What a control laboratory needs to teach so graduates can tune a real loop: the four process types, computer-controlled against manual rigs, calibration skills, and a purchase sequence that builds capability rather than a showroom.
A final-year student can derive a second order transfer function, sketch a root locus and state the Ziegler-Nichols rules from memory. Put them in front of a level loop that is oscillating and ask them to fix it, and they change the setpoint. That gap is the entire justification for a process control laboratory, and it is not closed by simulation software alone.
A control laboratory is unusual among teaching laboratories because the equipment is comparatively small and the pedagogy is comparatively hard. Getting it right is mostly about choosing processes whose dynamics differ from each other, and about deciding early how much of the loop the student is allowed to touch.
Write down what a student should be able to do unaided on the last day of the course. This list is the specification. Everything else is procurement detail.
Items three, four, nine and ten are the ones that separate a graduate an industrial employer wants from one who has only done coursework. They also happen to be the items that pure simulation cannot deliver.
Flow, level, pressure and temperature are not four examples of the same thing. They are four fundamentally different dynamic characters, and a student who has only tuned one of them has not learned control.
| Loop | Dynamic character | What it teaches | Common student error |
|---|---|---|---|
| Flow | Fast, self-regulating, noisy signal, very short dead time | Filtering, why high controller gain is possible and why derivative action is a mistake here | Adding derivative and amplifying the noise into valve chatter |
| Level | Integrating, no natural steady state, slow | Why a proportional-only controller leaves the level drifting, and where averaging level control is actually preferred to tight control | Tuning for tight control on a surge tank and fighting the process |
| Pressure | Fast in gas, very fast in liquid, strongly affected by volume | How process capacity changes the time constant, and the difference between compressible and incompressible behaviour | Assuming the tuning transfers between a large and a small vessel |
| Temperature | Slow, significant dead time, often non-linear | Why dead time limits achievable performance, and why integral action is necessary but slow | Increasing gain to speed it up and producing a sustained oscillation |
A single unit that presents all four on one frame, such as the EDIBON FLPTU flow, level, pressure and temperature control unit, is the efficient way to cover this ground, because the controller, the transmitters and the final control element are common and only the process changes. Students see directly that identical hardware behaves completely differently depending on what it is connected to.
Underneath that, a fundamentals unit such as the EDIBON RYC PID control and regulation unit earns its place by stripping the problem down: a controller, a simple process, and no distractions, so the effect of each term can be seen in isolation before anyone touches a plant.
This is the decision procurement committees ask about most often and the one with the least honest answer in most brochures. Both formats teach; they teach different things.
| Panel or manual controller | Computer-controlled unit | |
|---|---|---|
| What the student manipulates | A physical controller faceplate, a real valve, a real transmitter | Software blocks and setpoints on a screen |
| Strength | The loop is visible and physical; a student can put a hand on the valve and feel it move | Repeatable, exportable data, easy comparison of tuning runs, supports larger classes |
| Weakness | Data capture is manual and slow; comparing two tuning attempts is tedious | The rig can become a black box that students drive without understanding |
| Assessment | Suits viva and observed practical work | Suits report-based assessment and analysis of logged traces |
| Class throughput | Lower | Higher, and remote or shared access becomes possible |
| Maintenance | Fewer software dependencies, long service life | Depends on the supplier keeping software current with operating systems |
The workable answer for most Bangladeshi departments is a mixed laboratory: fundamentals taught on a manual or panel-controlled rig where the student can see the mechanism, and applied work done on computer-controlled units such as the EDIBON APC applications for process control range, with acquisition through EDAS/VIS so tuning runs can be logged, exported and compared in a report.
One question to put to any supplier of computer-controlled equipment before signing: what is the commitment on software compatibility over the life of the equipment, and can the rig still be operated if the supplied computer dies? A control rig that becomes unusable because a driver no longer installs is a common and avoidable waste.
Industrial control is rarely a single isolated loop, and a laboratory that stops there produces graduates who are surprised by their first plant.
Control is half the syllabus. The other half is measurement, and it is the half most laboratories under-equip because transmitters and calibrators look like accessories rather than teaching equipment.
A student should leave able to work with a 4 to 20 mA current loop as a matter of routine: identify two-wire and four-wire devices, calculate the loop resistance a supply can drive, understand why current is used rather than voltage over distance, and inject a signal to prove a loop from transmitter to display without the process running. Buy a loop calibrator. It is inexpensive relative to a rig and it converts a demonstration laboratory into a working one.
Sensor characterisation experiments are worth timetabling in their own right: the non-linearity of an orifice plate, the response time of a bare against a sheathed thermocouple, the effect of an air gap in a pressure impulse line, hysteresis in a control valve positioner. These produce the intuition that later makes a student say the loop is fine but the sensor is lying, which is very often the correct diagnosis.
Two Bangladesh-specific points to write into the specification. First, supply quality: control rigs contain switched-mode supplies and computers, and campus voltage fluctuation shortens their life. Put conditioning ahead of the laboratory distribution board and specify what happens to a running experiment when the supply drops. Second, service response: ask for a stated response time for a fault, and ask who in Bangladesh holds the diagnostic skills, because a control fault is rarely fixed by shipping a spare part blindly.
Any laboratory can show a student a step response. The laboratory that is worth the money is the one where the student can find out why the response is wrong.: Vvon Technologies commissioning notes