Step-by-step guide to digitizing and modeling vacuum tube characteristics
This tutorial will guide you through the complete process of creating a SPICE model for a vacuum tube triode from published characteristic curves. The workflow consists of two main phases:
The end result is a set of model parameters that can be used in SPICE circuit simulation.
First, you'll use the Characteristic Curve Digitizer to extract data points from a published triode characteristic curve image.
Find a triode plate characteristics graph showing Ia vs Vak curves at various Vgk values. Good sources include:
Save the graph as a PNG or JPG image file.
Open the Characteristic Curve Digitizer and click "Choose File" to load your image. Enter a project name (e.g., "300B" or "12AX7").
You need to define three calibration points to establish the coordinate system:
Enter the actual values for the X and Y calibration points in the input fields.
Each curve at a different Vgk value becomes a separate series. Click "Add Series" and enter the grid voltage (Vgk) for that curve. For triodes, Vgk is typically 0, -10, -20, -30V, etc.
With a series selected, click along the curve to add data points. Work from left to right along each curve, clicking to capture points at regular intervals or at key features.
Add a new series for each Vgk curve and digitize the points. The more curves you include, the better the model fit will be.
Click "Download CSV" to save your digitized data. The file will be named based on your project name (e.g., "300B.csv").
The exported CSV file has this format:
Each series header contains "Series N: Vgk" where Vgk is the grid voltage, followed by rows of "Vak,Ia" data pairs.
Now use the Triode Curve Fitter to find the optimal model parameters.
Click "Choose File" and select the CSV file you exported from the Characteristic Curve Digitizer. The data summary will show how many curves and points were loaded.
Choose between two triode models:
| Model | Equation | Best For |
|---|---|---|
| Child-Langmuir | Ia = K(Vgk + DaVak)1.5 | Simple approximation, 2 parameters |
| Koren | Complex multi-parameter model | High accuracy, better for simulation |
The default initial values work well for most tubes, but you can adjust them if you have prior knowledge of the tube's characteristics.
For the Koren model, set the amplification factor (μ) based on the tube's datasheet value if known.
Click "Run Curve Fit" to start the Nelder-Mead optimization. The progress bar shows the fitting progress, and the plot updates to show the fitted curves.
Check the results panel for:
If the fit is poor, try adjusting initial parameters or switching models.
The SPICE Model Output panel shows a ready-to-use subcircuit definition. Click "Copy to Clipboard" or "Download .txt" to save the model for use in your SPICE simulator.
| Parameter | Description | Typical Range |
|---|---|---|
| K | Perveance constant | 1e-7 to 1e-5 |
| Da | Inverse amplification factor (1/μ) | 0.01 to 0.5 |
The Koren model provides a more accurate representation, especially in the saturation region:
| Parameter | Description | Typical Range |
|---|---|---|
| μ (mu) | Amplification factor | 2 to 100 |
| X | Exponent (usually ~1.5) | 1.0 to 2.0 |
| kG1 | Grid coefficient | 100 to 3000 |
| kP | Plate coefficient | 10 to 1000 |
| kVB | Knee voltage parameter | 1 to 500 |