VNA Calibration
A vector network analyzer fresh from the factory — or fresh out of the box when you received it — has not been calibrated to your cables. Every cable, adapter, and connector between the VNA's physical ports and your device under test introduces errors: reflections from connector mismatches, phase shift due to cable length, and attenuation from cable loss. Without calibration, these errors add to your measurement and cannot be distinguished from the behavior of the device you are trying to measure.
VNA calibration is the process of measuring known standards — a short circuit, an open circuit, and a matched load — to characterize the VNA's internal errors and the test port cables up to the measurement reference plane. Once calibrated, the VNA mathematically subtracts these known errors from every subsequent measurement, leaving only the response of the device under test. This is not a factory calibration — it is a user calibration performed every time you change the frequency range, cables, or test fixture. It takes about two minutes with a NanoVNA.
Sources of VNA Measurement Error
Three categories of systematic errors affect VNA measurements. "Systematic" means they are repeatable — the same error occurs each time you measure under the same conditions. Because they are repeatable, they can be measured and corrected by the calibration process.
Directivity error: The directional coupler inside the VNA that separates incident and reflected waves is not perfect. It allows some of the incident signal to leak into the reflected-signal measurement path. This makes a perfectly matched device appear to have a small reflection. A VNA with 30 dB of raw directivity cannot measure return losses better than about 25–28 dB without calibration. After calibration, effective directivity rises to 40–50 dB or better.
Source match error: The signal source inside the VNA is not a perfect 50 Ω source. It has a small impedance deviation that affects measurements of devices with high reflections. This is corrected by the "open" and "short" calibration steps, which allow the VNA to characterize the source impedance accurately.
Port match and tracking error: The cables and connectors between the VNA and the calibration plane have length (which adds phase shift) and loss (which adds amplitude error). These errors change the phase and magnitude of every measurement at every frequency. The SOL calibration measures this path and corrects for it.
Calibration Standards: Short, Open, Load, Through
Four physical standards are used in VNA calibration. They are called SOLT — Short, Open, Load, Through. For a one-port measurement (S11 only), you need only Short, Open, and Load (SOL). For two-port measurements, you also need a Through connection.
Short (S): The center conductor and ground of the test port connector are directly connected together (zero length). This creates a total reflection of the incident signal — all power reflects with a 180° phase reversal. The VNA expects to see |Γ| = 1 at −180°. Any deviation from this is recorded as an error and corrected.
Open (O): The test port connector is left unconnected (or has only a mechanical "open cap" to keep connector geometry consistent). This also creates a total reflection — all power reflects with 0° phase shift. The VNA expects to see |Γ| = 1 at 0°. In practice, the open end of a connector has a small parasitic capacitance (0.02–0.1 pF) that causes the phase of the reflection to differ slightly from 0°. Quality calibration kits include a polynomial model of this capacitance, stored in the VNA's memory, that is applied during the calibration calculation.
Load (L): A precision 50 Ω resistor matched to the reference impedance. It should have no reflection — the VNA expects S11 = −∞ dB (or practically, S11 below −40 dB for a good load standard). The load is the "zero point" for reflection measurements.
Through (T): Port 1 and Port 2 are connected together (zero-length through, or a known short-length "thru"). This establishes the 0 dB reference for S21 transmission measurements and corrects for the difference in cable lengths between the two ports.
VNA calibration standards: Short (silver), Open (gold), 50 Ω Load, and Through. These four known standards allow the VNA to characterize and correct systematic measurement errors in the test cables and internal hardware.
View LargerThe SOL Calibration Process
During SOL calibration, the VNA measures the reflection coefficient (S11) of each standard at every frequency in the sweep. It knows exactly what the measurement should be for each standard (short: |Γ| = 1, ∠ = −180°; open: |Γ| = 1, ∠ = 0°; load: |Γ| ≈ 0). By comparing measured values to expected values, it solves for the three error terms at each frequency:
- Ed — directivity error
- Er — reflection tracking error (the response of the reflected signal path)
- Es — source match error
These three complex error terms at each frequency completely describe the one-port systematic errors. With these errors known, every subsequent S11 measurement is corrected by the inverse of these errors. The mathematics involves solving a system of three equations from the three measured standards.
Think of the three error terms as three unknowns. You need at least three independent equations to solve for three unknowns. Each calibration standard (short, open, load) provides one equation (the difference between the measured S11 and the known ideal S11). Three standards → three equations → three unknowns solved. With only one or two standards, the error correction would be underdetermined and only partially valid.
nanovna-procedure">Step-by-Step NanoVNA Calibration
- Set your frequency range first. Calibration is valid only for the frequency range used when the calibration was performed. If you want to measure from 1 MHz to 50 MHz, set that range before calibrating. After calibration, do not change the start or stop frequency, or you must recalibrate.
- Attach your test cables. Connect the cables and adapters that will be used in the actual measurement. The calibration must be performed with everything in the signal path except the DUT itself. If you will use a SMA-to-BNC adapter during measurement, include it during calibration.
- Navigate to: CALIBRATE. On the NanoVNA, press the menu button, select CALIBRATE. You will see options: OPEN, SHORT, LOAD, THRU, DONE, RESET.
- Attach the OPEN standard to Port 1 (leave Port 2 unconnected for now). Select OPEN on the menu. The NanoVNA sweeps the frequency range and records the open measurement. Wait for it to complete (usually 2–5 seconds).
- Attach the SHORT standard to Port 1. Select SHORT. Wait for completion.
- Attach the 50 Ω LOAD to Port 1. Select LOAD. Wait for completion.
- Connect Port 1 to Port 2 with the THROUGH connection. Select THRU. Wait for completion. If you will only make one-port S11 measurements, you can skip this step.
- Select DONE. The calibration data is applied immediately. The NanoVNA can also save calibration to one of its memory slots (C0–C4) — do this if you want to reload the calibration after power cycling.
The Calibration Reference Plane
The calibration reference plane is the physical point in the test setup where calibration was performed — the point where you connected the short, open, and load standards. All measurements are referenced to this plane. The VNA has no knowledge of anything between the calibration plane and the DUT — that length of cable, that adapter — is invisible to the VNA only because it was included in the calibration.
If you calibrate at the VNA's physical port and then add a 12-inch cable to reach the DUT, the 12-inch cable is not corrected — its phase shift and loss will appear in your measurements. Always calibrate at the measurement reference plane, meaning at the point where you will connect the DUT.
You want to measure the impedance of a 40m dipole feedpoint located 30 feet away. You run a 30-foot coaxial cable from the NanoVNA to the feedpoint.
Wrong approach: calibrate at the NanoVNA port, then connect the 30-foot cable, then connect to the antenna. The 30-foot cable is not corrected and introduces 3–5° per foot of phase error at 7 MHz — about 100–150° of total phase error. The impedance reading will be completely wrong.
Right approach: run the 30-foot cable from the NanoVNA to the feedpoint first. Then connect the short, open, and load calibration standards at the far end of the cable (at the feedpoint). Calibrate with all three standards there. Then disconnect the standards and connect the antenna. Now the 30-foot cable is included in the calibration reference plane.
Verifying Your Calibration
After calibration, perform these quick checks before trusting any measurements:
Load verification: Connect the 50 Ω load standard again. S11 should be below −40 dB across the entire calibrated frequency range. If S11 is higher than −40 dB with the load standard connected, the calibration is poor — usually caused by a dirty or damaged connector, or use of the standards out of order.
Open verification: Connect the open standard. S11 should be between −0.1 and +0.1 dB (essentially 0 dB). The phase should show a smooth curve that starts near 0° at low frequencies and shifts slightly (due to the open's capacitance model) at high frequencies.
Through verification (for S21): Connect Port 1 directly to Port 2. S21 should be 0 ± 0.5 dB across the entire frequency range. S11 should be below −30 dB with the ports directly connected.
Common Calibration Mistakes
| Mistake | Effect on Measurements | How to Avoid |
|---|---|---|
| Calibrating with no cables, then adding cables for measurement | Phase and amplitude errors proportional to cable length, worst at high frequencies | Always calibrate with all cables attached |
| Changing frequency range after calibration | Calibration invalid outside original range; measurements outside that range are uncorrected | Set frequency range before calibrating |
| Using low-quality calibration kit | Poor short/open definitions introduce fixed errors, typically visible as ripple in S11 | Use the cal kit that came with the VNA; better kits improve accuracy at higher frequencies |
| Loose or dirty connectors | Variable contact resistance causes irreproducible measurements | Clean connectors before each cal; do not overtighten (especially SMA: 0.5 N·m / 5 in·lb) |
| Temperature change after calibration | VNA components drift with temperature, invalidating error correction | Allow VNA to warm up 10–15 minutes; recalibrate after large temperature changes |
Frequently Asked Questions
Does the NanoVNA's built-in calibration save permanently?
The NanoVNA stores calibration in volatile memory by default — it is lost when the device powers off. To make it permanent (until recalibrated), after calibrating, go to SAVE and save to one of the memory slots (C0–C4). On next power-up, select RECALL and load the saved calibration. Note that a saved calibration is only valid for the same frequency range and the same cables as when it was performed. If you change cables or frequency range, you must recalibrate even if you have a saved calibration.
Why does the load need to be 50 Ω exactly?
The load standard establishes the zero reflection reference. If your "50 Ω load" is actually 52 Ω, the VNA's calibration will use 52 Ω as its reference, and all impedance measurements will be systematically off by a proportional amount. For amateur radio measurements, a 1–2% resistance error in the load is usually acceptable. Precision calibration kits specify the load resistance to within 0.1 Ω. A common trick for verifying a load is checking its S11 before calibration — it should show −35 dB or better if it is truly close to 50 Ω.
Can I use a BNC calibration kit on an SMA port?
Not directly — the calibration plane is defined by the physical connector of the calibration standard, and mixing connector types introduces adapter effects. You would need to either use a BNC-to-SMA adapter and include it in the calibration (calibrating at the BNC side), or purchase SMA calibration standards that match the NanoVNA's SMA ports. The NanoVNA V2 typically includes SMA calibration standards; older versions may include a mix. Using mismatched connector types is a common cause of poor calibration in amateur setups.
Test Your Knowledge
Answer the questions below to check your understanding. Every answer can be found in the lesson above.