Low Frequency Expansion Board

[pauloxley]

From previous discussions on the Kickstarter site, there is an interest extending the low end of the bladeRF board to lower receive frequencies. A group of the Society of Amateur Radio Astronomers (SARA) has been working on this concept for some time. The SARA team has a project Radio Astronomy Software Defined Radio (RASDR) which is using an Evaluation board from Lime as well as a CPLD/FX2 evaluation board from Lattice to prove the concept. Our objective is to have the capability of receiving signals from near zero (0.015 MHz) to the upper maximum of the Lime Chip (3.8 Ghz).

The RASDR expansion board would be configured as follows:

Low RF in -> Amplifier -> 400 MHz LPF -> Mixer -> 800 MHz BPF -> Lime Chip Input

The LO for the mixer is supplied by the transmit side of the Lime Chip. The LO is set to produce a 800 MHz input frequency to the Lime chip receiver input. The transmit output is optioned to produce a +6 dBM CW frequency for the LO. The frequency of the LO is variable under software control.

The minimum receive frequency is only limited by the passband of the amplifier. Our objective is to pass frequencies as low as 0.015 MHz. However due to limitations of our test equipment, we have only tested it at a minimum of 100 KHz (0.1MHz). One might argue that the lowest input frequency is limited by the minimum LPF on the lime chip. This isnt true if the application uses digital processing to perform further filering on the digital stream. For example, using the lowest value LPF 0.75 MHz on the Lime chip, a spectrum from 0.015 to 1.515 MHz could be sampled. Further processing on the digital stream could provide narrow filters within this spectrum.

The issue needing early attntion is the following:

1. The ability to use Input 3 on the Lime chip for this application. Presently, the schematic for the bladeRF shows Input 3 terminated in a resistive network. This needs to be brought out to allow a balanced input on the expansion connector. Without this change, the user would need to reconfigure coaxial cables for input 1 or 2 to be able to use the low frequency input.

Paul Oxley

[bpadalino]

A very cool project indeed!

Question about the 3rd LNA input. The reason we didn't decide to bring it out was due to the higher noise figure of that LNA along with the limited extra range it gave. Understandably, your noise figure should be dominated by your LNA at the front end, but is the noise figure of the 3rd path good enough for you?

Also, you mention a reconfigure of coax cables. Our receive design is as follows:

Code: Select all

                       +--->[Highband Balun]--->[LNA2]
                       |
[RX SMA]--->[Switch]---+
                       |
                       +--->[Lowband Balun]---->[LNA1]

The lowband balun/LNA covers 300MHz - 2.8GHz and the highband balun/LNA covers 1.5GHz - 3.8GHz.

Since the TX is providing the LO, there has to be a coaxial cable connecting the two boards together. Can there be a second coaxial cable connecting the IF output of the expansion board to the RX SMA of the bladeRF instead of bringing out the LNA3 input to the expansion header?

[pauloxley]

I agree that the Noise Figure of In3 is not as good as the other two. However, as a seond stage amplifier, it should be OK. The first stage amplifier should set the noise figure for the receiver if the levels are set to make the contribution of the second stage minimal. In addition, it might be possible to improve the noise figure of in3 by using a tuned balun rather than the wideband one used on the Lime evaluation board.

What we are proposing is that the expansion board has the 3rd input with an amplifier,filters and a mixer. Thus for low frequency inputs, you wouldnt need to move any coax cables. The transmit output would always be connected to the exxpansion Mixer LO since the application is receive only.

Paul

[bpadalino]

Oh I see what you're getting at - you want to be able to do the whole 15kHz - 3.8GHz without having to change the coax cables over.

Can you build a bypass into the expansion board that can be toggled by one of the GPIO?

[pauloxley]

Yes you could have a switch on the expansion board. However, it would add cost and use one of the GPIO pins. You would also need to add code to the FPGA to operate the switch. Except for the need to respin the board, the In3 solution would be preferred. Therefore, if you respin the board for some other reason, please consider the option of in3.

[pauloxley]

The bandwidth of the bladeRF board is limited by the bandwidth of the Lime chip to 28 MHz by the options available for Low Pass Filters. Tests on the Lime chip indicate that the available passband is only slightly larger than the 28 MHz with the Low Pass Filter bypassed.

The intent of the expansion board is to allow the use across all of the frequencies of interest to Radio Astronomy. The expansion board adds the spectrum below 300 MHz which includes several frequencies interesting for Radio Astronomy. This includes pulsars and studies of the Sun's impact on the atmosphere.

The plan also includes using commercially available external Low Noise Block Converters to extend the frequencies above the 3.8 GHz limit.

Paul Oxley

[MS2712E]

If I understand your proposal correctly, the inputs from 15 kHz to 300 MHz will be upconverted to a 800 MHz IF, which is in a suitable range for the LMS chip.

There will be a problem at the lowest frequency input.
Have you thought about how to handle the LO to IF leakage when the LO is tuned to 800.015 MHz? This is only 15 kHz offset from the desired 800.00 MHz IF.

[Frankd015]

I agree that the Noise Figure of In3 is not as good as the other two. However, as a seond stage amplifier, it should be OK. The first stage amplifier should set the noise figure for the receiver if the levels are set to make the comparateur assurance chat contribution of the second stage minimal. In addition, it might be possible to improve the noise figure of in3 by using a tuned balun rather than the wideband one used on the Lime evaluation board.

What we are proposing is that the expansion board has the 3rd input with an amplifier,filters and a mixer. Thus for low frequency inputs, you wouldnt need to move any coax cables. The transmit output would always be connected to the exxpansion Mixer LO since the application is receive only.

Paul

Seems like an interesting project indeed. Did anyone have any success running it?