When using master-slave control, the master controller takes the signal from its own thermocouple and delivers power to its heater in the normal way until the temperature comes to setpoint.  The slave controller takes the signal representing the temperature difference from the two thermocouples connected back-to-back.



Suppose you have a press with top- and bottom-heated platens, each with its own temperature controller. You want to be able to set the temperature on one controller -- say the top -- as the master, and have the bottom, or slave, adopt the same setting.

Method 1: Back-to-Back Thermocouples

To achieve master-slave control, make the connections shown in figure 1. With this setup, the master controller takes the signal from its own thermocouple and delivers power to its heater in the normal way until the temperature comes to setpoint.

The slave controller takes the signal representing the temperature difference from the two thermocouples connected back-to-back. If you set its setpoint to 0oF, the slave will controller decrease its output power for upscale readings, increase output for downscale readings and will not be satisfied until it achieves its setpoint of 0oF. This means that the two thermocouples and, therefore, the two platens, end up at the same temperature. Note that the slave controller must be configured to have no cold-junction compensation and to indicate in a range of say -100 to 100oF. Most modern controllers can be reconfigured in this way.

The circuit in figure 1 uses the North American color code for the Type J (iron/constantan) thermocouple, where red denotes the negative wire. Don't be misled. Although an internationally agreed upon color code has been out for a few years, its adoption is going slowly in North America.



Figure 1. This is one way to achieve master-slave control of two heated platens.

How This Method Works

You have to study the thermocouple polarity logic to understand how the slave circuit does its job and why two positive wires go into the positive and negative terminals of the slave controller.

Unless you specify otherwise, your thermocouples may come with their hot junctions welded to the metal sheath. When they are installed, the welded hot junctions become grounded and connected to each other. A grounded junction does no harm when only one control loop is involved, but with back-to-back connections, the difference signal becomes totally unpredictable and inaccurate. So, make sure that you specify isolated hot junctions and avoid grounding any part of the thermocouple circuits.

You can see from the circuit that the master thermocouple output acts as the setpoint for the slave controller. At cold startup, the master setpoint, being inside the controller, is at the target temperature right from power switch-on, and the master controller will be delivering full power. The slave controller must wait for the master's thermocouple signal to climb and bring a negative deviation to the slave controller so it can start delivering power. This makes the slave control's temperature lag behind that of the master zone on startup. You will see the same kind of lag following the master after setpoint changes and process disturbances. Keep this lag in mind when making control parameter (PID) adjustments on the slave controller.

There is no theoretical limit to the number of slave zones you can connect to one master controller. The combined input resistance of many slave controllers, if in the megohm region, could affect thermocouple signal accuracy. Also, the ground leakage of many thermocouples and wiring in parallel (again, many megohms) could degrade accuracy. In the absence of specific figures, I would put the limit at fewer than ten zones.

In the next issue, I will look at transmission of the master setpoint as a high-level analog signal and alternatively as a digitally coded signal.



Links