The TDR trace on channel 47 was off. Not by much—about 12 picoseconds longer than the mean of the other 127 channels in the bundle. That's 12 trillionths of a second. On the screen it looked like a tiny blip, barely outside the ±1% tolerance band. But when we plugged that cable into the customer's 128-channel ultrasound beamformer and imaged a test phantom, there it was: a subtle brightness artifact at the focal zone, exactly where channel 47's delayed signal was constructively interfering at the wrong depth.
Phase matching in micro coaxial cable bundles is the kind of engineering discipline that's invisible when it's done right and painfully obvious when it's not. Every channel in a multi-coaxial bundle must carry its signal at the same velocity, within extremely tight tolerances, so that the beamforming electronics can coherently combine signals from dozens or hundreds of transducer elements. Get the phase wrong and you get beam steering errors, focal degradation, and ultimately—in medical imaging—diagnostic image quality that doesn't meet clinical requirements.
Why Phase Matching Matters for Beamforming
Ultrasound beamforming works by applying precise time delays to the signals from each transducer element. These delays steer and focus the acoustic beam—and the system calculates them based on the assumption that the electrical path from each element to the beamformer is identical. Any variation in cable propagation delay adds an unintended time offset that the beamformer doesn't know about.
The math is unforgiving. At 7.5 MHz (a common mid-range ultrasound frequency), one wavelength in tissue is about 0.2 mm. A 1% phase error across a 2-meter cable corresponds to approximately 0.15 ns of timing error, which is about 5% of a wavelength at 7.5 MHz. That's small enough to be acceptable for most imaging Applications . A 3% error pushes you to 15% of a wavelength, and at this point lateral resolution visibly degrades—the beam focus broadens, and fine structures in the image start to blur.
For phased array probes doing beam steering (not just focusing), the tolerance is even tighter because steering errors accumulate across the aperture. A systematic phase gradient across the array—which can happen if the cable channels aren't properly randomized in the bundle—doesn't just blur the focus; it shifts the entire beam, creating geometric distortion in the image.
Inside the Manufacturing Process: Where Phase Matching Is Won or Lost
Phase matching isn't something you "add" to a cable at the end of the process. It's the cumulative result of controlling every manufacturing step from conductor drawing to final bundle assembly. Here's where each step contributes—and where things go wrong.
Step 1: Conductor Drawing and Plating
The center conductor determines the inner boundary of the coaxial geometry. For a 42 AWG 7-strand conductor , each strand is drawn to approximately 50 AWG (0.025 mm) and then silver-plated to 2–5 μm thickness. Conductor OD needs to be held to ±0.003 mm across the entire length of the cable—a 0.006 mm total variation window on a 0.064 mm conductor. Any diameter drift shifts the characteristic impedance and, more relevantly for phase matching, changes the VoP at that point along the cable.
We measure incoming conductor diameter with a laser micrometer at 100 mm intervals during incoming inspection. Spools that drift more than ±0.002 mm get set aside for single-channel Applications where phase matching isn't a concern. This is one of those things you only figure out after 10,000 terminations—the conductor you use for phase-matched bundles needs to be selected, not just accepted from the wire supplier's standard tolerance.
Step 2: Dielectric Application
This is the single most critical step for phase matching. The dielectric layer—whether ePTFE, FEP, or PFA —determines the velocity of propagation through its dielectric constant (Dk). VoP = c/√Dk, so any variation in Dk directly translates to VoP variation, which is phase mismatch.
For FEP melt extrusion, the key control parameters are melt temperature (±2°C), line speed (±0.5%), and die geometry (concentricity ≤5 μm). A well-tuned extrusion line produces dielectric wall thickness variation of ±0.005 mm, which translates to roughly ±0.3% VoP variation. That's well within our ±1% phase matching target for individual elements.
For ePTFE, the challenge is different. The dielectric is applied by wrapping a thin ePTFE tape around the conductor under controlled tension. The porosity of the ePTFE—which determines its effective Dk—depends on the stretch ratio, tension, and temperature during the original tape expansion process and again during wrapping. We've measured Dk variations of ±0.05–0.10 within a single ePTFE tape roll, which corresponds to ±0.5–0.8% VoP variation before you even consider wrapping uniformity. Achieving ±1% phase matching with ePTFE requires sorting cable elements by measured VoP and grouping matched elements into bundles—an additional step that adds time and cost.
Step 3: Shield Application
The outer conductor (shield braid or spiral wrap) has a secondary effect on impedance and VoP. Shield coverage, braid angle, and mechanical tension during braiding all influence the effective outer conductor diameter. In practice, shield variation contributes less than 0.1% to VoP spread—it's there, but it's small compared to the dielectric contribution.
That said, we learned this the hard way on one project: if the braiding tension isn't uniform, the braid can compress the dielectric unevenly, creating localized VoP disturbances. The TDR trace looked clean when measured on a straight cable, but when bent to 20 mm radius, channels with tighter braid showed 0.5–0.8% VoP shift while loose-braid channels barely moved. The bending wasn't creating the phase error—it was revealing a pre-existing non-uniformity masked by the cable's straight resting state.
Step 4: Cabling (Bundling)
Individual coaxial elements are assembled into a multi-channel bundle using a planetary cabling machine. This step introduces two phase-relevant effects that aren't obvious from the spec sheet.
Geometric path length difference. In a concentric bundle (1+6+12+18+...), the outermost layer channels follow a longer helical path than the innermost channels. If the cable isn't designed with helical transposition—where channels periodically swap between inner and outer positions—the outer channels can be 0.5–1.5% electrically longer than the inner channels. This is a systematic error that doesn't average out and directly violates phase matching specs.
Cabling tension-induced VoP shift. The act of wrapping channels around each other applies radial pressure to the inner layers. This pressure slightly compresses the dielectric, raising its effective Dk and reducing VoP. In extreme cases (too-high cabling tension on soft dielectric), the inner channels can run 0.3–0.5% slower in VoP than the outer channels. We control this by setting cabling tension to the minimum needed for bundle integrity—typically 50–150 grams per element depending on gauge and dielectric type.
TDR Testing: The Quality Gate for Phase Matching
Every phase-matched bundle we ship gets 100% TDR tested—every channel, measured against every other channel. No sampling, no statistical inference. If you're buying a 128-channel cable for ultrasound beamforming and your supplier is testing only a random sample of channels, ask why.
Our TDR setup uses a Tektronix DSA8300 sampling oscilloscope with a 80E10B TDR module, giving us 4.5 ps rise time and sub-1 ps timing resolution. Each channel is measured for electrical length at 23±1°C (temperature-controlled room, because PTFE-based dielectrics have a VoP temperature coefficient of roughly 50–100 ppm/°C—a 5°C ambient swing can shift phase by 0.03–0.05%, which matters when your tolerance is ±1%).
What the TDR Data Tells You
Beyond pass/fail, the TDR data distribution reveals process health. A well-controlled manufacturing process produces a tight Gaussian distribution of electrical lengths—all channels clustered near the mean with small standard deviation. When we see the distribution skewing or developing tails, it signals a process drift before any individual channel actually fails the tolerance.
Honestly, the TDR measurement itself is sometimes harder than the manufacturing. The cable fixture has to make consistent contact on all 128 channels simultaneously, each with less than 50 mΩ contact resistance. A dirty or worn fixture can add 2–5 ps of measurement uncertainty per channel, which eats into your tolerance budget when you're trying to hold ±1% on a 13 ns cable (that's ±130 ps total tolerance). We replace our TDR fixtures every 2,000 measurements and verify calibration against a NIST-traceable delay standard daily.
| Failure Mode | TDR Signature | Root Cause | Corrective Action |
|---|---|---|---|
| Systematic fast/slow channel groups | Bimodal distribution—inner channels slow, outer channels fast | Excessive cabling tension compressing inner channel dielectric | Reduce cabling tension; verify with in-process TDR at cabling step |
| Single outlier channel | One channel outside ±1% while rest pass | Dielectric void, conductor splice, or localized contamination during extrusion | Replace individual element; trace element to source spool; inspect extrusion logs |
| Gradual drift across all channels | All channels trending toward one end of tolerance band | Dielectric extrusion temperature drift; ePTFE tape Dk batch shift | Check extrusion PID settings; verify incoming ePTFE tape Dk per-roll |
| Phase shift under bending | Channels pass straight but fail when cable bent to service radius | Non-uniform braid tension causing position-dependent dielectric compression | Tighten braid tension tolerance; implement bent-cable TDR as production test |
| Temperature-dependent phase drift | Channels pass at 23°C but fail at 35°C | Channel-to-channel variation in dielectric density creating differential thermal coefficients | Tighten dielectric density control; sort elements by measured VoP temperature coefficient |
VoP Control: The Number Behind Phase Matching
Phase matching is ultimately VoP matching. The velocity of propagation through each coaxial element must be consistent to the same tolerance as the phase matching spec. For a ±1% phase matching target on a 2-meter cable, you need VoP consistency of ±1% across all channels—which means the effective Dk of every channel must match within ±2% (since VoP goes as 1/√Dk, the Dk tolerance is roughly double the VoP tolerance).
In the 3,000+ phase-matched bundles we've shipped for ultrasound Applications over the past five years, we've found that the VoP distribution narrows significantly when we implement two practices that most competitors skip: incoming dielectric material VoP screening (measuring actual VoP on short test cables from every material spool before using it in production) and in-process TDR at the cabling step (catching geometric-induced phase errors before final assembly and termination).
These steps add about 8–12% to manufacturing cost. Every customer who's compared our phase matching Cpk data against competitors' has decided the premium is worth it. Your mileage may vary, but for OEMs whose image quality reputation depends on consistent beamforming performance across every probe they ship, the math is clear.
The Relationship Between Phase Matching and Channel Count
Achieving ±1% phase matching on a 16-channel cable is routine—most competent manufacturers can do it. At 64 channels, it requires good process control and 100% TDR testing. At 128 channels, it's genuinely difficult. At 256 channels... we've done it, but it took us three development iterations and a significant investment in process automation to get first-pass yield above 85%.
The difficulty scales non-linearly because the probability of having at least one outlier channel increases with channel count. If individual element VoP has a 1% probability of exceeding the ±1% tolerance (a 2-sigma process), the probability of zero failures in a 128-channel bundle is about 28%. Put differently—you'd reject 72% of bundles at final test. That's why phase-matched production requires at minimum a 3-sigma process on individual elements, and preferably 4-sigma for 128+ channel bundles.
If you need a phase-matched cable assembly for an ultrasound probe or PAUT transducer and you're not sure what tolerance your beamformer actually requires, contact us with your transducer specifications and operating frequency —we can calculate the maximum allowable cable phase error based on your imaging requirements and recommend a specification that balances performance against realistic manufacturing yield.
Frequently Asked Questions
What phase matching tolerance is required for ultrasound beamforming cables?
Most OEMs specify ±1% electrical length matching, corresponding to approximately ±5 ps at 5 MHz over a 2-meter cable. Premium 3D/4D matrix array probes may require ±0.5%. The tolerance is set by allowable beam steering error—each 1% of mismatch shifts the focal point by roughly 0.2 mm at typical imaging depths.
How is phase matching measured?
TDR (Time Domain Reflectometry) with sub-5 ps resolution is the primary method. Each channel is measured individually for electrical length and compared against all other channels. Measurements are taken at controlled temperature (23±2°C) because PTFE dielectrics shift VoP at approximately 50–100 ppm/°C. VNA phase measurement at operating frequency provides complementary verification.
Can phase mismatch be corrected electronically?
Modern beamformers include per-channel delay compensation, but corrections beyond ±3% introduce beam pattern artifacts. Additionally, if cable phase mismatch drifts with temperature or flex position, the factory calibration becomes inaccurate during clinical use. The practical approach is minimizing cable phase error to ±1% and using electronics only for fine-tuning residuals.
So when you're evaluating a cable supplier's phase matching capability, ask for their TDR data from the last 50 production lots—not a cherry-picked best-case report. The distribution shape and Cpk index will tell you whether their process actually holds ±1% in production or whether they're relying on final-test screening and high scrap rates to ship compliant product. It's the difference between a manufacturer who controls phase matching and one who merely inspects for it. And if you've been through a beamformer calibration nightmare that traced back to cable phase inconsistency, you already know which one you'd rather work with.
Related Products
FRS Technology builds these assemblies to order - core count, gauge, jacket, connector and length to your drawing. Products relevant here:
Contact us with your application requirements. We reply with a technical proposal and pricing within 48 hours.
