Trace Weaver: Dual-Branch Architectures

• Trace Weaver is a class of dual-branch signal architectures that manipulate trace properties in C*-algebras and enable precise image selection in RF circuits.
• It combines mathematical trace analysis with practical phased-array design, reducing LO tuning ranges and optimizing bandwidth for E-band transceivers.
• The approach unites abstract trace space selection in operator algebras with engineered shared IF mixer architectures for enhanced spectral efficiency and scalability.

Trace Weaver denotes a class of dual-branch signal architectures and their associated trace analysis in two distinct domains: (1) the construction of $C^\ast$-algebras with prescribed trace spaces as developed in counterexamples to Naimark’s problem, and (2) the Weaver image-selection architecture for phased-array transceivers employing shared intermediate-frequency (IF) image-rejection mixers. Both leverage the manipulation and selection of “traces” in their respective mathematical or circuit frameworks—either as convex affine invariants of operator algebras or as frequency-converting signal paths with explicit control over bandwidth, image rejection, and local oscillator (LO) utilization.

1. Conceptual Foundations and Definitions

In the context of operator algebras, traced in Vaccaro’s and Akemann–Weaver’s work, the trace space $T(A)$ of a unital $C^\ast$-algebra $A$ refers to the closed convex set of tracial states: functionals $\tau$ such that $\tau(ab) = \tau(ba)$ for all $a,b \in A$ and $\|\tau\| = 1$. These constitute the fixed-point set under the action of the unitary group $U(A)$ by inner automorphisms. The structure of $T(A)$—a Choquet simplex under mild hypotheses—classifies $C^\ast$-algebras up to affine homeomorphism.

Conversely, in millimeter-wave phased array transmitters, the “Weaver architecture” specifies an analog signal-processing structure enabling frequency conversion with reduced LO tuning range, explicit selection between sideband images, and robust impedance/power matching for large-N-element arrays. Central to this architecture is the bidirectional shared image-selection IF mixer and the orchestration of signals through quadrature paths, exploiting phase inversion to toggle between upper (USB) and lower (LSB) sidebands (Ebrahimi, 2019).

2. Trace Spaces and Counterexamples to Naimark’s Problem

A counterexample to Naimark’s problem is a unital $C^\ast$-algebra, not isomorphic to the algebra of compact operators $K(H)$ for any Hilbert space $H$, which nevertheless has only one irreducible representation up to unitary equivalence ($U(A)$ acts transitively on $P(A)$, the extreme points of the state space $S(A)$). Under Jensen’s diamond principle ($\diamondsuit$), Akemann and Weaver constructed such algebras using a transfinite tower indexed by $\omega_1$, amalgamating pure states at each successor stage via asymptotically inner automorphisms to collapse all inequivalent pure states into a single orbit in the limit.

Crucially, the trace space $T(A)$ of these nonseparable counterexamples can be structured with maximal flexibility: for any metrizable Choquet simplex $X$, there is a simple, unital, nuclear, non-type I $C^\ast$-algebra of density $\aleph_1$ whose $T(A)$ is affinely homeomorphic to $X$. Through an appropriate selection of amalgamations and “trace-splitting” at successor stages, even nonmetrizable (i.e., nonseparable) Choquet simplices are realized as trace spaces (Vaccaro, 2017).

3. Weaver Image-Selection Architecture in Phased Arrays

The Weaver architecture for E-band (71–76/81–86 GHz) phased-array transceivers is defined by the following key components:

• Each of the $N=16$ antenna elements is equipped with a bidirectional PA/LNA, RF mixer, and IF amplifier.
• IF signals from all elements are combined through a 3 GHz-wide bandpass filter, then routed to shared IF mixers responsible for final up/down conversion in the Weaver stage.
• The LO system is centralized: a single 19.5 GHz injection-locked oscillator (ILO), digitally phase-shifted and multiplied ×4 locally, generates the ≈78 GHz LO required for RF mixers.
• Sideband selection across the full 10 GHz E-band is achieved by a 1-bit image-select inverter at the shared IF stage, flipping the sign on one quadrature branch, thus toggling between LSB and USB operation without broad LO retuning (Ebrahimi, 2019).

This organization permits a minimal (3 GHz) LO tuning span for comprehensive E-band coverage, simplifies the LO phase distribution problem, and enables narrowband, low-distortion phase steering ($\pm30^\circ$).

4. Frequency Mapping, Mixer Topology, and LO Management

Signal routing and frequency planning in the Weaver image-selection architecture exploit the symmetric placement of sub-bands (71–76 GHz and 81–86 GHz) around a central LO ($f_0 = 78$ GHz). By formulating the mixing as:

• $$f_{\mathrm{IF1}} = f_{\mathrm{LO}} - f_{\mathrm{RF}}$$ for LSB,
• $$f_{\mathrm{IF1}} = f_{\mathrm{RF}} - f_{\mathrm{LO}}$$ for USB,

and combining both under a shared IF passband, it becomes possible to keep $|f_{\mathrm{RF}} - f_{\mathrm{LO}}|$ within a single 3 GHz-wide window. The 1-bit quadrature phase inversion implements the image selection. Direct conversion would require a 15 GHz LO tuning range (≈15% FBW); two-point sliding-IF schemes halve this to ≈7.5 GHz (10%), but the Weaver approach reduces this further to 3 GHz (4% FBW) by leveraging dual image access without extra LO complexity (Ebrahimi, 2019).

The bidirectional shared IF mixer, realized as a Gilbert-ring, operates in both transmission (active splitting) and reception (active combining) modes, maintaining S₁₁ < –10 dB over the 3 GHz IF, with careful impedance matching across chips and PCB-level distribution.

5. Performance Metrics and Implementation

The 16-element phased array exploiting Weaver image-selection demonstrates:

• Full E-band beam steering ($\pm30^\circ$) with stabilized amplitude/phase balance (±1 dB / ±2.5° over 3 GHz).
• EIRP of 28–32 dBm, average ≈30 dBm, with per-element $P_{\text{sat}} \approx 8$ dBm.
• RX conversion gain ≈32 dB over the 10 GHz band.
• Measured EVM ≤ –19 dB for 16QAM (8 Gb/s), ≤ –24 dB for 64QAM (9 Gb/s) modulations; with variation across bands limited to ±2 dB.
• Power efficiency: $P_{\mathrm{DC}}$ per element = 250 mW (TX), 160 mW (RX); total array $P_{\mathrm{DC}} \approx 4$ W, yielding $EIRP/P_{DC} \approx 25\%$ (best published at E-band).
• Implementation: 90 nm SiGe BiCMOS, four 2×2 dies, on-chip LO/IF/antenna distribution, aperture-coupled balun with $<0.5^\circ/0.02$ dB differential error (Ebrahimi, 2019).

The architectural advantages derive primarily from the LO tuning reduction and shared IF processing, enabling greater scalability and lower mixer/LO dissipation.

6. Mathematical and Physical Implications

In operator algebra, the capacity to realize any metrizable or even nonmetrizable Choquet simplex as a trace space challenges classical intuitions drawn from finite-dimensional theory or type I algebras. Transitivity of the unitary action, which implies at most one fixed point (trace) in these settings, fails in the nonseparable, non–type I regime constructed by Akemann–Weaver and Vaccaro. This exemplifies a dramatic expansion of possible trace simplex structures, including Bauer, Poulsen, Cantor, and nonmetrizable simplices, thus refuting naive extensions of Glimm’s dichotomy in the nonseparable arena (Vaccaro, 2017).

In electronics, the Weaver architecture traces the dualities of image selection and signal path symmetry to minimize hardware complexity and maximize spectral efficiency. This interplay of theory and hardware enables performance regimes previously limited by LO, IF, and phase-shifter design constraints.

7. Broader Impact and Future Directions

The set-theoretic independence (dependence on $\diamondsuit$) of certain operator algebraic constructions suggests rich boundary phenomena between analysis, topology, and logic, particularly in the classification programs for operator algebras. Similarly, the physical realization of large phased arrays at E-band, leveraging shared image-rejection and traceable LO/IF paths, points to new paradigms in scalable RF circuit design.

A plausible implication is that further exploration of trace selection—whether in operator algebra or RF circuits—may yield insights into both the mathematical landscape of state space convexity and the engineering of high-performance multi-band front-ends. The demonstrated flexibility, efficiency, and compactness of the Weaver paradigm are likely to inform both theoretical frameworks (in trace theory and nonseparable $C^\ast$-algebra analysis) and future block-level circuit architectures for microwave and mmWave systems (Vaccaro, 2017, Ebrahimi, 2019).