How to Size a PLC for High-Speed Data Logging Applications

Defining the Data Logging Scope

High-speed data logging differs from standard process control. While a standard control loop might sample a temperature at one hertz, a high-speed logging application may capture servo position, motor current, and pressure at thousands of hertz. The PLC must process this data without missing cycles. This changes the sizing calculation significantly. The controller must not only read inputs but also buffer data, timestamp events, and transmit it to a historian or edge server. The first step is to define exactly what data is being captured and how fast it must be captured. Engineers must list every analog and digital input required. They must also list every output, including control valves, relays, and status indicators. The output list is often shorter than the input list in pure logging applications. However, any feedback loop that closes within the scan time must be treated as a critical control path. If the system simply records data without controlling anything, the I/O load is almost entirely on the input side.

Calculating I/O Point Density

Start with a complete list of all signals. Analog inputs require dedicated channels. Each channel typically supports one signal. Digital inputs can be grouped on a single card. The grouping depends on the card type. Some cards handle 32 points per slot. Others handle 64 or 128 points. The physical slot count limits the total number of points. A 24-slot chassis might support 1,000 to 1,500 digital points, depending on the card mix. Analog channels are more expensive and take up more slots. An 8-channel analog card might take one slot. A 16-channel card might take two slots. The physical footprint is a hard constraint. If the control panel is small, you cannot fit more cards. The I/O density calculation is a physical exercise. Count the points. Divide by the maximum points per slot. Add a 10 to 20 percent spare. This spare capacity is not optional. It protects against future changes. A new sensor or a diagnostic point added during commissioning can break a full chassis.

Determining Scan Time Requirements

Scan time is the time it takes the PLC to read all inputs, run the logic, and write outputs. For high-speed logging, scan time is the bottleneck. If the scan time is 10 milliseconds, you can log at a rate of 100 samples per second. If the scan time is 1 millisecond, you can log at 1,000 samples per second. The requirement is not just to log. It is to log fast enough that the data represents the physical process. If a vibration event lasts for 5 milliseconds and your scan time is 20 milliseconds, you miss the peak. The scan time must be shorter than the fastest event you need to capture. Engineers often underestimate this. They size for 10 millisecond scans based on standard control practice. High-speed logging often requires 1 millisecond or less. This drives the choice of processor speed. A slower processor cannot execute the logic fast enough. It will miss the scan deadline. The scan time calculation must consider the total number of points. More points mean more time to read and write. The processor must handle the I/O cycle plus any logic.

Assessing Processing Power and Memory

The processor must handle the I/O cycle and the data storage. High-speed logging generates large amounts of data. A single analog channel sampled at 1,000 times per second generates 1,000 data points per second. With 10 analog channels, that is 10,000 points per second. Over one hour, that is 36 million points. The PLC must store this data in memory or write it to a buffer before the next cycle. If the memory fills up, data is lost. The processor must write to a buffer at the same rate it reads from the sensors. This requires fast memory and a fast communication link. The internal memory of the PLC must be large enough to hold the buffer. External memory or a solid-state drive may be required. The processor speed must be high enough to manage this I/O. A standard 100 millisecond scan processor will fail. A high-speed processor with a fast bus is required. The communication interface is also critical. The PLC must send data to a historian or edge server. This traffic adds to the load. The Ethernet port must have enough bandwidth to handle the raw data stream.

Selecting Communication and Data Transfer

Data logging requires moving data from the PLC to a historian or a local edge server. The communication protocol matters. Ethernet/IP or Profinet are common choices. The bandwidth must handle the data rate. A high-speed logging application might generate 100 megabytes per second. A standard 100 megabit Ethernet port is insufficient. A gigabit port is required. The network switch must also be capable. The historian database must be able to ingest this data rate. The PLC is only the first link in the chain. If the network is slow, the PLC buffer will fill up. The PLC must have a mechanism to handle buffer overflow. Some controllers can drop data. Others can extend the buffer. The data transfer method must be designed with the PLC sizing. The communication card must support the required data rate. A standard I/O card is not enough. A high-speed communication card is required. The total system load includes the I/O, the logic, and the communication. All three must fit within the scan time.

Common Sizing Mistakes

The most common mistake is sizing based only on the number of I/O points. Engineers count the points and buy a chassis with enough slots. They ignore the scan time. A chassis with 1,000 points and a slow processor will fail a high-speed logging task. The processor cannot keep up. The second mistake is underestimating the data volume. Engineers calculate the sample rate but forget the storage requirement. The PLC buffer fills up in minutes. The third mistake is ignoring communication load. The PLC reads the data, but the network cannot carry it. The fourth mistake is not leaving spare capacity. A full chassis is fragile. A new point added during commissioning causes a redesign. The fifth mistake is assuming standard I/O cards work. High-speed logging often requires dedicated high-speed I/O cards. These cards have different bus requirements. They may not fit in a standard chassis. The sizing must consider the card type, not just the point count.

Final Verification Method

Verification is the last step. Do not rely on the initial calculation. Run a simulation. Load the actual logic into the PLC. Connect the I/O. Run the system at full speed. Monitor the scan time. It should remain stable. If the scan time drifts, the processor is under-sized. Check the data buffer. It should not fill up. Check the communication link. The bandwidth should be sufficient. The verification step catches errors that the calculation missed. It confirms that the hardware works together. It also tests the worst-case scenario. If the system passes verification, it is ready. If it fails, the sizing is wrong. Adjust the processor speed, add memory, or improve the network. Do not proceed to commissioning until the scan time is stable and the data is complete. Verification is the only true test of PLC sizing.

Quick Reference Table

Parameter	Typical Range for High-Speed Logging	Impact on Sizing
Scan Time	0.5 to 10 milliseconds	Determines processor speed
I/O Density	500 to 2,000 points	Determines chassis size
Sample Rate	100 to 1,000 Hz	Determines data buffer size
Data Volume	10 to 100 MB per second	Determines communication bandwidth
Memory	128 MB to 1 GB	Determines storage capacity

Step-by-Step Methodology Summary

1. List all analog and digital I/O points.
2. Calculate the total number of points including a 20 percent spare.
3. Determine the required scan time based on the fastest process event.
4. Select a processor that can execute the logic within the scan time.
5. Calculate the data volume generated per second.
6. Select a communication interface with sufficient bandwidth.
7. Verify the configuration with a full simulation before commissioning.

The sizing of a PLC for high-speed data logging is a systems problem. It is not just about counting points. It is about managing data flow. The processor, the I/O, the memory, and the network must all work together. If one part is weak, the whole system fails. Use the step-by-step method above to ensure the controller is correctly sized. The result will be a reliable data logging system that captures every important event.