Pi-Pico RX - Breadboard Version

A couple of years ago, I built a basic yet capable radio receiver using a Pi Pico, and while I originally designed a custom PCB for it, this time I’m building an even simpler and cheaper version that can be built on a breadboard using (mostly) through-hole components.

I wanted to build a very minimal (but useful) design that I could use as a platform for experiments, tweaks and upgrades.

If you are interested in the original design or want to find out more about the technical details you can find all the info here. Most of the technical details are the same, so I will focus on the new features and improvements here.

Quick Links

You can find the code on the GitHub page including a precompiled firmware in .uf2 format.

What Can This Receiver Do?

The receiver covers frequencies up to 30MHz, including commercial broadcasts on Longwave, Medium Wave, Shortwave, and the HF amateur radio bands. What’s great about this design is that it’s completely standalone—it doesn’t need a PC or sound card and can run for hours on just three AAA batteries.

Here are the specs:

• 0 - 30MHz coverage

• CW/SSB/AM/FM reception

• OLED display

• simple spectrum scope

• headphone output

• 500 general-purpose memories

• runs on 3 AAA batteries

• less than 50mA current consumption

Parts List

Item	Value	QTY	Description	Example
R1, R2	10k Ohm	2 off	Resistor (Metal Film)
R3	1k Ohm	1 off	Resistor (Metal Film)
R4	100 Ohm	1 off	Resistor (Metal Film)
R5, R6, R7, R8	82 Ohm	4 off	Resistor (Metal Film)
R9, R10	56k Ohm	2 off	Resistor (Metal Film)
C1, C3, C5	100n Farad	3 off	Capacitor (Ceramic)
C7	10n Farad	1 off	Capacitor (Ceramic)
C9	100u Farad	1 off	Capacitor (Electrolytic)
C10, C11, C12, C13	56n Farad	4 off	Capacitor (Ceramic)
C14, C15, C8	10u Farad	3 off	Capacitor (Ceramic)
C16	470n Farad	1 off	Capacitor (Ceramic)
C17, C18	220p Farad	1 off	Capacitor (Ceramic)
IC1		1 off	MCP6022 Dual Operational Amplifier	https://tinyurl.com/3zb279aa
IC2		1 off	Raspberry Pi Pico (or pico 2)	https://tinyurl.com/mpjc2pzn
IC3		1 off	74CBTLV3253 4:2 Analog Multiplexer	https://tinyurl.com/kuvmardh
SW1		1 off	Rotary Encoder with push switch	https://tinyurl.com/sepnzwwb
SW2, SW3		2 off	Push Button Switch (Momentary)	https://tinyurl.com/d7s957vz
L1/FB1	100u Henry	1 off	Inuductor/Ferrite Bead	https://tinyurl.com/y94cnnr6
Headphone Connector		1 off	3.5mm Stereo Jack	https://tinyurl.com/59cfmcxz
Display		1 off	128x64 I2C OLED Display 0.96 inch SSD1306	https://tinyurl.com/4w7f4p9a
Antenna Connector		1 off	BNC or SMA female connector	https://tinyurl.com/yrca5dsk

The Design Walkthrough

At the heart of this receiver is a Tayloe detector, which is popular for its simplicity and performance. The detector converts high-frequency RF signals into lower-frequency IQ signals that the Pi Pico’s Analog-to-Digital Converter (ADC) can sample. This design handles frequencies up to 30MHz, although the ADC has a bandwidth of only 250kHz.

The local oscillator, necessary for the detector to function, is generated directly by the Pi Pico using its PIO feature. This oscillator drives a 4-way analogue switch, which samples the incoming signal in four different paths, each covering a quarter of the local oscillator cycle.

The I and Q signals generated by the op-amps contain all the information needed to demodulate the signals. With both I and Q signals, we can determine the amplitude, phase, and frequency, including whether the signal frequency is higher or lower than the local oscillator.

The audio interface in this design uses a simple PWM method. Although it’s basic, it performs surprisingly well. The RC low-pass filter removes the PWM ripple and the output is strong enough to drive headphones directly or even a small speaker, though an external amplifier is recommended for better performance.

Improvements and Tweaks

Since building the original design, I’ve received a lot of feedback and made several key improvements to enhance both performance and usability.

1. Capacitors to Prevent Op-Amp Saturation

One of the issues that has been addressed was the saturation of the op-amps at higher frequencies. One or both op-amps in the Tayloe detector would saturate, leading to poor rejection of aliased signals. I quite a few changes to remedy the solution, but in the end the solution came from the truSDX transceiver. The addition of these two capacitors solved all the issues of op-amp saturation on the high bands and I am getting much better performance.

2. Improved Frequency NCO resolution by changing the system clock frequency on the fly.

Another major improvement involves the frequency accuracy of the Numerically Controlled Oscillator (NCO). In the original design, I used the fractional dividers in the Pi Pico’s PIO peripheral to generate a local oscillator frequency close to the desired frequency. However, this method only allowed me to get within about 100kHz of the target frequency. While this was sufficient given the 250kHz bandwidth of the ADC, I was able to get even better resolution by making small changes to the system clock frequency.

The firmware originally ran at 125MHz, but by tweaking the PLL, the system clock frequency can be varied between 125MHz and 133MHz. 133MHz is the maximum frequency without overclocking. There are 23 possible system clock frequencies in this range. Choosing the best combination of system clock and PIO divider gives a resolution of ~ +/-8kHz of the desired frequency.

This improvement also allows for a narrower bandwidth of about +/-12kHz, now we can oversample the I and Q signals by a factor of 10, which greatly improves the rejection of alias signals which was a weakness in the original design.

Note: Although the PIO-based local oscillator has a resolution of +/-8kHz (previously ~+/-60kHz). The receiver achieves an overall frequency resolution much better than 1Hz. The software implements a second, very-high-resolution, NCO and mixer in the front end of the receiver to do the fine-tuning.

3. Switching to a More Affordable Op-Amp

With the improved frequency accuracy and reduced bandwidth, it is now possible to switch to a more affordable operational amplifier. The LT6231 op-amp, is a popular choice in Tayloe detector designs due to its exceptionally low noise performance. The original design required a gain-bandwidth product (GBP) of 60MHz, well within the 215MHz limit of the LT6231. However, the LT6231 is relatively expensive, costing about twice as much as the Pi Pico.

Thanks to the improvements in frequency accuracy and the ability to oversample, the new design now requires a GBP of less than 10MHz. This allowed me to switch to an MCP6022 op-amp, which has a GBP of 10MHz but costs less than half as much as a Pi Pico. It doesn’t have the same low-noise performance as the LT6231, but it is sill good enough that it doesn’t limit the receiver’s performance. This change not only reduces the overall cost of the project but improves the receiver’s overall performance.

Antenna and Enclosure

A random wire antenna in a high location preferably outdoors, or in the attic would be ideal for this type of receiver. If you want an indoor antenna, or something a bit more portable, I have had pretty good results using a you-loop antenna, it is a clever design that cancels noise.

The only downside is that the output level tends to be quite low, so you need a sensitive receiver or some kind of pre-amplifier. I’m just using a cheap wideband LNA I bought online. I have used this setup for all the experiments in this demo.

Enclosure

For the enclosure, I designed a 3D-printed case in FreeCAD. Since this is an experimental receiver, I will be leaving the lid off!

Testing the Receiver

I’ve tested the receiver by tuning into various broadcast stations and exploring the HAM bands. I also connected it to a PC soundcard to experiment with digital modes like FT-8. I have been pleased with the results, with successful signal reception from multiple continents, even using an indoor antenna.

Check out the video of the receiver in action.

Conclusion

This SDR receiver would be a great project for anyone getting started in home-brew radio construction, or perhaps for someone looking for something fun to build with their pi-pico. While it might not live up to expensive commercial radios, its low cost and simplicity make it a valuable tool for receiving signals from around the globe.

Breadboard Radio - Part 2

Exploring the Pi Pico 2

I recently received my Pi Pico 2 and was able to integrate it into the SDR with minimal code changes. Despite initial expectations of only marginal improvements, the performance boost was impressive.

The key enhancements in the Pi Pico 2 are the Floating Point Unit (FPU) and the DSP co-processor. Although the SDR code is written in fixed-point arithmetic, the faster clock frequency of the Pico 2 provides a significant reduction in CPU usage—from around 80% on the original Pico to approximately 40% on the Pico 2.

The Pico 2 offers a choice of processors: the ARM Cortex M33 or the RISC-V Hazard 3. I tested both and found their performance similar in this fixed-point application. The success of the RISC-V processor suggests that we may see more of these processors in future projects.

One curiosity was whether fixing the well-known ADC bug in the Pico would affect the SDR’s performance. After testing the noise floor across various bands, I found no measurable or noticeable difference. Given that the SDR averages hundreds of samples, any occasional bad ADC readings likely get lost in the noise.

The increased clock frequency of the Pi Pico 2 also allows for finer resolution of the local oscillator, enhancing the SDR’s performance, particularly on higher frequency bands. While the improvement is modest—just a couple of kHz—it contributes to a more predictable performance on the high bands.

Receiving Weather FAX

I received a comment asking if the SDR could be used to receive weather FAX. Although I hadn’t tried this before, I gave it a go using a sound card and Fldigi on a PC. The setup worked without any issues, successfully downloading weather maps. This is a really interesting technical solution, it is quite impressive that a narrow band HF channel can be used to transmit FAX data, but it is also a testament to the skill of the people who can read and interpret these maps.

Operational Amplifier Alternatives

Component availability, particularly for op-amps, has been another concern. Fortunately, the op-amps used in this project are not particularly special, and suitable substitutes are widely available. Here are the key specifications to look for in an op-amp for this SDR.

Operating Voltage

The design uses the 3.3V output from the pico to drive the Tayloe detector. This is mainly so that the voltage stays constant regardless of the battery level. It also means that we don’t have to worry about over-driving the ADC which only works at up to 3.3V. A starting point is to select a dual output with a minimum supply voltage of 3V or less.

Gain-Bandwidth Product:

A product of around 10 MHz is sufficient, given the reduced bandwidth of the detector (12 kHz) and a gain of 600.

Noise Performance:

Dan Tayloe’s paper provides some guidance on the noise specification and presents a formula to calculate the Minimum Detectable Signal based on the op-amp performance.

At HF frequencies there is a great deal of noise on the bands, RECOMMENDATION ITU-R P.372-7 tells us the expected level of man-made noise at different frequencies.

The plot compares the expected receiver performance with the levels of band noise. We can see that an amplifier with a noise density of 9 nV/√Hz should offer comparable performance to a typical receiver, and would not limit the performance of the receiver under most circumstances.

I selected a few of the cheapest operational amplifiers that meet this specification and tested them in the receiver.

	GBP (MHz)	Noise Vol age Density (nV/√Hz)	Cost (GBP)	DIP package
MCP6292	10	8.7	0.82	Yes
MCP6022	10	8.7	1.44	Yes
MCP662	60	6.8	1.25	No
OPA2607	50	3.8	1.26	No
OPA1662	22	3.3	1.44	No
LT6231	215	1.1	6.25	No
LTC6227	420	1	7.53	No

All of these devices were tested and worked without any issues. There was no noticeable difference in performance, so it isn’t worth using expensive devices in this design. If you struggle to find the right op-amp, consider adapting the design to work with 5V devices, which broadens the range of available components.

Additional Improvements

I’ve also explored adding an external amplifier and speaker to the SDR. While low-cost PC speakers work well, a built-in speaker could be more convenient, especially for portable use. There are various low-power amplifier options available, allowing you to tailor the setup to your specific needs and budget.

I have opted to use an inexpensive PAM8403 module. These can be obtained cheaply and can provide up to 3W output into a 4 ohm load. It uses a class-D design which allows efficiencies of up to 90% which makes it ideal for portable applications. As supplied, the module uses 10k ohm input resistors, this sets the gain much too high for this application, so I have replaced these with 100k ohm resistors to give a gain of around 2x.

I have included a simple switch to isolate the amplifier when the speaker is not required, but this could also be achieved using a switch in the headphone jack.

There are numerous small speakers to choose from and the quality can be variable. In any case, a sealed enclosure can improve the quality of the sound. Speakers in this price range don’t come with detailed specifications, usually just the impedance and power rating, so it isn’t possible to design anything sophisticated. This page gives some simple rules of thumb.

I used some self-adhesive felt on the rear of the enclosure to help absorb echoes.

Conclusion

This SDR project continues to evolve, with numerous upgrades and improvements planned for the future.