Pi Pico Rx – A crystal radio for the digital age? | Hacker News Reader

Pi Pico Rx – A crystal radio for the digital age?

101-things.readthedocs.io

145 pointsnolist_policy24 comments

Neat project I do wonder what more could be done with the RP2350.

Note that the RP2350, being M33 based has square root, and can add 16MB of PSRAM behind the XIP that will help with the flash writing problem, The solder party RP2350 Stamp XL has a SOP8 pad to add PSRAM and is actually shipping for anyone who is looking to play with them, no connection with them, just a user.

The extra PIO block would also allow for S2C output which may be nice rather than a TPA2012D2, which will require most people to buy a breakout board anyway.

The rp2040/rp2350 series is incredible. The DMA/SIO/PIO features are just so powerful and even the esp32/stm32 seem restrictive to me to go back to. It is amazing what people are doing with them.

(2023)

Very nice. It's not a "crystal radio", though. It's a direct sampling receiver. That is, the processing takes place at the RF frequency directly. Most receivers use heterodyne conversion to down-convert a desired chunk of spectrum to a convenient working frequency. This doesn't. It just runs the raw RF through an A/D.

As A/D devices and processing have improved, the frequency upper limit for direct conversion receivers has moved up. There are low-cost direct conversion receivers going up to 50MHz, but they have to use an FPGA rather than doing all the work in software.

The comparison is not with the crystal but with the low parts count and simplicity.

It looks like it is a Direct Conversion (Zero IF) rather than a Direct Sampling Receiver [0] - ie. it uses does a quadrature LO to down convert to baseband I/Q (in this case a Taylor QSD as a mixer). The RP2040 ADC bandwidth is only 250KHz.

[0] https://www.analog.com/en/resources/technical-articles/a-rev...

Oh, you're right. Affordable direct sampling receivers for low bands do exist, but this isn't one.

Very cool write up! I'm amazed that it's running on AAA batteries.

The introduction to SDR (software defined radio) is much appreciated.

Edit: defined, not designed

software defined radio

The staggered quadrature sampling trick also comes up in the Crank-Nicholson method for solving e.g. the Schrodinger equation, optics etc.

The LTC6227, used in this design, is "last time buy." The replacements aren't as good: either higher noise or lower GBW, and they too will vanish in a few years. That is, of course, deliberate, forcing designers to climb the curve of ever more expensive, exclusive components. Another component in this design, the FST3253, is also a figment of history: no longer produced. Electronics is a nasty rat race, with AD, for example, having bought literally all of their competitors except TI, we have an perfect ADC/DAC/op-amp oligopoly, with everything getting worse with each passing day. So the notion of a cheap, easily reproduced "crystal radio for the digital age" isn't actually workable: nothing you can dream up will be buildable more than 2-3 years after you design it as the components you've chosen cease to exist.

Not sure I would have gotten very far with this as a 7 year old, but as a 62 year old it looks quite interesting :-)

That's really cool, but as long as AM exists, crystal radio will be crystal radio

Indeed very simple and within reach of a beginner.

Another illustration how RPi 2040 PIO is an exceedingly helpful thing. Without it, the RF-related circuits would have to use either bit-banging, or a bunch of small discrete components.

Certainly within reach of a beginner to build, but I wonder how many beginners would struggle to understand the details of a SDR. Knowledge of processors, memory, IQ sampling, DSP, etc. are not typical subject matter for beginners.

The original crystal radios were very simple, with a crystal (or later, a 1N34 diode), an optional (low Q) bandpass filter, and a piezoelectric earphone -- no batteries required. The principles of operation could be explained to a beginner in just a few minutes.

When I was a kid, I optimized my first "crystal" radio down to four components:

1) 1N34 diode.

2) 22k Ohm resistor (impedance matches the earphone).

3) Piezoelectric earphone.

4) Antenna wire (with alligator clip).

It worked well because there was only one radio station nearby. (KNX 1070, 50KW was about a mile away.)

By this token, the physics of semiconductors, which is required to explain how the crystal detector receiver works, is also uncommon knowledge among beginners!

Both computers and semiconductors can be explained qualitatively, bringing in the ideas like a one-way valve, or following written instructions. This, of course, would gloss over a ton of things, much like the optimized crystal radio blissfully lacks the RF filter in the audio frequency circuit: the mechanical properties of the earphone suffice.

"All models are wrong, but some are useful", as they say.

When I look at that crystal radio kit, my fingers hurt. I had a kit like that as a kid in the 90s - not sure if it was a radio or more generic electronics - with the springs to hold wires together. At some point, made a short circuit and burned my fingers. Immediately upon seeing that, the memory comes back. I guess it's burned into my memory.

I loved those spring electronics kits, Radio Shack always sold a bunch. And yeah, the D batteries could get the springs glowing and wires smoking. A lot of my early electronics projects were smoke machines.

Surely it wasn’t a crystal radio you were building, since they have no power source?

For any powered circuit though, a few AAs can dump a surprising amount of current and heat into naive fingers.

I was looking at digital fm chips a few years back that specifically pick up digital fm and more importantly to my project at the time digital text that stations broadcast.

Deviation aside this is a cool project and impressive as someone said that it runs on alkaline batteries.

I’ve been looking at using the RDA5807M ic to handle the RDS and audio output. Seems to be pretty simple and cheap.

Was looking into Meshtastic recently, as a means to have basic communication in disaster situations (hurricanes, war, etc.) but haven't picked a device yet. Any recommendations?

I think the Meshtastic site mentions a few boards, but it probably depends too on what features you want. I picked up a couple Heltec boards with tiny displays to tinker with, though haven't spent enough time there yet.

I've got a Lilygo T-Beam with a 18650 battery for backup power, an external battery, and a 3D printed case. It's fine for a stationary setup at home.

For more portable use, you might go with something like a Lilygo Meshtastic T-Echo.

https://www.ebay.com/itm/186858359561?mkcid=16&mkevt=1&mkrid...

Had one late last year, but had to return it because the firmware bricked within a few hours of use and no support was forthcoming. Wouldn't leap in just yet. The issue was triggered by running a repeater, whereas client-only use was fine.

Just get a radio? Not to be glib, they're well built, mature tech that can go many tens of km from elevated positions. Don't get triangulated in a conflict situation, but it's the thing most likely to work and remain durable/portable under extreme conditions.

Reply on news.ycombinator.com