Featured Articles
Announcing Pi
Technical Specifications
Raspberry Pi FAQs
Win a Pi
Raspberry Pi Wishlist
Actions
Forum Thread Details
  • Replies 4 replies
  • Subscribers 685 subscribers
  • Views 440 views
  • Users 0 members are here
Related

Trying the Pi Pico RX SDR Project

scottiebabe

The element14 community (shabaz) pointed me to a very neat SDR project using a RPI Pico.

I have a RPI Pico and lots of generic and jellybean electronic components.

So, I will have a go at building a copy of the open source project.

All of the projects schematics and documentation are available here:

https://101-things.readthedocs.io/en/latest/radio_receiver.html

I don’t have all the exact components used in the project.

I made the following substitutions with similar components I had on hand

image

The frontend low rdson analog multiplexer was replaced with a 74HC066 quad analog switch (plus a one hot decoder).

I also lowered the gain of the op-amp difference amplifiers.

On power up, the picorx software defaults to an LO frequency of 1.4085 MHz.

I used my signal generator to apply a 150 mVpp carrier with a frequency that sweeps from 10 kHz below the pico’s LO to 10 kHz above the pico’s LO.

I get wonderful sinewaves on the output I/Q channels:

image

The quadrature detector has significantly more bandwidth than I expected.

You will notice that for frequencies above the local oscillator the Q signal leads the I signal (and vice versa for frequencies below the LO).

The QSD’s response within 1 kHz of the LO is being partially rejected by the AC-coupled difference amplifiers.

The amplitude response of the I/Q channels is not very well matched (I will have to think about that for a bit)

More to experiment with, but impressive results for the simplicity of the project.

  • Here is a spot test of the image rejection performance at 1.4 MHz

    First I apply an AM RF signal 6 kHz above the LO

    image

    Signal Generator:

    image

    The recorded I,Q samples from the o-scope

    image

    The magnitude of the complex signal I + jQ

    image

    The FFT of the I,Q samples

    image
    There is a DC offset in the I,Q samples so some power will be seen at 0 Hz. An image of the + 6 kHz signal exists but is attenuated by approximately 30 dB. On a linear scale the image is less noticeable
    image

    Adding a unmodulated carrier 8 kHz below the QSD’s LO

    image

    The measured I,Q samples

    image
    image

    In the frequency domain the -8kHz tone (noisy, relaxation oscillator, not crystal controlled) has an image at +8 kHz, but attenuated by 30 dB

    image
  • in reply to scottiebabe

    Nice results! There may be noticeable difference with the various substitutions, because of the higher switch resistance (and I don't know how matched they are, although they should be near-identical since they are all on the same chip), plus the load on the sampling capacitors is different for the different inputs of the op-amp (but then again, it was different for the original circuit too). I guess it would be nicer to also have a quad op-amp serving as a same-load buffer for each sampling capacitor, but maybe the effect is negligible (I've not seen any circuits have any buffers either admittedly).

    Regarding the switches, In the past I tried a SD5400CY, perhaps the 3253 is better (it's seen a lot in online circuits, whereas the SD5400 is rarely seen) simply because I thought there was less inside to be mismatched or cause any other switching noise from all the additional MOSFETs in the silicon in the 3253), but I've never compared. SD5400CY is harder to find and too pricey now (in the past it was very cheap) so those are negatives too.

  • Spent some time putting the UI bits on a front panel. Most of the configuration options are only available via the physical UI, not the usb serial interface.

    image

    The author did a nice job on the UI, it looks sharp.

    image