I was asked if I had any thoughts about spectrum for amateur experimentation on LTE.
Yes.
The RF Talent Pipeline
Without coordinated and explicit, carefully designed incentives for allowing spectrum use for experimentation and non-commercial networks, the US will continue to lack a robust RF talent pipeline familiar with modern cellular standards specifically and a hardware talent pipeline generally, and will destroy little by little what meager trickle remains due to the ISM bands.
Friction Sums
Every additional restriction or roadblock adds up, as amply demonstrated by any data-based ecommerce or marketing campaign. Simple things like not having your ‘buy now’ button on the very first page are enough to make significant fractions of visitors with intent to purchase fail to purchase. The same is true even with the relatively low burden of an amateur radio license. What’s the burden of getting usable cellular spectrum for real experimental networks?
Low Friction -> High Velocity
MeshCore and Meshtastic are an example of the motivated volunteer demonstrating significant capability even with a serious handicap. These are functional mesh networks covering huge swaths of land and population without any of the benefits in propagation or transmit power other spectrum might otherwise have.
These unlicensed networks are experiencing some of the fastest innovation in radio today due to accessibility. Anyone that’s interested can join and start experimenting. This is an exciting technology with nearly no barrier to entry, and its wide and growing popularity and capabilities are a testament to light-touch regulation being an effective tool in encouraging innovation. In talking to new hams, I’ve found that many started on unlicensed technology, then ‘graduated’ to licensed services as their interest and experience developed.
Lower up-front barriers -> more innovation, time and time again. Lower barriers also implies faster cycle times.
Self-regulation for experimenters works, because they have the self-interest to make it work (reference the flood vs routing differences between Meshtastic and MeshCore!).
Missing Rung
Modern technologies like 4G, 5G and soon 6G are among the current top rungs on the ladder from unlicensed experimentation to industry employment, but there’s almost no spectrum for the amateur experimenter to learn on despite easy access to open source software implementations of much of the stack.
Ham radio 4G/5G are limited by amateur access to spectrum, and the little spectrum we have is entirely by accident thanks almost exclusively to China. LTE Band 40 is used in China for telecom, but overlaps with US amateur privileges. Just about all cellular modems are global and not region locked, which provides a path forward for US ham experimentation - but only by accident, only with limited spectrum, and for off the shelf eNodeBs/gNodeBs, only by using Chinese hardware intended for their domestic market. A deliberate US domestic path for spectrum would open up the US domestic used-hardware market, and would end that dependence.
Existing Approaches Don’t Fit
CBRS shows why this matters. It was easier and cheaper for me to find a regulatory path forward on Band 40, source LTE Band 40 hardware, and make significant software changes to open source stacks to comply with Part 97 rather than use off the shelf CBRS equipment and deal with the ongoing regulatory and financial burden, because I’m not making money off this. On top of ongoing SAS fees, many installs require a roughly $600 Certified Professional Installer certification. CBRS for real deployed nodes at usable installations is effectively only for commercial ventures, not the hobbyist. It’s the spectrum management that drives that ongoing cost. Hardware is a one-time cost, and unlike a business I typically buy on the used market.
Ofcom-style is too burdensome. Every roadblock that would protect spectrum from experimentation makes it effectively worthless to volunteers and defeats the purpose of opening it up.
An Idea
First: tiny carve-outs on the band edges of US cellular bands, always active, kept for amateur experimental and disaster response networks. That would ensure ham radio LTE and experimentation had a clear and continuous path forward, as well as provide a way and means for the various organizations like modem manufacturers to actually support it. Because it’s in the same bands, COTS hardware would ‘just work’.
Second: the ability to expand from those carve-outs into spectrum that is genuinely unused in an area. By ‘unused in an area’ I am referring to how cell coverage in NH often only actually works by the major roads, and the cities are the only truly dense network coverage. There is ‘coverage’ over most of New Hampshire, but typically only a single band per carrier, if there’s even more than one. There are huge swaths of spectrum that are effectively unused outside of cities because they are uneconomical in low density areas.
Both parts use one rule: a modern, self-regulating, DFS-style scan. Amateur networks expand bandwidth only where they detect no licensee network, and contract on detection of one appearing, whether that’s a commercial network returning after a disaster or a carrier building out a new area. Expansion should be slow and conservative; retreat should be immediate. There’s lots of precedent for sense-and-avoid, to translate an FAA-ism, including 5 GHz Wi-Fi DFS and LTE-LAA listen before talk.
There are already build-out requirements. This would be a new type of build-out requirement for the spectrum owner, but a self-enforcing one: the spectrum owner reasserts their own usage simply by deploying, as population density increases in a given area and new deployments become feasible. No applications, no database, no coordination fees.
Narrow bandwidths that are economically unfeasible to use for spectrum owners are disproportionately more valuable for hams because we don’t need it to be economically viable - it’s donated hardware and airtime. This could also provide genuine economic incentive to allow more agility and wider ranges in cellular bandwidth and carrier aggregation support.
Contingencies
For anything sensing misses, having performance-based standards for response to notification of interference to spectrum holders could go a long way to reassuring without adding significant additional burden. There are possible technological solutions to this too, such as requiring a receiver that listens for a cryptographically signed ‘killswitch’ request - but the key is it has to be simple, and has to be reactive to keep the initial barrier low.
The primary reason to have a requirement to transmit callsigns is so you can identify the transmitter in case there’s an issue. Identification is what makes the reactive model work, and it’s already required.
Beyond Spectrum
The biggest non-spectrum blocker on all this is the lack of a ham radio body that can advocate for getting a ham radio E.164 prefix, a PLMN assignment, carrier configuration on major mobile phones, and all the other complexities that require interfacing with giant commercial companies and coordinating with standards bodies of various types. I think that’s a critical next-5-years goal for ham LTE.
Resilience
Ham radio networks are fantastically resilient, in no small part because they are run by volunteers with the technical skills required, who are already geographically distributed and personally impacted when the infrastructure that they share goes down. Hams already show up to disaster areas like Puerto Rico after Maria in 2017, North Carolina in 2024 after Helene, and many others, with personally owned gear with which they have extensive experience.
Commercial interests and heavy regulations are not sufficient to handle all situations, because being ready for anything is very expensive and there’s always a tradeoff between spend and preparedness in a for-profit venture. Generalists with a broader skillset and a wider array of tools carry the day against the unknown. Building that corps of capable volunteers is only possible in the cellular space if there is room to learn.