mrtnmcc

You might also want the topic files to be compressed against each other to get a baseline matrix and then multiply any results by the inverse, assuming equal priors on the topics.

Reminds me of Jobs saying.. the problem with phones these days is the bottom 1/3.

Some still disagree.

Notice the log scale on the year axis. There is far more probability mass during older years because of that.

A way to get back at Texas for datacenter restrictions?

Ten pads and thousands of flights per year sure will accelerate Golden Dome, err... I mean "going to Mars"

Just use -- to signal you're a human. That and two spaces after periods :)

This UDRP panel only had jurisdiction over domain name disputes. They still threatened going to court over trademark and we had more important things to do. The new domain is https://ScaleRF.com/

We had a similar result when a big U.S. defense company (Kratos) tried to take our open source project's domain name: open.space

The panel ruled in our favor, that their OPENSPACE trademark is probably invalid because it is descriptive.

https://domainnamewire.com/2026/04/08/u-s-defense-contractor...

A good tutorial on sigma delta is https://www.analog.com/en/resources/technical-articles/sigma...

For a deeper dive, check out any of Richard Schreier's lectures.

Tx/Rx isolation is about -35dB to -40dB depending on the Tx/Rx polarization selection. The QEC image rejection is about -50dBc in Rx, and about -40dBc in Tx (one-time cal, without background cal enabled). The Tx and Rx are perfectly synchronized. This allows you to easily do things like near-field 4x4 MIMO radar: https://www.youtube.com/watch?v=gVqESemU_AI

Yes once you have runtime IQ compensation figured out in digital, its essentially free and easily puts your image rejection ratio into the -50 to -60dBc range. I don't have a write up on the RxQEC, but it's rather trivial: you periodically compute the 2x2 IQ correlation matrix, then take the sqrt (e.g., Cholesky) and multiply your incoming Rx samples by that inverse. You can extend that further for widerband by using FIR taps but here the baseband matching is good enough to go with "single tap"

We don't support passive or active radar beyond basic near-field sensing. We also proactively submitted a detailed report to the State Department earlier this year showing we don't exceed any of the USML criteria for ITAR controls. Separately we have an ECCN determination under EAR. This does rule us out from exporting to a few places (Cuba, Iran, Russia, North Korea, Syria, and some regions of Ukraine). But we are able to ship to most countries.

Thanks for the support! C-band is really the sweet spot in terms of affordability and compact size (the scale of the antennas and array spacing increases proportional to the wavelength). Maybe at some point we'll develop something at 2.4, but more devices are moving into 5GHz these days.

Yes open-space was a short-lived name: https://domainnamewire.com/2026/04/08/u-s-defense-contractor... (good story though) We finally settled on Scale RF for the company and Quad RF for the product (and Moon RF for the bigger phased arrays!). Yes that incident kind of made our brand a mess..

Yes-- but it's open sauce!

Yes, I actually designed the I/Q calibration for many of Analog Device's transceivers (AD93x), and indeed it is a fun problem. If you're interested in what was done for QuadRF, you can read: https://QuadRF.com/cals/txqec.html (Warning: Math!)

We've got the switching noise nailed down. Fortunately the LVDS jitter doesn't affect the sigma delta too badly because that impacts proportional to baseband frequency which is largely filtered out by the decimation filters beyond 40 MHz.. With a total of eight ADCs per QuadRF, you can see we are getting huge savings by being custom! While the per-ADC ENOB is 7-8 bits, another nice thing about phased arrays is that the quantization/ADC noise averages away between elements, so with 8 ADCs in QuadRF we pick up another 1.5 bit giving 8.5-9.5 ENOB, which is frankly better than most SDRs. For the bigger phased arrays that improves further quickly.

We should have a video about MoonRF once we finish the QuadRF mass production, look for it ~ early September! Will be legendary! You can read on how the multi-tile synchronization and calibration works here: https://QuadRF.com/docs/#phased-arrays

Absolutely :) we're working on documenting an awesome Meshtastic demo. Should have a writeup next week to add to the Crowd Supply updates page. Also Roy on our team will be demoing it (along with the RF augmented reality) at Teardown 2026 in Portland if you're in that neck of the woods.

It really depends on the transmitter strength, but if you set the Rx gain high on the QuadRF, we get within 2dB of the thermal noise detection limit.. so about as good as is possible with a receiver this size. I believe a few km is easily doable with a consumer drone but we haven't focused on it.

check out https://QuadRF.com/docs/

Realtime sound visualization was actually a project I did 20 years ago as a freshman and that (probably?) inspired me to build this AR app in QuadRF.

On balance, I would say this RF version was 200x harder.

QuadRF creator here. Happy to answer questions!

We have a quick demo video as well: https://m.youtube.com/watch?v=QvniJk3uNyA

Along with a deeper dive video: https://m.youtube.com/watch?v=zdJ9Tbm8ALg

We didn't give Jeff great direction on camera alignment calibration or setting the radio gain but he seemed to mostly figure it out. We're improving the UI based on his suggestions (it's open source so you can customize it too)

The RF augmented reality is just one of many applications of this brand new 4x4 MIMO software-defined radio built from the ground up. The AR uses a web app to stream RF points that your phone/laptop browser then live-merges with your local camera in the browser. I've been obsessed with low latency and high frame rate to make it a truly AR experience. More technical details at https://QuadRF.com/

That's some extreme paranoia. Don't you think the devs are competent or ethical enough to call out an ask like that?

What does it even mean to "get updates from China"?

Software these days is distributed and globalized in virtually every sense. Polestar is headquartered in Sweden and much of their software development is in the UK.

$499 with Raspberry Pi 5 included.

Crowd Supply campaign: https://www.crowdsupply.com/scale-rf/quadrf#

I'm mostly joking.. LNT is always the best policy, even if only (sometimes) for the feeling of pristineness.

Honestly orange peels are incredible, the smell, the robustness. It reminds me of the joke of the plastic cup at Whole Foods filled with orange slices. If only there was a natural packaging alternative...

Suppose this is the article to give that person who chides you for dropping an orange peel on a hike.

Yes the FPGA does digital beamforming and there will be published FPGA pin constraint files so you can write your own Verilog.

For IQ samples, you can stream a continuous 60 MSPS IQ beam over the Ethernet, but for custom beamforming algorithms, the processing needs to be distributed across the FPGAs (for MoonRF it is a SERDES daisy chain, using the same FFC connectors as the links to the RPi-5).

Yes I'll put some newer block diagrams on the website this week. Thanks for the reminder.

That post is truly hilarious though. If everyone talked like that on the nets, the hobby would surely grow a lot faster.