tinySA Input Impedance: 50 Ohms, Attenuation, Overload and Spurs
tinySA Input Impedance: 50 Ohms, Attenuation, Overload and Spurs
“The input is 50 ohms” is incomplete. Model, connector, frequency range, mode, attenuation, LNA state, firmware and reference plane all belong in the sentence.
Fifty ohms is a defined reference condition, not a myth. Nominal input impedance, attenuation setting, overload testing and calibrated S11 measurement answer different questions. The official tinySA documentation describes several different front ends, so a result from one path cannot be generalized to every tinySA.
Start With the Exact Hardware and Path
The official comparison page distinguishes tinySA Basic, tinySA Ultra ZS405, Ultra+ ZS406 and Ultra+ ZS407. On the Basic, LOW and HIGH are different receiver paths. Ultra and Ultra+ use a normal range plus an extended Ultra mode whose upper frequency and calibration boundary depend on the hardware version. Record the rear label where applicable and capture CONFIG → VERSION, connector, frequency range, attenuation, LNA, reference level, RBW, detector/averaging and spur-removal state before interpreting a trace.
| Path | Current official boundary | What “50 ohms” means here | Level caution |
|---|---|---|---|
| Basic LOW | 100 kHz–350 MHz; manual/automatic internal attenuation 0–31 dB | The Basic specification says 50 ohms when internal attenuation is at least 10 dB. | +10 dBm absolute maximum at 0 dB attenuation; +0 dBm is the suggested maximum with automatic attenuation; the specification advises below −25 dBm for best measurements. |
| Basic HIGH | 240–960 MHz; no user-settable step attenuator | The Basic specification explicitly says impedance is frequency dependent and deviates from 50 ohms. | +10 dBm absolute maximum without attenuation. This direct-IQ path has limited mirror suppression and is easily overloaded; out-of-band signals also matter. |
| Ultra/Ultra+ normal range | 100 kHz–800 MHz for ZS405; 100 kHz–900 MHz for ZS406/ZS407; 0–31 dB attenuation with LNA off | The Ultra specification says 50 ohms when internal attenuation is at least 10 dB. | +6 dBm absolute maximum at 0 dB attenuation, +20 dBm short-term peak only with 30 dB attenuation, ±5 V maximum DC, and +0 dBm suggested with automatic attenuation. |
| Ultra mode | Extended range and level-calibrated upper limit depend on ZS405/406/407 | Do not transplant a normal-range result across the whole extended range. The input low-pass filter is bypassed above its passband. | Mirror/spur rejection uses repeated measurements; short, scanning, complex or wideband signals can be missed or misplaced. RF-port LO leakage can reach −10 dBm. |
Those are operating boundaries from the cited official pages as checked on 29 August 2026, not a substitute for the documentation matching the unit in hand. Current official figures are path- and condition-specific: published Basic HIGH and comparison values differ with front end, tone spacing and LNA state. A bare product name and one intercept number are therefore insufficient. Cite the hardware, firmware, mode, spacing, level and document revision with the result.
What 50 Ohms Actually Says
At a defined frequency and reference plane, the complex input impedance Zin can be expressed through the reflection coefficient referenced to Z0 = 50 Ω:
Γ = (Zin − Z0) / (Zin + Z0)
Return loss = −20 log10|Γ|
VSWR = (1 + |Γ|) / (1 − |Γ|)
An official “50 ohms at 10 dB or more” condition identifies the intended front-end state; it does not promise zero reflection at every frequency. The measurable quantities are S11, return loss or VSWR over frequency, with uncertainty and the calibration plane stated.
Attenuation can improve the match seen upstream
A well-matched attenuator leaves the bare impedance behind it unchanged while reducing the reflected wave that returns to the source. For an ideal matched pad with A dB one-way loss:
|Γin| = |Γload| × 10−A/10
An ideal 10 dB pad therefore reduces the load-reflection magnitude by a factor of ten, improving the apparent return loss by 20 dB.
Real pads add their own mismatch, frequency limit, power limit and uncertainty. They also raise the analyzer-referred noise floor by their insertion loss. An internal attenuator can be the reason the connector presents a better nominal match, because its physical location is part of the front-end topology. “The downstream load is unchanged” and “the source sees a better termination through the pad” are both true when their reference planes are stated.
Measure S11 If the Question Is Input Match
A spectrum trace cannot establish connector-plane impedance. Use a vector network analyzer and a controlled one-port reflection measurement:
- Freeze the tinySA state. Select the intended connector, mode, frequency range, fixed attenuation and LNA state. Disable automatic changes that could switch the RF path during the sweep.
- Set a safe VNA stimulus. Check the model-specific tinySA input and DC limits, and also protect the VNA from any tinySA LO leakage. Start at low source power. A characterized isolating pad or DC block may be appropriate, but it must be included in the calibration or de-embedded.
- Calibrate at the intended plane. Perform a one-port calibration with suitable open, short and load standards at the end of the cable or adapter that will mate to the tinySA. Keysight's calibration guidance describes this fixed reference plane and its systematic error correction.
- Verify the calibration. Reconnect a known load or verification standard before the DUT. Connector repeatability, adapters and a flexing cable can dominate a good-looking trace.
- Sweep the useful band. Save S11 or return loss for every relevant fixed attenuation/LNA/mode state, with hardware and firmware identity. If an external pad is left between the calibration plane and DUT, the result is the padded assembly unless that pad is characterized and removed mathematically.
The tinySA self-test and input-level calibration serve different purposes. Official level-correction instructions use a calibrated generator for input amplitude and a calibrated power meter for output amplitude. They do not turn the tinySA into a VNA, move a reflection reference plane or calibrate S11.
Attenuation Stepping Diagnoses Overload, Not Impedance
The tinySA attenuation pages recommend increasing attenuation and watching a displayed fundamental and its harmonics. Because the display compensates for internal attenuation, a genuine external component should remain approximately stationary while the noise floor rises. If a product falls as more attenuation is added, it was at least partly generated inside the analyzer. That is a valuable overload check—but still not an input-match measurement.
Use the test carefully:
- Hold frequency span, RBW, detector/averaging, reference level, LNA and spur-removal settings constant.
- Confirm that the source itself does not change with load, level or time. For transmitter work, sample a rated dummy-load path rather than using the tinySA as the load.
- Repeat with an external attenuator at the connector. Internal attenuation protects stages after its position; an external pad also reduces the signal before the connector and first internal device.
- Expect the noise floor to rise with added loss. A disappearing product may have moved below the floor, not ceased to exist.
- In Ultra mode, repeat with a suitable external preselector and documented spur-removal state when mirrors or wideband/transient behavior are plausible.
One tone is not a two-tone IMD3 test
A single applied tone can reveal harmonics, source contamination, images, LO/mixer products and unrelated internal spurs. The third harmonic is a third-order nonlinear product, but calling every single-tone mystery line “IMD3” confuses it with the standard two-tone intermodulation test.
A two-tone IMD3 test applies tones at f1 and f2 and looks near them at 2f1 − f2 and 2f2 − f1. In the cubic small-signal region, third-order product power grows approximately 3 dB for each 1 dB rise in tone power. A defensible test uses two clean, isolated sources, a suitable combiner and filters, verifies the source/combiner products without the DUT where possible, and changes analyzer attenuation to prove that the analyzer is not creating the result.
“Dynamic Range” Needs a Measurement Definition
Displayed range, measurement range, DANL, compression-to-noise range, phase-noise-limited range, image rejection and spur-free dynamic range are not synonyms. Keysight's spectrum-analyzer guidance defines dynamic range as the ratio between simultaneously present large and small input signals that can be measured to a stated uncertainty. Rohde & Schwarz likewise shows that attenuation prevents overload but raises the analyzer-referred noise floor; phase noise can dominate close to a carrier.
So a tinySA SFDR number belongs with its model, path, frequency, RBW, signal level, attenuation, LNA state, tone spacing and spur-removal method. The current official comparison table itself reports Ultra-family IIP3 differently for 500 Hz and at least 2 MHz tone spacing. A clean trace or a low displayed floor is useful evidence only for the documented setup, not a universal “dynamic range” specification.
A Safe Transmitter-Sampling Chain
- Calculate transmitter power in dBm and include modulation peak, overshoot, coupler tolerance, attenuator tolerance, cable loss uncertainty and fault cases.
- Rate the load, sampler, connectors, cables, attenuators and DC block for frequency, peak/average power and duty cycle. A small SMA attenuator is not automatically a transmitter-power attenuator.
- Verify sampler output independently with a suitable power sensor at low transmitter power. Entering an EXT GAIN value only offsets the display; it does not measure the actual path.
- Start with generous external attenuation and the tinySA LNA off. Approach the intended analyzer level from below, staying within current model/path guidance rather than using the absolute maximum as a target.
- Perform the attenuation-step check. For harmonic work, the official tinySA guide also recommends verifying that displayed products remain stationary after another 5 or 10 dB is inserted.
The official Safe Operation page gives minimum external-attenuation examples of 40 dB for 1 W, 50 dB for 10 W and 60 dB for 100 W, and calls for a DC block plus at least 3 dB attenuation above 5 V DC. Treat those as minimum damage-prevention examples, not a complete uncertainty, spectral-purity or component-rating design.
A Repeatable Bench Record
- Identity: genuine hardware model/revision, connector and firmware shown by VERSION.
- State: mode, Ultra enablement, frequency/span, RBW/VBW, points, detector/averaging, reference level, attenuation, LNA, AGC and spur removal.
- RF chain: generator/transmitter, filters, combiner/coupler, every cable/adapter/pad/DC block, their calibration data and reference planes.
- Level evidence: independently measured fundamental/peak power, insertion losses, tolerances, warm-up and repeatability.
- Classification test: attenuation sweep, source-off/background trace, changed span/RBW, alternate preselection and—in an IMD claim—a proper two-tone setup.
- Match evidence: separate calibrated VNA S11 sweep in the same fixed front-end state.
Authoritative sources checked
- Official tinySA model comparison—current Basic, ZS405, ZS406 and ZS407 identity, ranges, input limits and condition-dependent linearity figures.
- tinySA Basic specification, Basic level menu, Basic HIGH input guidance and Basic limitations.
- tinySA Ultra/Ultra+ specification, Ultra level menu, Ultra-mode operation and Ultra limitations.
- Official Safe Operation table, transmitter-output measurement guide, attenuation/overload check and harmonic-measurement guide.
- Official firmware-update notes and Ultra/Ultra+ level-correction menu—version capture, pre/post-update self-test and amplitude-calibration boundaries.
- Keysight, Optimizing RF and Microwave Spectrum Analyzer Dynamic Range—dynamic-range definitions, attenuation, noise, distortion and phase-noise limits.
- Keysight, Specifying Calibration Standards and Kits for Vector Network Analyzers—systematic error correction and fixed calibration reference planes.
- Rohde & Schwarz, Understanding Basic Spectrum Analyzer Operation and Understanding Third Order Intercept—front-end overload, attenuation and two-tone IM3 verification.
Mini-FAQ
- Is every tinySA input always 50 ohms? No. The official 50-ohm condition applies to Basic LOW and Ultra/Ultra+ inputs with at least 10 dB internal attenuation. Basic HIGH is explicitly frequency dependent and deviates from 50 ohms.
- Does 10 dB internal attenuation guarantee a perfect 50-ohm match? No. It selects the state for which the official specification says 50 ohms; actual return loss remains frequency-, unit- and setup-dependent and requires an S11 measurement.
- Can a matched attenuator make the source see closer to 50 ohms? Yes. It reduces the load reflection on the round trip, although it does not change the bare load, adds loss and contributes its own mismatch and uncertainty.
- Can a changing spur prove that the input impedance changed? No. Spur movement during an attenuation sweep is evidence about overload or an internally generated product, not connector-plane S11.
- Is a single-tone mystery peak IMD3? Not by itself. A standard IMD3 result needs at least two tones and products at 2f1 − f2 and 2f2 − f1; a one-tone artifact may instead be a harmonic, image, spur or source impurity.
- What is the safest way to sample a transmitter with a tinySA? Use a power-rated dummy-load or through-line path, a characterized directional sampler, enough external attenuation for worst-case peak power and a rated DC block when needed; verify the level independently before keying at full power.