hardware configurations
Justin Kelly
features would be desired in the first revision OSRP
Farhad Abdolian
I am a new comer on GR discussion and I am afraid I started a round of angry discussions around.
I have already started 3 projects sponsored by my own company to create the foundation for a open HW platform for GR.
The first project is for internal use, it is an adapter board for Xilinx FPGA boards to make it possible to use Ettus RF boards. This will be used as the basic ground for our future designs that we will make available under GNU license free of charge for non-commercial use.
The 3rd project is actually 2 different one and it has not started yet, we are thinking about 2 different version of the design, one low cost/low-end and one for more serious projects with a much larger FPGA and memory and faster ADC and DAC.
Here are the main highlights of the 2 units we have thought about:
- Altera Cyclone 4 FPGA in 484 or 672 ping FBGA package, supporting FPGA from $40 to $300
- 2 high speed DDR2 SDRAM memory path from 64megx16 to 512megx32 costing between $12 to $80
- 2 x 12 bits ADC and 2 x12 bits DACs at 80-250MHz or twice as many on the high-end version ($40 to $140)
- Support for 480Mbits/s USB2 (low end) and 5Gbit/s USB 3.0 (high-end) ($10 to $30)
- Support for 530MHz ARM9 based Friendly ARM expansion board (low-end)
or Dual 1GHz Cortex A8 based Panda board (high-end)
- TFT touch screen for diagnostic and control (high-end), Small TFT screen for diagnosting on the low-end version
- Possible support for PCIe (10Gbits/s) or eSATA (3Gbits/s) on the high-end unit
- 4 layer board for low-end and 8 later board for high-end version
And for the radio interface (not my strong point here):
- Both versions will have RX and TX similar to the simple RX and simple TX from Ettus plus a dual expansion socket on one of the channels
We can take care of the HW design and the FPGA firmware design as well as the PCB design, prototyping and testing. But we don't have a good knowledge of RF and Linux SW so that is something that can be done by the GR community to get this a real open source HW project.
Please let me know what you think of the idea and if you would be OK to participate in the pre-spec discussions before I send it to the whole group.
Thanks in advance and best regards,
Farhad Abdolian
PS. You can access my linkedin profile at: http://fr.linkedin.com/in/farhadabdolian
,
From: Justin Kelly <free...@gmail.com>
To: Open Software Radio Peripheral <open-software-r...@googlegroups.com>
Sent: Sun, January 23, 2011 9:20:32 PM
Subject: hardware configurations
Euripedes Rocha Filho
I'm very interested in work in this project. I'm a new college graduate engineer and I'have worked on put gnuradio and an old XtremeDSP board together. I can help with FPGA code and LINUX software. But my main goal is to provide a low cost plataform to the community. My personal goal is to achieve more experience in this kind of project. I have some experience in signal processing and FPGA development.
Farhad, I saw you are hiring to work on the change of the Cyclone FPGA, just don't remember where. May be on linkedin.
With regards.
Euripedes
Euripedes Rocha Filho
Euripedes
Farhad Abdolian
Yes, I got a freelancer working on the schematic changes of the USRP1 and porting it to Altium designer (our PCB designer prefer having everything in Altium). We are planning to have the first prototype by end of February.
This is really exciting, we want to be able to use this platform and bring it with us to the schools we are building in rural areas of Africa (see our fund raising page on Orphan Heroes page) to give low-cost mobile phone access to the locals.
Best regards,
Farhad
From: Euripedes Rocha Filho <rocha.e...@gmail.com>
To: open-software-r...@googlegroups.com
Sent: Sun, January 23, 2011 10:21:06 PM
Subject: Re: hardware configurations
Justin Kelly
I think you hit the nail on the head with your initial specs.
I am more inclined to use PCIe, E/SATA, and Firewire800, over USB.
Starting with the Ettus Research RF board foot print is a good start.
The only other thing I recommend is making the design modular, (Think Lego blocks) so the user can pick the interface type+FPGA+ADC/DAC combo+ARM+RF board they want.
The reason I bring up the modular topology is there are 24+ bit ADC/DACs on the market now.
Cheers
--Justin
Moeller
> Hello everyone,
> I am a new comer on GR discussion and I am afraid I started a round of angry discussions around.
Hello Farhad!
I like the discussions about alternative hardware. It's very productive.
> I have already started 3 projects sponsored by my own company to create the foundation for a open HW platform for GR.
> The first project is for internal use, it is an adapter board for Xilinx FPGA boards to make it possible to use Ettus RF boards. This will be used as
> the basic ground for our future designs that we will make available under GNU license free of charge for non-commercial use.
I'm not actively involved in FPGA development, but I used FPGA for glue logics before,
made some tutorials about VHDL and some tiny designs. My spare time is very limited,
but maybe I could do some small contributions.
Do you really want to restrict it to non-commercial use?
The GPL/LGPL license is open for every use, only that you have to publish the sources.
I think it would need a special open hardware license.
It is really productive to restrict it to non-commercial use?
Linux got really great contributions from commercial users and big companies (IBM, Intel, ...)
If it is used only by hobbyists, it's unlikely to get valuable contributions from professional users.
Of course the academic community could contribute, but there is also a large group of
commercial research institutions that would be excluded.
> The second project is based on the USRP1, replacing the Cyclone 1 with a modern Cyclone 4 FPGA and removing the second AD device and replacing it with
> a general purpose connector for future expansion. It also adds a TFT display to the board for diagnostic and debugging and possibly adding some memory
> to the FPGA. The supply voltage will be changed to 12V. It will also has a socket for high-accuracy crystal oscillator on the board. The goal is that
> this board can be produced for a price less than $150.
I would buy such an affordable alternative to the expensive USRP1.
But I would not need a TFT. All control and diagnosis is done with Gnuradio software from PC.
>
> The 3rd project is actually 2 different one and it has not started yet, we are thinking about 2 different version of the design, one low cost/low-end
> and one for more serious projects with a much larger FPGA and memory and faster ADC and DAC.
Fast ADC is interesting, also DSP-like FPGA (Virtex) for on-board signal processing.
However, it's not easy to develop signal processing for such dedicated architectures.
> - 2 high speed DDR2 SDRAM memory path from 64megx16 to 512megx32 costing between $12 to $80
Yes, SDRAM would be nice for taking signal snapshots (digital scope application)
> - 2 x 12 bits ADC and 2 x12 bits DACs at 80-250MHz or twice as many on the high-end version ($40 to $140)
> - Support for 480Mbits/s USB2 (low end) and 5Gbit/s USB 3.0 <http://www.everythingusb.com/superspeed-usb.html>(high-end) ($10 to $30)
I'm interested in a low-cost version. But for research and commercial users,
high-end would also be interesting.
> - Support for 530MHz ARM9 based Friendly ARM <http://www.friendlyarm.net/products/mini2440?lang=en> expansion board (low-end)
> or Dual 1GHz Cortex A8 based Panda board <http://pandaboard.org/> (high-end)
Not bad. A linux OS on the board would be great.
> - TFT touch screen for diagnostic and control (high-end), Small TFT screen for diagnosting on the low-end version
If the radio is controlled via PC, I don't see any benefit in a TFT.
Only if Gnuradio is running on the embedded CPU. There you have the problem
of lower performance of embedded CPU.
> - Both versions will have RX and TX similar to the simple RX and simple TX from Ettus plus a dual expansion socket on one of the channels
Best would be to use a compatible interface.
So, USRP daughterboards could be used, or USRP could use new boards from this project.
> We can take care of the HW design and the FPGA firmware design as well as the PCB design, prototyping and testing. But we don't have a good knowledge
> of RF and Linux SW so that is something that can be done by the GR community to get this a real open source HW project.
That's very helpful. PCB development is very difficult and risky for software people.
I suppose you will get much software support from the Gnuradio community, if there is
a cheaper hardware platform as USRP. Many software modules from Gnuradio could be reused.
Even FPGA algorithms (DDC etc.) are open source (GPL) and can be used.
> Please let me know what you think of the idea and if you would be OK to participate in the pre-spec discussions before I send it to the whole group.
I think it's a great idea. But it should be done step-by-step.
Complex designs with FPGA signal processing and CPU will be difficult to develop.
Large parts of the Gnuradio software would have to be adapted.
Most wanted in the Gnuradio community is an affordable alternative to the USRP 1/2.
This could be done with the least effort, reusing large parts of the Gnuradio sofware.
Moeller
> This is really exciting, we want to be able to use this platform and bring it with us to the schools we are building in rural areas of Africa (see our
> fund raising page on Orphan Heroes page <http://orphansheroes.com/Orphans_Heroes/Donate__Andokope_School.html>) to give low-cost mobile phone access
> to the locals.
My idea is to use a software defined radio for electronics education.
The kids have so many electronic consumer products.
Blinking LED and a mic amplifier experiments are nothing special to a child
that owns a PC already. But with SDR you can do really fascinating
experiments like a small RADAR, spectrum analyzer, watching electronic
emissions, listening to satellite signals ...
An affordable SDR is also interesting for development countries.
Imagine, in Africa a school using a spectrum analyzer for electronics
education. With the Gnuradio-like SDR this could be realized with <$500.
Professional devices from HP, Agilent or R&S would cost in the $10 000 -$100 000 range.
Of course it's a different class of device, but you can teach the principles of
spectrum analyzers with the simple SDR, too.
It's a matter of access to higher education and welfare for these countries.
I'm afraid with my little spare time I can't help very much with
the SDR PCB development.
But I would like to contribute experimental setups for electronics education,
Matlab/Octave scripts for computation, analysis and visualization.
I also think of some hardware extras, like a scope frontend (voltage protection,
attenuation). Or FPGA modifications for RADAR experiments
(needs synchronisation for TX/RX signals)
My vision is to provide the whole chain of open hardware, open software
and experimentation manuals (description of the setup, possibly some extra
hardware, special application and analysis programs, with Octave postprocessing
incl. nice Gnuplot graphs).
Justin Kelly
--Justin
Farhad Abdolian
What an interesting and lively discussion. It is really nice to real all these wonderful ideas!
Just a few notes from my side:
By free for non-commercial use, I mean the design can be used, duplicated and reproduced by anyone who is not going to use it in a commercial application. companies who want to use the design have to either purchase the boards or pay a fee to use the design. The reason for this is that we are going to spend 100s of hours of work and resources on this project, we would be more than happy to share it with universities, hobbyists and people who want to use this in their research or for learning. But if a company wants to use this in a product, then they should respect the work we put on this and pay for some of the charges. This will help us work on future designs as well.
I like the modular idea a lot, that is why the board will have minimum amount of 'blocks' and multiple connectors, here are the ones I had in mind:
- 2 USRP compatible connectors
- 1 IDE connector (one of the crazy ideas is to use the 200MBytes/s ability of the EIDE in some applications)
- 1 eSATA connector
- 1 USB3 connector
- 1 general purpose connectors (2x20 pin or similar tbd)
- LCD/TFT connector that can be used as a GPIO
- Possible PCIe on high-end
I am not in favor of FireWire 800, as much as I loved it when it came, but it is not fast enough to justify the efforts. eSATA and USB3 are definitely the way to go for a design that will be on the market in about a year.
The TFT modules I wrote about are very small units (1.5" and cost less than $3 in production) the larger units (2.5' or 4" QVGA) add an extra $20 to the price which is nothing for a $1300 board. But it can be let out and the connector can be used for something else.
I believe the biggest pottential is for the very low cost design and if we can make the board available for less than $150 then the usage of GR will spread into a much larger audience than it is today.
Best regards,
Farhad
Sent: Mon, January 24, 2011 2:20:10 AM
Subject: Re: hardware configurations
Euripedes Rocha Filho
In the hardware discussion, I want to start to work in the cheapest solution. But first we need to set the tools. In an open source project more people = more quality and features, so we must choose the tools in a way that people can colaborate easily. To me, this means: use always open source tools when you can do this. So I think we can choose between gEDA e kicad for schematics and board design. I like kicad but I don't know gEDA so I have a biased opinion. I want to start to design a small library with the symbols and packages we'll use to make easy to start developing, but I want to know your opinion.
Regarding the license I think that an open source license without a comercial use restriction is better. I understand Farhad concerns about people using his design without paying nothing but may be a company who will place new features on the design. What about a restriction for comercial use if you don't give something for the community? Anyone can use the design but must return something as some design effort or financial donation. I guess we must open a topic just for this discussion.
Euripedes
Justin Kelly
To Farhad,
It's funny I was googling for the ATAPI 6 specs lastnight (A.K.A. IDE/
EDIE)
ATAPI 6 is the 100Mb/s spec, it was the widely adopted one before the
introduction of SATA, it may provide an ultra cheap interface for
people who are willing to use a desktop machine, and don't mind
letting it all hang out...
When you say "2 USRP compatible connectors" do you mean 2 RF daughter
board expansion slots?
For the general purpose expansion connector I think you should combine
this with the LCD/TFT I/O pins, and just make it a single port.
After this gets going I think looking into doing some Die on Board
designs for making Mini PCIe cards is in order.
To: Euripedes
I think the licensing restrictions are fine the way they are for now,
but think a man hour barter system should be established later on.
To: Everyone
Have you received the notification of the RF Board discussion?
That's all for now.
N2TOH
> Sent: Mon, January 24, 2011 2:20:10 AM
> Subject: Re: hardware configurations
>
> another thought is to include an I2C or similar interface to control external
> frequency converters, (In Ham speak a transverter.)
>
> --Justin
>
>
> Hi Farhad,
>
>
>
> >I think you hit the nail on the head with your initial specs.
>
> >I am more inclined to use PCIe, E/SATA, and Firewire800, over USB.
>
> >Starting with the Ettus Research RF board foot print is a good start.
>
> >The only other thing I recommend is making the design modular, (Think Lego
> >blocks) so the user can pick the interface type+FPGA+ADC/DAC combo+ARM+RF board
> >they want.
>
> >The reason I bring up the modular topology is there are 24+ bit ADC/DACs on the
> >market now.
>
> >Cheers
>
> >--Justin
>
Farhad Abdolian
>ATAPI 6 is the 100Mb/s spec, it was the widely adopted one before the
>introduction of SATA, it may provide an ultra cheap interface for
>people who are willing to use a desktop machine, and don't mind
>letting it all hang out...
http://cgi.ebay.com/New-Mukii-TiP-Q120U3Si-SATA-IDE-USB-3-0-Adapter-/250711599289?pt=LH_DefaultDomain_0&hash=item3a5f9368b9
Best regards,
/Farhad
Justin Kelly
It's Justin not Justine
If something like the device you linked to on Ebay will do the job,
and make developing a first Generation OSPR easier then I say go for
it.
What type of connector were you thinking of for the OSRP board? I was
thinking a male 40 pin PCB connector, and just hang the board off of a
PCI expansion card to start with.
I have a few cards like this one,
http://cgi.ebay.com/VIA-4-Port-SATA-and-IDE-Expansion-PCI-Card-Box-CHWPEC02-/120669617732?pt=LH_DefaultDomain_0&hash=item1c18783e44
Justin
On Jan 24, 11:39 am, Farhad Abdolian <f.abdol...@yahoo.com> wrote:
> Hi Justine,
>
> >ATAPI 6 is the 100Mb/s spec, it was the widely adopted one before the
> >introduction of SATA, it may provide an ultra cheap interface for
> >people who are willing to use a desktop machine, and don't mind
> >letting it all hang out...
>
> Actually, for those with modern laptops, IDE interface can be good too if they
Moeller
> Regarding the license I think that an open source license without a comercial use restriction is better. I understand Farhad concerns about people
> using his design without paying nothing but may be a company who will place new features on the design. What about a restriction for comercial use if
> you don't give something for the community? Anyone can use the design but must return something as some design effort or financial donation. I guess
> we must open a topic just for this discussion.
I also prefer that the developers benefit from commercial users.
But it will be very difficult to organize. How should the income be shared
between developers? And how to define non-commercial use? Is it commercial if
the radio is used in paid research projects? What about research institutes owned
partially by university, partially by a company (outsourced research groups),
what about military users, what about somebody selling his hobby equipment later ...
I think you will loose a significant large potential user base with that
restriction.
Commercial users just wanting to buy a device will choose the USRP more likely.
Users that want to modify a design could use it, but with a free and open
GPL or LGPL-like license they would have to put the modifications under
a free license again, so contributing back to the project.
If you want to charge for the design, the contributing companies would also have
a right to charge for the modifications. All in all, very complicated.
Justin Kelly
another discussion by sending an email to open-software-radio-
perip...@googlegroups.com with the subject as [Discuss-OSRP] "insert
you new topic here"
Justin N2TOH
Justin Kelly
An arbitrary request that may seem unnecessary at this point, but I'll
make it anyway.
In reguards to Printed Circuit board outlines, I'm requesting that
they try to conform to already established standards. Such as size
restrictions for PCI, and PCIe boards, and within the confines of
established dimensions for the the 2.5", 3.5" and 5.25" disk drives.
By keeping these suggestions in mind it will be easier going forward
it integrate whatever we come up with to scale into bigger systems.
I have already committed to a drawing of the VIA chipset ATAPI 6 card
I mentioned, and will see if I can scare up a set of standardized
drawings for the other drive types.
Thank you and good day,
Justin N2TOH
Euripedes Rocha Filho
Euripedes
--
You received this message because you are subscribed to the Google Groups "Open Software Radio Peripheral" group.
To post to this group, send email to open-software-r...@googlegroups.com.
To unsubscribe from this group, send email to open-software-radio-p...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/open-software-radio-peripheral?hl=en.
Farhad Abdolian
Sorry but I am very busy these days and can't participate much in discussions. Just a few notes and corrections
Justin, Sorry for the typo, I hate when people do that to me, I can't believe I did that myself. I used to work with a guy in New York who keeps using "Farhead" as my name :|
The main reason I suggested the PATA is:
1 - It is pretty easy to implement
2 - For those who don't want to use it, can use the simple 2x20 ping 0.1" connector for other purpose than ATA
3 - Makes it possible for those who want extremely low cost version
Second, USB 3 is not USB 2, it is much more complicated protocol and at the moment there are not USB3 device ICs available on the market, the only ones available are USB3 to PCIe converters but they are all for Host controllers and not useful for us. Using a USB3 <-> PATA or even SATA device makes it possible to get the board to the market faster and with minimum amount of efforts (USB3 IP costs more than $70k not including the royaltee).
There are companies using eSATA for high speed communication, GateRoket is one of those and they are very successful in their designs.
As I mentioned before, in my company, we will develop some products for internal use and then will release them for general public. For our internal use, we need something very low cost to support multiple GSM frequencies and possibility to have the whole OpenBTS/Gnu Radio SW and HW on one module (a combination of Panda board and our own adapter board with Ettus WBX or something similar). We want to use these solutions for our schools in Africa and other parts of the world and don't mind sharing it with hobbyists and private users and students but I don't think we will give up the rights for the designs completely to 'for profit' companies.
One important clarification here is that we will put a note in all our designs that the designs made by our company can not under any circumstances be used for or by any military person/group/organization/company or be part of any product used by the military of ANY country for ANY application or be used by individuals using these products for research of products for any kind of military usage. Almost all our team members have a very strong support for this.
Best regards,
Farhad
Subject: [Discuss-OSRP:25] Re: [Discuss-OSRP] Re: Discuss-OSRP Re: hardware configurations
Euripedes Rocha Filho
But I also want to develop an open source board without any restrictions that can be used even comercially. I will start the plans with some hints from our previous discussions using kicad to draw boards and schematics.
Best regards
Euripedes
Justin Kelly
Farhad, NP we all make mistakes.
I'm to the point now I need to focus more time roughing out a simple prototype, and getting stuff done.
I like to treat my Life experiences like a Chinese Buffet, pick what I like and pass on the other stuff. On that note I recall something from my childhood, something to the tune of do into other as....
In reference to Farhad's quote "One important clarification here is that we will put a note in all our designs that the designs made by our company can not under any circumstances be used for or by any military person/group/organization/company or be part of any product used by the military of ANY country for ANY application or be used by individuals using these products for research of products for any kind of military usage. Almost all our team members have a very strong support for this."
I would like to take it one step further and say under no circumstances should any of the OSRP IP be used to harm sentient lifeforms.
Justin N2TOH
Justin Kelly
on that note I'm working towards a mostly unrestricted design, the two current limitations are NO Military Applications EVER!, and must cause no harm.
Justin N2TOH
Farhad Abdolian
Best regards,
Farhad
PS. Love the NO Military Applications EVER!, but "must cause no harm" can be tricky specially if you thing of radiation issues :).
Sent: Tue, January 25, 2011 1:52:13 PM
Subject: Re: [Discuss-OSRP:30] Re: hardware configurations
Justin Kelly
Justin
Farhad Abdolian
Just got the first schematic of the new board with Altera Cyclone IV done. We are still working on the added features and would like to hear what you think of it. As promised, the prototypes are planned for mid march 2011.
All design files will be available on our web site (http://www.seemaconsulting.com) later this year
Best regards,
Farhad
Justin Kelly
signal?
I would also like to encourage you to look and the Anderson Power
poles for the DC power connector.
Justin
> 2166KViewDownload
Farhad Abdolian
There will be 2 oscillators on the board, one similar to USRP1 and one socket for your own. I am not sure about the Anderson power poles, are these the types you are talking about?
http://home.comcast.net/~nt9e/TES4410GBP.JPG
The connector will be a simple 12V power connector (http://i00.i.aliimg.com/photo/v0/301629764/SMT_DC_JACK_DC_POWER_socket_DC.jpg)and possibly a PC power supply connector like this http://home.arcor.de/r69/computer/trex/gfx/pwrcon.gif (I am not sure if it is necessary since we don't need that much power on this board)
Best regards,
Farhad
To: Open Software Radio Peripheral <open-software-r...@googlegroups.com>
Subject: [Discuss-OSRP:37] Re: First schematic of the new board done
You received this message because you are subscribed to the Google Groups "Open Software Radio Peripheral" group.
To post to this group, send email to open-software-r...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/open-software-radio-peripheral?hl=en.
Euripedes Rocha Filho
Euripedes
To unsubscribe from this group, send email to open-software-radio-p...@googlegroups.com.
Farhad Abdolian
The main reason for Cyclone IV is the fact that I dislike Xilinx after working with them for 20 years. Their tools are very unreliable and was the cause of many failed projects for us in the past. Altera has a much more robust set of tools, the only problem is that their prices are usually much higher than Xilinx which makes it more difficult to includethem in low cost products. The device we chose is about $15 in 100s which is a good price. It is also possible to use a larger FPGA on the same footprint if it is necessary. The price will then go up to about $70.
We are in discussion about adding a PCIe connector or ATA-7 compatible connector to the board, we are still not sure about the choice since we can not have both of them with the chosen FPGA.
Best regards,
Farhad
Subject: Re: [Discuss-OSRP:39] Re: First schematic of the new board done
Euripedes Rocha Filho
Hi,
The main reason for Cyclone IV is the fact that I dislike Xilinx after working with them for 20 years. Their tools are very unreliable and was the cause of many failed projects for us in the past. Altera has a much more robust set of tools, the only problem is that their prices are usually much higher than Xilinx which makes it more difficult to includethem in low cost products. The device we chose is about $15 in 100s which is a good price. It is also possible to use a larger FPGA on the same footprint if it is necessary. The price will then go up to about $70.
I don't like Xilinx tools also. About the tools, in my design only devices avaiable by free tools will be considered.
We are in discussion about adding a PCIe connector or ATA-7 compatible connector to the board, we are still not sure about the choice since we can not have both of them with the chosen FPGA.
If your plans include a PCIe, why not go for the Cyclone IV GX family that have a hard PCIe core embedded?
Justin Kelly
I posted a link with better prices an more options here, http://groups.google.com/group/open-software-radio-peripheral/browse_thread/thread/434a7a76967882c7
The company I have been buying them from is here, http://www.connex-electronics.com/
Development boards, The only reason I'm looking at Xilinx is on paper this board looks good and is cheap. http://www.xilinx.com/products/devkits/EK-S6-SP601-G.htm
Justin
To unsubscribe from this group, send email to open-software-radio-p...@googlegroups.com.
Justin Kelly
Justin