel extension when this device is connected and pass a reference to the device object. The driver is now responsible for initializing the device and create all necessary user space information that is necessary for the IOKit framework (the user space library) to communicate with the device. Usually Apple has default IOKit drivers for most USB device classes (that's why you don't need an extra driver for every USB mouse or keyboard for example), but if there is a more specific driver found, then this driver will be used instead. And when creating the user space data, of course the driver may lie about the vendor-/product-ID, causing the user space program to see false values. Since IOKit is written in C++ and heavily depends upon objects and object inheritance, it is not even necessary to write the driver from scratch, but instead it can inherit from Apple's default driver and just override some methods and otherwise rely upon the super implementation. However, writing IOKit kernel drivers is not that easy and even people developing Mac software for years may have no idea how to do that and the IOKit documentation is one of the worst ones Apple is offering.
So to answer your question: Is it possible? Yes. Is it fairly easy? No. Can non-programmers do it? Certainly not, unless someone else has written a ready to use extension for them, that they only need to install.
It's actually rather EASY, and I'm definitely not a programmer!
Found the guide here
http://rolande.wordpress.com/2012/10/25/getting-the-aten-usb-serial-adapter-working-with-mac-osx-lion
Thanks, Rolande!
I was able to change the IDs of a USB Ethernet adaptor I bought on eBay to use another driver with Mac OS X 10.8.
Since nobody has answered so far, I've stumbled across a page which basically says that you can get an Arduino and program it to return any VID/PID/serial that you want. I still can't find any software-only way to do this.
I can create A OCX in Delphi code to do that
And supply infomation to send/receive data to the USB with other states like:- is it still connected.
Natraly I need a little more detail than that in case that same chip is connected from another product.
example frendly name, device desription.............