The 6G spectrum stack: frequency layers and TRX chains — 4 Minutes Ago

No single slice of radio spectrum can do everything a mobile network needs. Low frequencies travel far and slip through walls but carry little data. High frequencies carry enormous data but fade after a few hundred metres and stop at the first obstacle. Every generation of mobile technology has answered this the same way — by stacking complementary bands — and 6G formalises the idea into what Huawei calls a multilayered frequency band framework.

The framework is really two questions layered on top of each other. First: which bands, and what is each one for? Second: once you pick a band, how do you claw back the coverage and capacity that physics keeps trying to take away? The answer to the second question is beamforming — and that is where TRX chains come in. Let’s take the stack first, then the transceivers, then the part most explainers get wrong.

01The layer stack

Read the diagram from the bottom up, low frequency to high. Each layer buys more bandwidth than the one below it and pays for it in range and penetration. The colours track wavelength — the same physics that makes low bands „warm“ and optical „violet“ is what sets their coverage.

Low- and mid-band stay the backbone. Low bands (700–900 MHz) and mid-bands (3–5 GHz) remain the most cost-effective way to blanket an area, and Huawei expects them to be as vital in 6G as in 5G. The interesting new tenants are the upper-mid bands — 6 GHz (5925–7125 MHz) and 10 GHz (10–13.25 GHz), the range 3GPP labels FR3 — which the industry is treating as the sweet spot for early 6G: far more bandwidth than sub-6 GHz, far friendlier propagation than mmWave.

mmWave supplies the raw bandwidth for peak rates and the resolution for centimetre-level sensing, with the E-bands (71–76 and 81–86 GHz) as prime candidates for big contiguous blocks. Terahertz goes further still — roughly 230 GHz of spectrum sits in the 100–450 GHz range — but only over short hops. And at the very top, optical / visible light trades radio entirely for LEDs and photodiodes: no electromagnetic interference, which makes it attractive for radiation-sensitive settings like healthcare and manufacturing.

02What a TRX actually is

TRX is shorthand for a transmit/receive chain — one complete signal path feeding the antenna array. In a massive-MIMO radio, the TRX count is the number of independently controllable paths. More chains means the radio can form narrower beams and push more simultaneous data streams to different users at once (higher-order spatial multiplexing).

That is the whole trick for making high frequencies usable. A higher band loses energy fast, so instead of broadcasting weakly in every direction you concentrate it into a tight beam aimed at each user. Shorter wavelengths also let you pack more antenna elements into the same physical panel — so the higher layers can, in principle, support far more elements than sub-6 GHz ever did.

The move from 64 transceivers to 256 is what turns FR3’s extra megahertz into actual cell capacity.

Real deployment data makes the jump concrete. A typical commercial 5G massive-MIMO cell runs 64 TRX over 100 MHz. Modelling from recent FR3 research scales 6G radios to 128 or 256 TRX over 200–400 MHz — the extra chains are exactly how you convert wider bandwidth into higher beamforming and spatial-multiplexing gain rather than just a bigger, weaker pipe.

03TRX per layer — the comparison

Here is where each layer lands. Read the TRX / RF chains column together with the one beside it — because at the top of the stack, the two decouple.

* Hybrid beamforming: element count runs into the hundreds or thousands, but the number of RF chains (TRX) is deliberately kept lower — analog phase shifters steer sub-arrays that share a chain. Antenna elements and TRX are not the same number here.

04The part explainers get wrong

The tidy version of this story is „higher frequency, more antennas, more TRX, forever.“ It breaks at mmWave. Fully digital massive MIMO — one dedicated transceiver per element — is affordable at mid-band and stretches into FR3 at 128–256 chains. Push into mmWave and terahertz and giving every element its own chain becomes impossible on power and cost grounds, because reduced power-amplifier efficiency at those frequencies is already a headline problem.

So the architecture switches. Hybrid beamforming lets a modest set of RF chains drive a much larger array through analog phase shifters, with each chain steering a sub-array. Element counts explode; TRX counts stay disciplined. That is why the table shows FR3 as the peak of the raw TRX curve, not mmWave — and why anyone quoting „thousands of TRX“ at terahertz is usually counting elements, not transceivers.

05The bill arrives as watts

None of this is free to run, and the cost lands where operators feel it. If 6G radios scale toward 128–256 chains and much wider carriers, per-radio power draw at low load can push past a kilowatt. That makes aggressive power management — TRX sleep modes, carrier shutdown when a cell is quiet — a design requirement rather than a nice-to-have. Today those features are underused in live networks, mostly because operators worry about coverage and reliability the moment a chain goes to sleep. Expect that reluctance to be one of the real fights of early 6G rollout.

06The stack, in one line

The frequency layers and the TRX counts are two axes of a single design. Climb the stack and each layer hands you more bandwidth but worse propagation; the compensation is more transceiver chains and bigger arrays forming tighter beams — right up until mmWave, where the compensation stops being „more TRX“ and becomes „smarter arrays with fewer chains.“ Low bands stay lightly beamformed to hold the coverage floor. The high layers lean on that beamforming to make their bandwidth usable at all. Get both axes right together and 6G works; treat either one in isolation and the numbers stop adding up.