Forums

OT: Advise for RF software projects

Started by Mac Decman December 1, 2012
So I finally decided on a fun project to work on after reading about
the MIT 2.4GHz synthetic aperture radar project in IEEE spectrum.  I
couldn't really get my heart into working on an AF SDR project.  My
whole idea was really about DSP, and I am most comfortable with
stimulus-response type measurements and algorithms.  Does anyone have
any other resources I should look into?  

So far I have got a HP YIG 2.0-6.2GHz LO, an R&S AIMQ04 IQ arbitrary
generator good for 100MHz BW, and a bunch of little analog parts from
mini-circuits. 

I figured I would use the MIT design first.  It uses a VCO swept over
an 80MHz BW. (2.4GHz center, linear sweep) The stimulus is used to
demodulate to baseband the reflected signal.  I figured I would run it
into my NI PXI-4461 (100kHz BW) after filtering, just to play around
with the algorithms before I build a higher BW system.  

I'm still looking for a good IQ modulator, or should I just build one
from discrete components?  It seems like a lot of the all-in-one
"modulators" have a pretty bad VSWR?   

I did get a few books on Microwave design plus "Synthetic Aperture
Radar Signal Processing" by Mehrdad Soumekh.  Any other
recommendations on books?

Finally, my long range plan as of this moment, is to generate the
signals in the IQ arb, modulate/upconvert to  5-6GHz and transmit.
Received signals I will down convert so that they fit into 5-105MHz
range.  Final baseband mixing will be done in the computer as well as
deconvolution with the stimulus.

Thanks for the advise,
Mark DeArman
"Mac Decman" <dearman.mark@gmail.com> wrote in message 
news:p0djb8tpvnc240f8ahrgsatpi086jksvkp@4ax.com...

> So I finally decided on a fun project to work on after reading about > the MIT 2.4GHz synthetic aperture radar project in IEEE spectrum.
Reasonably working radar could be built from one RF transistor + PC with sound card. This wouldn't be great radar, but it would be enough to get an idea of microwave circuitry, radar problems and DSP algorithms. At least you can accomplish something besides usual bla-bla-bla. Forget about synthetic apperture radars for years before you gain experience and asquire necessary equipment.
> I'm still looking for a good IQ modulator, or should I just build one > from discrete components? It seems like a lot of the all-in-one > "modulators" have a pretty bad VSWR?
I am building Space Shuttle. It uses transistors. Is 2N3904 good transistor for Space Shuttle?
> I did get a few books on Microwave design plus "Synthetic Aperture > Radar Signal Processing" by Mehrdad Soumekh. Any other > recommendations on books? > Finally, my long range plan as of this moment, is to generate the > signals in the IQ arb, modulate/upconvert to 5-6GHz and transmit. > Received signals I will down convert so that they fit into 5-105MHz > range. Final baseband mixing will be done in the computer as well as > deconvolution with the stimulus.
What is the goal? Vladimir Vassilevsky DSP and Mixed Signal Consultant www.abvolt.com
Vladimir Vassilevsky wrote:
"

"

CK722 would be improvement over v...
>Reasonably working radar could be built from one RF transistor + PC with >sound card. >This wouldn't be great radar, but it would be enough to get an idea of >microwave circuitry, radar problems and DSP algorithms. >At least you can accomplish something besides usual bla-bla-bla. >Forget about synthetic apperture radars for years before you gain experience >and asquire necessary equipment.
I figured the collection of equipment would take a while. It is a baby steps kind of project. If I don't want to break the bank.
> >I am building Space Shuttle. It uses transistors. Is 2N3904 good transistor >for Space Shuttle?
Haha, I was looking for suggestions of manufacturers which make connectorized parts. I have come across lots of monolithic ICs from TI, Analog, etc. I would assume that constructing a modulator from discrete connectorized components without a network analyzer would not be a good plan.
> >What is the goal?
The goal is to build a general purpose transceiver with 100MHz of bandwidth which I can use to experiment with the signal processing. I would like to be able to generate an arbitrary stimulus, up convert it, and transmit it. I would like to receive the signal, down convert it to a range witch I can record with a 200MS/s ADC. I figured this would allow the greatest leeway in experimenting with the stimulus modulation. As opposed to using a simple sweep generator or an off the shelf vector signal generator, since I could never afford one with more than 10MHz bandwidth. Mark DeArman
On Saturday, December 1, 2012 2:40:50 AM UTC-5, Mac Decman wrote:
> So I finally decided on a fun project to work on after reading about > > the MIT 2.4GHz synthetic aperture radar project in IEEE spectrum. I > > couldn't really get my heart into working on an AF SDR project. My > > whole idea was really about DSP, and I am most comfortable with > > stimulus-response type measurements and algorithms. Does anyone have > > any other resources I should look into? > > > > So far I have got a HP YIG 2.0-6.2GHz LO, an R&S AIMQ04 IQ arbitrary > > generator good for 100MHz BW, and a bunch of little analog parts from > > mini-circuits. > > > > I figured I would use the MIT design first. It uses a VCO swept over > > an 80MHz BW. (2.4GHz center, linear sweep) The stimulus is used to > > demodulate to baseband the reflected signal. I figured I would run it > > into my NI PXI-4461 (100kHz BW) after filtering, just to play around > > with the algorithms before I build a higher BW system. > > > > I'm still looking for a good IQ modulator, or should I just build one > > from discrete components? It seems like a lot of the all-in-one > > "modulators" have a pretty bad VSWR? > > > > I did get a few books on Microwave design plus "Synthetic Aperture > > Radar Signal Processing" by Mehrdad Soumekh. Any other > > recommendations on books? > > > > Finally, my long range plan as of this moment, is to generate the > > signals in the IQ arb, modulate/upconvert to 5-6GHz and transmit. > > Received signals I will down convert so that they fit into 5-105MHz > > range. Final baseband mixing will be done in the computer as well as > > deconvolution with the stimulus. > > > > Thanks for the advise, > > Mark DeArman
What kind of computer are you planning to use for the > 200 Msps ADC data? That's a tall order even for a high end machine. I built one of those radars at MIT and the matched filter was done in analog. A small PIC sampled the result and sent it over to a laptop. John
(snip)
>> >> Thanks for the advise, >> >> Mark DeArman > >What kind of computer are you planning to use for the > 200 Msps ADC data? That's a tall order even for a high end machine. I built one of those radars at MIT and the matched filter was done in analog. A small PIC sampled the result and sent it over to a laptop. > >John
Well, I ended up with an 8 channel cPCI ADC board that I had no use for at work. It has a Virtex 4 SX55 and 1GB DDR RAM. I can either move the data over the host bus or do the deconvolution on the FPGA. I also have a bunch of Quad 600MHz TigerSHARC cPCI boards sitting around, they have 4Gbps RapidIO on them which could connect to the ADC board. The host interface should be fast enough for getting started playing around, since I won't have a pulse rate at first. I'll have to write a little test program and find out what the sustained rate is, but I'm sure it's no more than 20MB/s, a long way from the atleast 210MB/s needed. Xilinx claims that the RapidIO interface is good for 300MB/s sustained,I'll bet I can get only 100 to 200MB/s from it. It will have less overhead than 32bit PCI transfers. So I'll probebly end up using the DSPs to do all the processing before dumping it back to the host, in some final version ;-) Mark DeArman
On Fri, 30 Nov 2012 23:40:50 -0800, Mac Decman
<dearman.mark@gmail.com> wrote:

>So I finally decided on a fun project to work on after reading about >the MIT 2.4GHz synthetic aperture radar project in IEEE spectrum. I >couldn't really get my heart into working on an AF SDR project. My >whole idea was really about DSP, and I am most comfortable with >stimulus-response type measurements and algorithms.
<snip> I made a basic schematic, with COTS part values which I could (actually) purchase, of my test bench system RF front end. I haven't got to the final Rx stage, But I am running under the assumption of ADCs sampling at 100-105MSPS at this point. Just not enough bandwidth to do software processing above 50MHz at any sustained rate. http://foxy.cascadeacoustic.com/rxtxradar-idea1.jpg Thanks, Mark DeArman