Large Spectrum Generation

[FR]

Dear All

I want to generate a Spectrum having bandwidth as large as  20Mhz . the hardware  I am using has realtime 56MHz Bandwidth . I am generaing a single tone (1Mhz) from DDS and then feed the signal to DAC and transmit the signal . Sweep is one of the option but because of time constraint , I want to explore other options .is it possible to use Frequency modulation or some other technique . yours suggestion will be highly appreciated .

 

BR

[xc6lx45]

Hi,

there are two important corner cases with regard to the power you're outputting:

- slow sweep (wobble / FM): the signal is continuous-wave, its peak-to-average is as low as physically possible. This will output the highest power within a given amplitude limit and (oversimplified) optimally use the dynamic range of the signal chain.

- single pulse: use a sample stream ... 0 0 0 1 0 0 0... (which has constant spectrum) or use the impulse response of a suitable bandpass filter. This results in the shortest possible signal within the limitations of the time-bandwidth product (see e.g. "band-limited pulse" on Wikipedia) but the signal has a high (bad) peak-to-average ratio, thus (oversimplified) requiring higher dynamic range in the signal chain.

Examples for the two methods are a conventional RF vector network analyzer (continuous wave, at least traditionally) and pulse-based time domain reflectometry, e.g. Teledyne LeCroy Sparq instruments.

What can work well is to use a bandpass-filtered (pseudo)random sample stream, possibly with additional clipping / filtering rounds to improve PAR. But at the end of the day, it depends on what statistics of the signal matter how much in your application, e.g. autocorrelation, amplitude PDF, actual spectrum shape.

Matlab (Octave) would be my tool of choice for algorithm design, not RTL

 

[hamster]

My first guess would be...

Take a pseudo-random stream of +1/-1 values, at about 1/4th your desired bandwidth. 

Lowpass filter that to half your desired bandwidth. 

Multiply that by your carrier samples

Send that to DAC.

Would be easy to do, but spectrum will be 'lumpy'.

 

[zygot]

8 hours ago, FR said:

I want to generate a Spectrum having bandwidth as large as  20Mhz .

It's probably harder to generate a signal constrained to 20 MHz. There's a whole discipline out there dedicated to EMI abatement. In the communications world where spectrum is tightly controlled constraining transmitters can comsume a significant part of the development process.  The "magic sauce" of RF design centers around frequency planning. I take it that you didn't like my reply to this same question in your post ADRV9361-z7035 SDR for generationg arbitrary waveform in a different section of the Diligent forum. I certainly endorse doing this in Octave but I've done it in VHDL (not the DAC of course) for commercial applications. I started with Octave simulation, then did VHDL simulation and confirmed performance on real hardware with calibrated RF test equipment. Certainly it's possible to do what you want to do. You just need the right structure and a lot of knowledge. Fortunately, knowledge can be obtained through experimentation, testing and analyzing.

[edit] In engineering we tend to want to confirm that what we did is correct. In science, I'd argue, the most important part is questioning concepts, tests, test measurement data, and the conclusions drawn from those things. In this regard engineers should try an be more like scientists.

[D@n]

@FR,

If you want really wide, throw out the hardware digitizer and use an OSERDES to create a 1-bit sample rate.  Most of the 7-series FPGAs are capable of outputting up to 800Mbps.  Treat this as your signal, and Voila!

Be aware, however, the devil is in the details.  The digital transmission lines on many boards aren't necessarily made for 100+MHz.  You may suffer from some serious distortion and/or attenuation.  It's probably good enough for some hobby type work, but certainly not anything I'd try to sell to a customer.

That said, I was able to use this approach to "broadcast" to a near-by FM radio using a XuLA2LX45.  The FM code is still posted.  It's actually pretty simple.  In that case, the output signal was generated at the system clock speed (80MHz), and if I ganged all of the spare I/O's together I was able to transmit for about 18 inches or so.  With a proper antenna, and some higher I/O speeds, I should have better luck.

Dan

[zygot]

1 hour ago, D@n said:

That said, I was able to use this approach to "broadcast" to a near-by FM radio

Nice! I've done that too on a few occasions. Unfortunately, none of them were intentional or were for projects having anything to do with "broadcasting". I assume that your transmitter, unlike mine, conformed to the right frequency mask and that yours didn't irk one of your neighbours.

My only reason for this comment is that while cheap SDRs are readily available that doesn't mean that people using them to transmit needn't have the training to use them responsibly, much the way Ham operators need a license. The same concept can be applied to drone operators.

I'm betting that a lot more experimenters using FPGA boards are doing this all the time, just not aware that they are. Poorly designed logic and or improperly terminated "digital" drivers is really all you need. I suppose that you get to have some bragging rights if most of your signal ends up close to where you designed it to be.

[zygot]

2 hours ago, D@n said:

the output signal was generated at the system clock speed (80MHz), and if I ganged all of the spare I/O's together I was able to transmit for about 18 inches or so.  With a proper antenna, and some higher I/O speeds, I should have better luck.

 I think that you overestimate what's required. In my cases ( the ones that I know of by chance ) I was able to drown out a stereo receiver connected to an outside antenna and there was no (intentional) ganging of drivers to to do so.

Now there's a good idea for your SDR, use it as a network analyzer for your digital projects.

[D@n]

@zygot

It wasn't just that I could get at least 18 inches ... I never managed to demonstrate the ability to go any farther.  I tried listening to the signal from the next room (my goal) but never managed to make it that far.

Did I stay within the band?  Probably not.  I was "broadcasting" on 90MHz from an 80 MHz system clock with a half inch board (i.e. an extremely disadvantaged "antenna".  Yours may have been better "matched")  Did I disrupt anyone else?  Not likely from a rural farmhouse.

A larger board, with a higher clock rate might be able to do a lot better.  Still not anything I would consider "professional" or even "sales worthy" for all the reasons you cite above.

Still, it's a fun story,

Dan

[zygot]

@D@n

I should mention that when I mentioned using the SDR as a network analyzer I wasn't directing that comment toward you... the person asking the question has the SDR. (But for very little money you could dive into the SDR world. There are a number of impressive one available these days). The contextual confusion is my fault.

There are a lot of ways to accomplish unintended things in unexpected ways. Two drivers trying to convince a load to see opposite digital levels can make quite a noise source. More so a bi-directional bus of them. I guess that the point of my last couple of posts is to be mindful and careful. I'm pretty comfortable with your abilities but you are an exception. Not all of us (sadly, that now includes me) live in an isolated rural farm house.

Anyway, while I encourage experimentation I also encourage owning the responsibility that is required to mess around with certain "toys". I know this guy, a friend of one of my cousins, a very bright lawyer, who really isn't mature enough to own a green laser pointer.... or perhaps most technology...

[xc6lx45]

21 hours ago, D@n said:

... and use an OSERDES to create a 1-bit sample rate.

(OT)

A curiosity, I've tried to use it for PWM once but the DC average observed on the brightness of a LED seemed to depend on the bit position. Has anybody made similar observations?

[zygot]

All of this is entertaining but it's all based on a pretty nebulous goal.

On 10/1/2019 at 1:27 AM, FR said:

generate a Spectrum having bandwidth as large as  20Mhz

Perhaps knowing what the purpose of the project is might help focus the discussion. Since an SDR has been mentioned I'm assuming that the bandwidth refers to the modulation bandwidth. To me this implies some bsndeidth limiting specification. Creating significant energy across a 20 MHz or greater band in the form of noise is one thing. Creating a signal meeting some minimal specification that you want to modulate onto a carrier is quite another. Is using an FPGA the best or even a good choice to accomplishing the project goal?

[edit] So I did take the time to look up the hardware that is the subject of the discussion, which is an AD9361 connected to a ZYNQ on a SOM. I'm not particularly interested in spending the time to see how flexible or modifiable the ADI supplied code support is on the PL level. There are other options for this transceiver if you want to play with HDL and SDR that might be more or less appropriate depending on whether you want to do MATLAB to signal or HDL to signal design. Sometimes unravelling the development framework a board provides is harder than just doing everything in an HDL. 

One thing for sure is that if you ask uninformed, nebulous or cryptic questions then you are likely to get a lot of uninformed and not particularly helpful answers... Oh, I know another thing too. If you choose a platform with limited vendor support it gets a lot harder to get answers to questions that the vendor assumes you are knowledgeable to figure out on your own...

When choosing a development platform you should ascertain whether or not the vendor supplied tools and documentation represent a walled-garden or an open experience. Does the platform come with a robust user community to interact with or are you forced to hope to run into someone else using your platform on a completely unrelated site like Digilent's?