English | 简体中文
Current target: .NET Framework 4.6.1 (net461). The application and
regression runner do not require .NET 8 to run.
PID_CSharp is a C# WinForms simulator for water-tank level, oven temperature and DC motor speed control. Change the setpoint and PID tuning, apply disturbances, and watch the process and controller output respond. The interface is available in English and Chinese.
The simulator is intended for learning PID control, practicing tuning and comparing the responses of different processes.
Everyone is welcome to contribute code.
SP is the setpoint, PV is the measured process value, and MV is the controller output. The controller adjusts MV based on the difference between SP and PV.
| Scenario | Setpoint and feedback | Controller output | Disturbances |
|---|---|---|---|
| Water tank | Target and actual water volume, L | Outlet-valve opening, 0–100% | Inlet-valve changes, unstable water supply |
| Oven | Target and measured temperature, °C | Applied heater output, 0–100% | Opening or closing the door, adding a cold workpiece |
| Motor | Target and actual shaft speed, rpm | Drive command, 0–100% | Applying or removing a 1.20 N·m load |
The tank controls its level by adjusting discharge: a high level calls for more outlet-valve opening. In the oven and motor scenes, increasing the output raises temperature or speed.
The tank has a capacity of 3,100 L. Its level display uses water volume in liters. The inlet valve is adjusted manually; in automatic mode, the PID controller adjusts the outlet valve. Change the inlet flow to observe the valve response and the return toward the setpoint. Switch to manual mode to adjust the outlet valve yourself and compare the two approaches.
The oven model includes the thermal inertia of the heater, chamber and workpieces. Once the temperature has settled, add a cold workpiece or leave the door open briefly before closing it. Watch how the heater output changes as the temperature falls and recovers.
Opening the door cuts heater power immediately and increases heat loss through air exchange. While the heater interlock is active, the controller tracks the actual zero output. Closing the door releases the interlock and resumes automatic control. The power readout and output trend show the applied heater output.
The 48 V DC motor model simulates armature current and electromagnetic torque, with rotor inertia and a current limit. Applying load initially reduces the speed, followed by an increase in drive output and current. Removing the load initially raises the speed; the controller then brings it back toward the setpoint.
The drive is unipolar. With a target of 0 rpm, the motor coasts down under friction, without reverse braking current.
Start with the default settings:
| Scenario | Mode | Setpoint | Tuning |
|---|---|---|---|
| Water tank | PI | 1500 L | GAIN=12, RESET=6 s, RATE=0 s |
| Oven | PID | 180 °C | Kp=2.4 %/°C, Ti=20 s, Td=8 s |
| Motor | PI | 1800 rpm | Kp=0.1 %/rpm, Ti=0.3 s |
The tank uses GAIN, RESET and RATE for its tuning parameters. GAIN/100 is the proportional gain in percentage points of valve opening per liter of error. RESET is the integral time Ti, and RATE is the derivative time Td.
The motor interface retains Td=0.03 s. It is inactive in PI mode and becomes active when PID mode is selected.
The oven and motor defaults produce some overshoot, so you can observe the initial rise, the peak and the return to the setpoint. When comparing P, PI and PID in these two scenes, click Reset after changing the mode to start from the same initial conditions. Once the process settles, change the setpoint or apply a disturbance. Adding derivative action does not guarantee a better response with the existing tuning; retune and compare.
The tank runs at 10× simulation speed, the oven at 8× and the motor at 1×. Ti, Td and trend times use simulated seconds. The Instructions dialog provides operating steps and suggested experiments for all three scenes.
All three scenes use the algorithm in
PID_helper.cs. Each has its own controller
instance and simulation state. Scene code handles units, control direction
and process models.
The controller supports P, PI and PID, with output limits, independent integral limits, optional measurement filtering, filtered derivative on the measurement, manual output tracking and online retuning. Anti-windup options are conditional integration (the default), back-calculation and disabled output-saturation feedback. Advanced options are configured through the helper API; the interface has no anti-windup strategy selector.
The anti-windup design follows concepts from
ROS control_toolbox.
This is an independent C# implementation with no ROS dependency. See
PID_CONTROL.md for equations, parameter conventions
and limitations.
The main form, frmMain, hosts three tabs. Each scene is a separate
UserControl: usWaterTank, usOven and usMotor. Page controls are declared
in their .Designer.cs files, so layouts can be edited in the Visual Studio
Designer.
The migration preserves the PID equations, parameter units, default tuning,
anti-windup strategies, manual/automatic transfer, oven door interlock and
process-model equations. Core source files have compatibility edits for APIs
that are unavailable in .NET Framework 4.6.1; the control algorithm was not
redesigned. The oven's temporary Span buffer is now a reusable array, and
explicit font and DPI settings preserve the existing page layout.
Validation on 2026-11-15:
- Debug and Release builds passed against the .NET Framework 4.6.1 reference assemblies.
- All 76 existing regression checks and 27 UI/target-framework checks passed.
- Runtime checks used .NET Framework 4.8 on the development machine. A machine with only .NET Framework 4.6.1 installed has not been tested.
Development requirements:
- Windows 10 or Windows 11.
- Visual Studio 2022 17.14 or later with the .NET desktop development workload installed.
- .NET Framework 4.6.1 Developer Pack / Targeting Pack for compilation.
- .NET SDK 9.0.315, pinned in
global.json, for the SDK-style projects and C# 12 compiler only. - .NET Framework 4.6.1 or a compatible later 4.x runtime to run the application.
Both the application and regression runner target net461. Running the
application requires neither .NET 8 nor a .NET SDK. In Visual Studio's
Developer PowerShell, run these commands from the repository root:
msbuild .\pid_loop_CSharp.sln /restore /p:Configuration=Release
.\pid_loop_CSharp\bin\Release\net461\pid_loop_CSharp.exeYou can also open pid_loop_CSharp.sln in Visual Studio, set the main project
as the startup project and run it.
Use the MSBuild supplied with Visual Studio for the application's WinForms
icon resources. A clean dotnet build cannot process these non-string resources.
Deploy the complete output directory, including pid_loop_CSharp.exe.config,
System.ValueTuple.dll, the en satellite-resource directory, pid_loop.cfg,
pid_sampling.cfg and the PID diagram under Assets. System.ValueTuple is
Microsoft's compatibility package that preserves the existing tuple code on
.NET Framework 4.6.1. The sampling settings in pid_sampling.cfg apply to the
tank scene. The real I/O period setting is reserved for future integration
and does not enable hardware communication.
The regression project is a standalone console runner that uses the shared PID and process-model source files directly. It currently contains 76 checks covering all three scenes, saturation recovery, filtering, output tracking, parameter validation and the oven door interlock.
Run it from the repository root:
msbuild .\tests\PidLoop.RegressionTests\PidLoop.RegressionTests.csproj /restore /p:Configuration=Release
.\tests\PidLoop.RegressionTests\bin\Release\net461\PidLoop.RegressionTests.exeThis project is not included in the main solution and must be run separately.
The process models and default settings are intended for teaching and simulation experiments. The oven uses effective heat capacities and a fixed airflow when the door is open. The motor model is also simplified. The curves should not be treated as predictions for a particular machine.
The project has no hardware I/O integration and has not been validated on real equipment. Using the controller with real equipment requires tuning for the actual sampling period, sensors and actuators. You will also need I/O drivers, data quality checks, watchdogs, actuator feedback and independent equipment protection.
The PID algorithm is encapsulated in a reusable controller; the author accepts no liability for any issues arising from its use in engineering projects.
The three simulation scenes are shown below. The water-tank screenshot is from an earlier version.
Packaged builds are available in appDownload/. Build from
source for the version described above.


