pith. sign in

arxiv: 2510.10647 · v1 · pith:DOR56NK4new · submitted 2025-10-12 · 📡 eess.SP

A Parametric Power Model of Upper Mid-Band (FR3) Base Stations for 6G

Emanuele Peschiera , Sangbu Yun , Youngjoo Lee , Liesbet Van der Perre , Fran\c{c}ois Rottenberg This is my paper

Pith reviewed 2026-05-21 21:11 UTC · model grok-4.3

classification 📡 eess.SP
keywords FR3 base stationspower consumption modelhybrid beamformingenergy efficiency6G networkspower amplifiersRF chainssignal processing
0
0 comments X

The pith

Parametric modeling shows hybrid beamforming improves energy efficiency by 1.4 times over fully digital designs in large FR3 base stations.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper develops a parametric power model for prospective 6G base stations in the upper mid-band from 7 to 24 GHz. It accounts for consumption from digital and analog processing, power amplifiers, supply, and cooling across data transmission, signaling, micro-sleep, and idle phases in both downlink and uplink. The work compares hybrid partially-connected beamforming to fully-digital beamforming for arrays up to 1024 antennas. At 30 percent load, the model identifies clear crossover points where power amplifiers dominate at low RF chain counts and processing dominates at high counts. Hybrid designs reach 1.3 Gbit/s per user while using 1.4 times less energy than fully digital approaches.

Core claim

The paper presents a parametric power model for FR3 base stations that quantifies the energy consumed by signal processing, power amplifiers, and auxiliary systems in both downlink and uplink. For arrays with 1024 antennas operating at 30 percent load, the model shows power amplifiers consume the largest share when using 64 or fewer RF chains, whereas digital and analog processing dominate when the number of RF chains reaches 512 or higher. Hybrid partially-connected beamforming achieves a downlink rate of 1.3 Gbit/s per user and provides 1.4 times higher energy efficiency than fully-digital beamforming.

What carries the argument

parametric power model that breaks down consumption into digital and analog processing, power amplifiers, supply, and cooling during data, signaling, micro-sleep, and idle phases for hybrid and fully-digital beamforming

If this is right

  • For 1024-antenna arrays at 30 percent load, power amplifiers are the main consumer with 64 or fewer RF chains.
  • Digital and analog processing becomes the dominant power consumer with 512 or more RF chains.
  • The digital plus analog processing consumes 2 to 4 times more than the PA in fully-digital beamforming.
  • Hybrid beamforming achieves 1.3 Gbit/s per user in downlink with 1.4 times better energy efficiency.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Designers could select RF chain count based on this model to minimize total power for given array sizes.
  • The crossover points suggest different hardware optimization priorities depending on chosen architecture.
  • Extending the model to uplink-heavy scenarios or real traffic patterns could refine deployment guidelines.

Load-bearing premise

The specific power consumption values and scaling rules used for digital/analog processing, power amplifiers, and cooling systems at FR3 frequencies accurately represent prospective base-station hardware.

What would settle it

Direct power measurements from a hardware prototype of an FR3 base station with 1024 antennas under varying RF chain configurations and load conditions, compared against the model's predictions.

read the original abstract

Increasing attention is given to the upper mid-band or Frequency Range 3 (FR3), from 7 to 24 GHz, in the research towards sixth-generation (6G) networks. Promises of offering large data rates at favorable propagation conditions are leading to novel FR3 base station (BS) architectures, with up to thousands of antenna elements and radio-frequency (RF) chains. This work investigates the power consumption of prospective FR3 BSs and its relation to the delivered data rates. We model the power consumed by digital and analog signal processing, power amplifiers (PAs), and supply and cooling during four phases (data, signaling, micro-sleep, and idle) in downlink and uplink. Hybrid partially-connected beamforming is compared to fully-digital one. Results show that, for BS arrays with $1024$ antennas at $30\%$ of load, the PA consumes most of the power when $64$ or less RF chains are utilized, while the digital and analog processing consumption takes over when the number of RF chains is $512$ or more. The digital plus analog processing consumes $2\times$ to $4\times$ more than the PA for fully-digital beamforming. Hybrid beamforming achieves $1.3$ Gbit/s/user in downlink while improving the energy efficiency by $1.4\times$ compared to fully-digital beamforming.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 1 minor

Summary. The manuscript develops a parametric power consumption model for prospective 6G base stations in the upper mid-band (FR3, 7-24 GHz). The model accounts for digital/analog signal processing, power amplifiers, and supply/cooling overhead across four phases (data, signaling, micro-sleep, idle) in downlink and uplink. It compares hybrid partially-connected beamforming to fully-digital architectures for arrays up to 1024 antennas, reporting that at 30% load the PA dominates power for 64 or fewer RF chains while processing dominates for 512 or more, with hybrid beamforming delivering 1.3 Gbit/s/user and 1.4× better energy efficiency than fully-digital.

Significance. If the underlying component values prove representative, the model supplies a useful tool for exploring energy-efficiency trade-offs in large FR3 arrays, particularly the crossover between PA and processing dominance and the relative advantage of hybrid beamforming. The parametric structure supports design-space exploration even if specific numerical outcomes require further grounding.

major comments (1)
  1. [Abstract and numerical results] Abstract and numerical results: The headline claims (PA dominance for ≤64 RF chains, processing dominance for ≥512 RF chains, and 1.4× EE gain for hybrid at 1024 antennas / 30 % load) are obtained by substituting specific per-component power expressions (digital/analog processing linear in N_RF and bandwidth, PA efficiency at FR3 frequencies, cooling overhead) into the four-phase model. These expressions are presented without new measurements, traceable derivations, or references to FR3 hardware data; a 2–3× shift in the processing coefficient or modest change in PA efficiency would alter the reported crossover points and EE ordering. This is load-bearing for the central quantitative conclusions.
minor comments (1)
  1. [Results] The 30 % load factor is treated as a fixed input; a brief sensitivity plot or table showing how the dominance thresholds and EE ratio vary with load would improve interpretability without altering the core model.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive review and for recognizing the potential utility of the parametric model for exploring energy-efficiency trade-offs in large FR3 arrays. We address the major comment in detail below.

read point-by-point responses

  1. Referee: [Abstract and numerical results] Abstract and numerical results: The headline claims (PA dominance for ≤64 RF chains, processing dominance for ≥512 RF chains, and 1.4× EE gain for hybrid at 1024 antennas / 30 % load) are obtained by substituting specific per-component power expressions (digital/analog processing linear in N_RF and bandwidth, PA efficiency at FR3 frequencies, cooling overhead) into the four-phase model. These expressions are presented without new measurements, traceable derivations, or references to FR3 hardware data; a 2–3× shift in the processing coefficient or modest change in PA efficiency would alter the reported crossover points and EE ordering. This is load-bearing for the central quantitative conclusions.

    Authors: We agree that the specific numerical outcomes depend on the selected parameter values and that these values are not derived from new measurements performed in this work. The manuscript is a modeling study that adapts established parametric expressions for digital/analog processing power (linear in N_RF and bandwidth) and PA efficiency from the existing literature, with adjustments for FR3 frequencies drawn from published hardware characterizations. We will revise the manuscript to (i) add explicit citations and traceable derivations for each key expression in a new appendix, (ii) include additional references to recent FR3-relevant hardware data where available, and (iii) incorporate a sensitivity analysis demonstrating how ±2–3× variations in processing coefficients or modest changes in PA efficiency affect the reported crossover points and the 1.4× energy-efficiency advantage. These additions will clarify the grounding of the quantitative results while preserving the parametric character of the model that enables design-space exploration. revision: partial

Circularity Check

0 steps flagged

Parametric model applies external scaling rules; no reduction of claims to inputs by construction

full rationale

The paper constructs a parametric power-consumption model for FR3 base stations and evaluates it across antenna counts, RF-chain counts, and load factors. The headline results (PA vs. processing dominance thresholds, 1.4× EE advantage of hybrid beamforming) are direct numerical outputs of that model once its per-component expressions and coefficients are inserted. No equation is shown to be defined in terms of its own output, no fitted parameter is relabeled as a prediction, and no load-bearing premise rests on a self-citation whose validity is presupposed by the present work. The derivation chain therefore remains open to the chosen parameter values rather than closing on itself.

Axiom & Free-Parameter Ledger

2 free parameters · 1 axioms · 0 invented entities

The central claim rests on a set of component power values and scaling assumptions that are not derived within the paper but taken as inputs to the parametric model.

free parameters (2)
  • Component power consumption values
    Power figures for digital/analog processing, PAs, and cooling are required to produce the reported numbers but are not derived from first principles in the abstract.
  • Load factor of 30%
    The 30% load condition is used to generate the headline results and is an input choice rather than an output.
axioms (1)
  • domain assumption The four operational phases and the hybrid partially-connected architecture accurately represent prospective FR3 base stations.
    These modeling choices are invoked to structure the power calculations.

pith-pipeline@v0.9.0 · 5798 in / 1410 out tokens · 54359 ms · 2026-05-21T21:11:20.048586+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Measurement-Based Massive MIMO Channel Characterization and Performance Evaluation at FR3 (8 and 15 GHz) Under Equal Physical Aperture

    eess.SP 2026-04 unverdicted novelty 6.0

    Under equal physical aperture, 15 GHz FR3 measurements show higher spectral efficiency than 8 GHz due to more antenna elements overcoming increased sparsity, despite a 3 dB coverage deficit.

Reference graph

Works this paper leans on

36 extracted references · 36 canonical work pages · cited by 1 Pith paper · 2 internal anchors

  1. [1]

    INTRODUCTION The upper mid-band, roughly ranging from 7 to 24 GHz, is one of the main candidate frequency bands to provide a10×capacity in- crease in sixth-generation (6G) networks [1]. This band is expected to trade-off the benefits of sub-6 GHz bands in terms of propaga- tion and coverage, with the large bandwidths offered by millimeter wave (mmWave) fr...

    work page

  2. [2]

    A Parametric Power Model of Upper Mid-Band (FR3) Base Stations for 6G

    SYSTEM MODEL As depicted in Fig. 1, we consider a BS equipped with2M ant antennas, PAs and low-noise amplifiers (LNAs), and2M RF radio- frequency (RF) chains that are equally split for use in DL and UL. The BS servesKsingle-antenna users using space-division multiple access (SDMA) and orthogonal frequency-division multiplexing (OFDM). The carrier frequenc...

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  3. [3]

    POWER MODEL FOR A FR3 BASE STATION In this section, we describe the main BS components and subcom- ponents, and their power consumption models. We subdivide the BS components in digital processing, analog processing, and PAs, and express the total BS power consumption averaged over the frame as Pcons = Pdig ηdig,s/c + Pana ηana,s/c + PPA ηPA,s/c (5) where...

    work page

  4. [4]

    We consider a 3GPP 39.801 Urban Micro (UMi) wideband channel model generated by the Sionna sim- ulator [30]

    NUMERICAL EV ALUATION In this section, we present the evaluations of power consumption and ergodic rates in a FR3 scenario. We consider a 3GPP 39.801 Urban Micro (UMi) wideband channel model generated by the Sionna sim- ulator [30]. The BS array is a uniform planar array (UPA) having up toM ant = 1024antenna elements and half-wavelength spacing in both di...

    work page

  5. [5]

    We consider time-division duplexing (TDD) operation withτ DL = 0.75, τUL = 0.25,τ DL,sig =τ UL,sig = 1/14andζ DL,sig = 1/12[8]

    The noise powers are set asσ 2 n,DL =σ 2 n,UL =−123dBm. We consider time-division duplexing (TDD) operation withτ DL = 0.75, τUL = 0.25,τ DL,sig =τ UL,sig = 1/14andζ DL,sig = 1/12[8]. We futher setη dig,s/c =η ana,s/c =η PA,s/c = 0.8. According to (1), the average physical loads are bounded as xDL ∈[0,1]and xUL ∈[0,1]. In Fig. 2, we show the power consump...

    work page

  6. [6]

    The consumption of each digital subcomponent is adapted to the hardware architecture, being this FPGA or ASIC

    CONCLUSION In this study, we proposed a power model for FR3 BSs that quantifies the consumption of digital, analog, PA, and supply and cooling com- ponents by averaging over different working modes in a time frame, from active to idle. The consumption of each digital subcomponent is adapted to the hardware architecture, being this FPGA or ASIC. Up-to-date...

    work page

  7. [7]

    6G takes shape,

    J. G. Andrews, T. E. Humphreys, and T. Ji, “6G takes shape,” IEEE BITS Inf. Theory Mag., vol. 4, no. 1, pp. 2–24, 2024

    work page 2024

  8. [8]

    Cellular wireless net- works in the upper mid-band,

    S. Kang, M. Mezzavilla, S. Rangan, A. Madanayake, S. B. Venkatakrishnan, G. Hellbourg, et al., “Cellular wireless net- works in the upper mid-band,”IEEE Open J. Commun. Soc., vol. 5, pp. 2058–2075, 2024

    work page 2058

  9. [9]

    New mid- band for 6G: Several considerations from the channel propa- gation characteristics perspective,

    J. Zhang, H. Miao, P. Tang, L. Tian, and G. Liu, “New mid- band for 6G: Several considerations from the channel propa- gation characteristics perspective,”IEEE Commun. Mag., vol. 63, no. 1, pp. 175–180, 2025

    work page 2025

  10. [10]

    Comprehensive FR1(C) and FR3 lower and upper mid-band propagation and material penetration loss measurements and channel models in indoor environment for 5G and 6G,

    D. Shakya, M. Ying, T. S. Rappaport, H. Poddar, P. Ma, Y . Wang, et al., “Comprehensive FR1(C) and FR3 lower and upper mid-band propagation and material penetration loss measurements and channel models in indoor environment for 5G and 6G,”IEEE Open J. Commun. Soc., vol. 5, pp. 5192– 5218, 2024

    work page 2024

  11. [11]

    6G prototyping in the upper mid-band (7–24 GHz),

    M. Mezzavilla, A. Rasteh, M. Zappe, E. Zappe, A. Dhananjay, and S. Rangan, “6G prototyping in the upper mid-band (7–24 GHz),” inProc. IEEE Wireless Commun. Netw. Conf., 2025, pp. 1–3

    work page 2025

  12. [12]

    Enabling 6G performance in the upper mid-band by transitioning from massive to gigantic MIMO,

    E. Bj ¨ornson, F. Kara, N. Kolomvakis, A. Kosasih, P. Ramezani, and M. B. Salman, “Enabling 6G performance in the upper mid-band by transitioning from massive to gigantic MIMO,” IEEE Open J. Commun. Soc., pp. 1–1, 2025

    work page 2025

  13. [13]

    Comparison of power consumption models for 5G cellular network base stations,

    A. M. Busch, K. Eger, and B. Richerzhagen, “Comparison of power consumption models for 5G cellular network base stations,” inIntelligent Distributed Computing XVI, M. K¨ohler- Bußmeier, W. Renz, and J. Sudeikat, Eds., Cham, Switzerland, 2024, pp. 199–212, Springer Nature Switz

    work page 2024

  14. [14]

    A parametric power model of multi-band sub-6 GHz cellular base stations using on-site measurements,

    L. Golard, Y . Agram, F. Rottenberg, F. Quitin, D. Bol, and J. Louveaux, “A parametric power model of multi-band sub-6 GHz cellular base stations using on-site measurements,” in Proc. IEEE Int. Symp. Pers. Indoor Mob. Radio Commun., 2024, pp. 1–7

    work page 2024

  15. [15]

    A flexible power model for mm-wave and THz high-throughput communication systems,

    C. Desset, P. Wambacq, Y . Zhang, M. Ingels, and A. Bour- doux, “A flexible power model for mm-wave and THz high-throughput communication systems,” inProc. IEEE Int. Symp. Pers. Indoor Mob. Radio Commun., 2020, pp. 1–6

    work page 2020

  16. [16]

    Energy efficiency of mmWave massive MIMO precoding with low-resolution DACs,

    L. N. Ribeiro, S. Schwarz, M. Rupp, and A. L. F. de Almeida, “Energy efficiency of mmWave massive MIMO precoding with low-resolution DACs,”IEEE J. Sel. Top. Signal Process., vol. 12, no. 2, pp. 298–312, 2018

    work page 2018

  17. [17]

    Capacity and power consumption of multi-layer 6G networks using the up- per mid-band,

    D. L ´opez-P´erez, N. Piovesan, and G. Geraci, “Capacity and power consumption of multi-layer 6G networks using the up- per mid-band,” inProc. IEEE Int. Conf. Commun., 2025

    work page 2025

  18. [18]

    On Optimizing Time-, Space- and Power-Domain Energy-Saving Techniques for Sub-6 GHz Base Stations

    E. Peschiera, Y . Agram, F. Quitin, L. Van der Perre, and F. Rot- tenberg, “On optimizing time-, space- and power-domain energy-saving techniques for sub-6 GHz base stations,”arXiv preprint arXiv:2505.15445, 2025

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  19. [19]

    Zynq UltraScale+ RFSoC ZCU216 Evaluation Kit,

    AMD, “Zynq UltraScale+ RFSoC ZCU216 Evaluation Kit,” Evaluation board user guide (UG1390), Dec. 2023

    work page 2023

  20. [20]

    FPGA architecture: Principles and progression,

    A. Boutros and V . Betz, “FPGA architecture: Principles and progression,”IEEE Circuits Syst. Mag., vol. 21, no. 2, pp. 4– 29, 2021

    work page 2021

  21. [21]

    Erik Dahlman, Stefan Parkvall, and Johan Sk ¨old,5G NR: The Next Generation Wireless Access Technology, Academic Press, New York, NY , USA, 2nd edition, 2020

    work page 2020

  22. [22]

    Complexity optimized digital predistortion model of RF power amplifiers,

    W. Cao, S. Wang, P. N. Landin, C. Fager, and T. Eriks- son, “Complexity optimized digital predistortion model of RF power amplifiers,”IEEE Trans. Microw. Theory Tech., vol. 70, no. 3, pp. 1490–1499, 2022

    work page 2022

  23. [23]

    A frequency hopping software-defined radio platform for com- munications and sensing in the upper mid-band,

    M. Mezzavilla, A. Dhananjay, M. Zappe, and S. Rangan, “A frequency hopping software-defined radio platform for com- munications and sensing in the upper mid-band,” inProc. IEEE Int. Workshop Signal Process. Adv. Wireless Commun., 2024, pp. 611–615

    work page 2024

  24. [24]

    Energy-constrained modulation optimization,

    S. Cui, A. J. Goldsmith, and A. Bahai, “Energy-constrained modulation optimization,”IEEE Trans. Wireless Commun., vol. 4, no. 5, pp. 2349–2360, 2005

    work page 2005

  25. [25]

    A 28-nm 8-bit 16-GS/s DAC with>60 dBc/>40 dBc SFDR up to 2.3 GHz/5.4 GHz using 4-channel NRZ-output-overlapped time-interleaving,

    S. Chen, C. Huang, L. Sun, Y . Liu, W. Tang, J. Fan, et al., “A 28-nm 8-bit 16-GS/s DAC with>60 dBc/>40 dBc SFDR up to 2.3 GHz/5.4 GHz using 4-channel NRZ-output-overlapped time-interleaving,”IEEE Trans. Circuits Syst. II, Exp. Briefs, vol. 72, no. 2, pp. 374–378, 2025

    work page 2025

  26. [26]

    ADC Performance Survey 1997-2025,

    B. Murmann, “ADC Performance Survey 1997-2025,” [On- line]. Available: https://github.com/bmurmann/ADC-survey

    work page 1997

  27. [27]

    A 10GS/s hierarchical time-interleaved ADC for RF- sampling applications,

    N. Markuli ´c, J. Nguyen, J. L. Benites, E. Martens, and J. Cran- inckx, “A 10GS/s hierarchical time-interleaved ADC for RF- sampling applications,” inProc. IEEE Symp. VLSI Tech- nol. Circuits, 2024, pp. 1–2

    work page 2024

  28. [28]

    A 10–36-GHz I/Q mixer-first RX using current-mode coupler cascade in 16-nm FinFET,

    M. Helal, A. Helaly, and G. M. Rebeiz, “A 10–36-GHz I/Q mixer-first RX using current-mode coupler cascade in 16-nm FinFET,”IEEE Microw. Wireless Technol. Lett., pp. 1–4, 2025

    work page 2025

  29. [29]

    A four-element true-time-delay slice-based receiver array for FR3 upper mid-band wireless,

    H. Abbasi, A. Slater, F. Beheshti, S. Poolakkal, and S. Gupta, “A four-element true-time-delay slice-based receiver array for FR3 upper mid-band wireless,”IEEE J. Solid-State Circuits, pp. 1–15, 2025

    work page 2025

  30. [30]

    A multi-band 5.3–18 GHz LNA for FR3 bands using hybrid switching with sub-2 dB NF in FDSOI,

    A. Afifi and G. M. Rebeiz, “A multi-band 5.3–18 GHz LNA for FR3 bands using hybrid switching with sub-2 dB NF in FDSOI,”IEEE J. Solid-State Circuits, pp. 1–11, 2025

    work page 2025

  31. [31]

    ADMV8818: 2 GHz to 18 GHz, Digitally Tunable, High-Pass and Low-Pass Filter,

    Analog Devices, “ADMV8818: 2 GHz to 18 GHz, Digitally Tunable, High-Pass and Low-Pass Filter,” Data Sheet (Rev. B), 2021

    work page 2021

  32. [32]

    ADF41510: 10 GHz, Integer-N/Fractional- N PLL Synthesizer,

    Analog Devices, “ADF41510: 10 GHz, Integer-N/Fractional- N PLL Synthesizer,” Data Sheet (Rev. A), 2024

    work page 2024

  33. [33]

    Energy efficiency performance of LTE dynamic base station downlink DTX operation,

    J.-F. Cheng, H. Koorapaty, P. Frenger, D. Larsson, and S. Fala- hati, “Energy efficiency performance of LTE dynamic base station downlink DTX operation,” inProc. IEEE 79th V eh. Technol. Conf., 2014, pp. 1–5

    work page 2014

  34. [34]

    Power amplifiers performance survey 2000–present,

    H. Wang, M. Eleraky, B. Abdelaziz, et al., “Power amplifiers performance survey 2000–present,” 2025, [Online]. Available: https://ideas.ethz.ch/research/surveys/pa-survey.html

    work page 2000

  35. [35]

    Analysis and design of a SiGe BiCMOS PA for 6G FR3 band with 29-dBm PPAE and 40.1% PAE,

    D. Pecile, A. Gambarucci, S. Kokorovic, and A. Bevilacqua, “Analysis and design of a SiGe BiCMOS PA for 6G FR3 band with 29-dBm PPAE and 40.1% PAE,”IEEE Trans. Microw. The- ory Tech., pp. 1–12, 2025

    work page 2025

  36. [36]

    Hoydis, S

    J. Hoydis, S. Cammerer, F. Ait Aoudia, M. Nimier- David, L. Maggi, G. Marcus, et al., “Sionna,” 2022, https://nvlabs.github.io/sionna/

    work page 2022