My bench is set up for independent verification across three measurement domains. Each instrument provides an independent path to validate the signal characteristics I care about.
Frequency range: 300 kHz – 40 GHz. Measures S-parameters (S11 return loss, S21 insertion loss). This is my primary tool for characterizing cable plants, connectors, and passive components.
Calibration: Full SOLT (Short-Open-Load-Through) before every measurement session. Calibration verification kit (CVK) run after calibration to confirm the standard is holding.
Acceptance: S11 > 15dB across operating band. S21 within channel loss budget per signal integrity standards (Q4503810).
Certifies copper cabling to TIA/EIA-568-C.2 and ISO/IEC 11801 Class EA / Cat 6A. Measures all channel parameters: insertion loss, NEXT, FEXT, PSNEXT, PSACR, return loss, propagation delay, skew.
Acceptance: Must pass "Margin" criterion—measured values must exceed the standard's limits by a sufficient margin to account for test equipment uncertainty and cable degradation over time.
4-channel, 400 MHz bandwidth. Used for protocol-level validation: eye diagram analysis, jitter decomposition, rise/fall time measurement. This is where I catch timing issues that a network analyzer can't see.
Acceptance: Eye opening must meet BER < 10⁻¹² for the target data rate. Random jitter + deterministic jitter must fall within the receiver's tolerance budget.
These are the thresholds I apply when certifying a cable plant for production use. Every run must pass all criteria before it's accepted.
| Parameter | Cat 6 (250 MHz) | Cat 6A (500 MHz) | Fiber OM4 (850nm) |
|---|---|---|---|
| Insertion Loss (max) | 20.8 dB @ 100m | 27.9 dB @ 100m | 3.8 dB @ 100m |
| Return Loss (min) | 14.0 dB @ 250MHz | 14.0 dB @ 500MHz | 25.0 dB |
| NEXT (min) | 23.5 dB @ 250MHz | 27.9 dB @ 500MHz | N/A |
| PSNEXT (min) | 20.5 dB @ 250MHz | 24.9 dB @ 500MHz | N/A |
| Propagation Delay (max) | 496 ns/100m | 496 ns/100m | 4.8 ns/m |
| Delay Skew (max) | 45 ns | 45 ns | 9 ns (100m) |
Source values derived from ISO/IEC 11801 channel specifications and signal integrity (Q4503810) reference data. These are working numbers from actual field certifications, not theoretical limits.
Instrument accuracy decays over time. My schedule:
| Instrument | Frequency | Procedure | Record |
|---|---|---|---|
| HP 8753E | Annual (NIST-traceable lab) | Full factory recalibration with certificate | Log: /bench/cert/8753e |
| Fluke DSX-5800 | Semi-annual | Calibration verification + adjustment if drift > 1% | Log: /bench/cert/dsx5800 |
| Keysight DSOX4000MS | Annual | Internal self-cal + external verification | Log: /bench/cert/dsox4000 |
| All instruments | Weekly (bench check) | CVK verification against known reference standards | Log: /bench/weekly/ |