PLC CPU Explained: How It Works, How to Choose, and How to Replace It

Most people searching for a PLC CPU are not students. They are in the middle of a job: choosing a controller for a new line, replacing a unit that just faulted, comparing two brands before a purchase order, or hunting for a part the OEM stopped making. The spec sheet in front of you lists clock speed and memory in megabytes, but those numbers rarely decide whether the project runs. What decides it is scan time, protocol compatibility, and whether you can actually get the part when you need it.

This guide walks the full chain a buyer or engineer goes through: understanding what the CPU does, choosing the right one across six major brands, replacing or upgrading an existing unit, troubleshooting common faults, and sourcing the part. Skip to the section that matches where you are.

1. What a PLC CPU Actually Does

The CPU is the part of a programmable logic controller that reads inputs, runs your program, and drives outputs, over and over, on a fixed cycle. Everything else in the rack (the power supply, I/O modules, and communication cards) exists to serve that loop. Before we get to selection and replacement, here are the three things worth knowing about how it behaves.

CPU vs Processor

People use "processor" and "CPU" interchangeably, and in casual conversation that is fine. The difference matters when you buy. The processor is the chip that executes instructions. The CPU is the complete control unit that houses that chip along with memory, communication ports, and the firmware that ties them together. When you order a replacement, you order a CPU module by its part number, not a bare processor chip. That distinction is also what governs the timing determinism of your line, and it is what you quote a supplier against.

The Scan Cycle in 4 Steps

A PLC CPU does not run your logic once. It runs continuously, in a repeating loop called the scan cycle. Each pass follows the same order:

1. Input scan. The CPU reads the state of every input and stores it in memory.
2. Logic solutions. It executes your program (ladder logic or another IEC 61131 language) using those stored input values.
3. Output update. It writes the results to the output modules.
4. Communication and diagnostics. It handles network traffic, updates status, and runs internal self-checks before starting again.

The cycle is deterministic, which is the whole point of a PLC: the same inputs produce the same outputs in the same predictable time. Inputs and outputs are only read and written at fixed points in the loop, not the instant a sensor changes. The time for one full pass is the scan time, and that is why scan time, not clock speed, decides real performance. We come back to that in the specifications section.

Picture a line cook who reads every ticket, cooks every dish, plates them, and then checks the pass before turning back to the tickets. What matters is how long one full round takes, not how fast the hands move on any single motion.

CPU Modes: Run, Program, and Remote

A CPU operates in one of a few modes, and knowing them prevents mistakes during maintenance:

• Run executes the program and controls the process normally.
• Program stops program execution so you can download or edit logic. Outputs are no longer being driven, which is exactly why you switch to it before servicing.
• Remote lets you change modes over the network instead of at a physical keyswitch, which is useful on distributed or hard-to-reach installations.

The point to remember: moving a live CPU to program mode stops control output. During a CPU swap you will deliberately stop the processor, which is one reason replacement is a planned task rather than a hot job.

2. Inside the CPU: Memory, Communication, and Diagnostics

Three internal systems drive most selection and troubleshooting decisions: how the CPU stores data, how it talks to the rest of the system, and how it reports its own problems.

Memory Types

A CPU uses more than one kind of memory, and each behaves differently when the power goes off:

• RAM is volatile working memory. It holds live values while the CPU runs and is cleared on power loss.
• Flash is non-volatile storage that keeps your program through a power cycle, so the controller boots back up with the logic intact.
• Retentive and battery-backed memory preserves selected values (counters, recipes, accumulated totals) across power cycles, so the process resumes where it stopped rather than from zero.

The failure mode worth flagging: retentive memory on many older platforms depends on a backup battery. When that battery dies, the controller can lose retained data and, on some units, the program itself. A common and costly mistake is assuming the program is safe while the backup battery has quietly been dead for months. This is why a spare battery belongs in your maintenance kit and why "Does it use a backup battery?" is a fair question to ask when sourcing a replacement. It also connects directly to the battery-loss symptom in the troubleshooting table further down.

Backplane and I/O Communication

The CPU connects to I/O modules through a backplane, the rack that physically and electrically links the modules in a chassis. Beyond the rack, it talks to drives, HMIs, other controllers, and SCADA over industrial networks. The protocols you will see most often are EtherNet/IP, PROFINET, Modbus, and EtherCAT, and they tend to cluster by brand ecosystem (EtherNet/IP with Allen-Bradley and PROFINET with Siemens, for example).

The reason this matters early: a CPU that does not speak your network's protocol is a non-starter, no matter how good the rest of its specs look. Protocol fit is a hard constraint on both selection and replacement. The brand comparison in Section 4 maps the main protocols to each platform.

Diagnostics, Watchdog, and Fault Handling

A CPU checks itself continuously while it runs. The watchdog timer is the core mechanism: if a scan takes longer than a set limit, or the processor hangs, the watchdog trips and forces a fault state instead of letting the machine run blind. Faults are surfaced through status LEDs on the front of the module, diagnostic buffers or logs you read with the programming software, and fault codes.

For now, the takeaway is simply that the CPU tells you when something is wrong and roughly what kind of problem it is. We turn those signals into a symptom-and-fix quick reference in Section 6.

3. PLC CPU Specifications That Actually Matter

Understanding the parts is one thing. Knowing which numbers to weigh when you choose is another. A handful of specifications carry the decision.

Why Cycle Time Matters More Than GHz

Clock speed sells consumer processors, but on a PLC it means little on its own. What you care about is scan time: how long the CPU takes to complete one full pass of input, logic, output, and communication. A faster clock can help, but program size, I/O count, and network load often matter more.

Here is the relationship, using illustrative numbers (treat these as an example, not a benchmark for any specific model):

Rule of thumb: your worst-case scan time should sit comfortably below the response time your fastest loop demands. Size the CPU to the scan budget, not the headline frequency.

Memory, I/O Count, and Protocol Compatibility

Three factors carry most selection decisions:

• Memory. Match program and data memory to the size of your logic, with room for future edits. Running near the ceiling turns every change into a fight.
• I/O count. Count the physical and networked I/O the application needs, then add roughly 20% headroom for expansion. It is cheaper to buy capacity now than to re-architect later.
• Protocol compatibility. The CPU has to speak the protocol your network and field devices use. This is the constraint that overrides the others.

With those criteria set, here is how the major brands line up.

4. How to Choose a PLC CPU: Brand by Brand

There is no single best PLC CPU, only the right fit for your application, your existing network, and the software your team already knows. Here is where each major platform tends to sit.

Allen-Bradley (CompactLogix / ControlLogix). Programmed in Studio 5000 with native EtherNet/IP. CompactLogix suits compact and mid-range machine control; ControlLogix scales to large, high-availability process and plant systems. Common on North American-spec lines, automotive, and food and beverage. Browse Allen-Bradley PLC modules.

Siemens (S7-1200 / S7-1500). Programmed in TIA Portal over PROFINET. The S7-1200 fits compact and OEM machines; the S7-1500 targets high-performance lines and process applications. Close to a default on European-spec equipment. Browse Siemens PLC modules.

Mitsubishi (MELSEC FX / Q). Programmed in GX Works. The FX series is compact; the Q series is modular for mid and high-end builds. Strong across Asia in packaging, material handling, and machine tools. Browse Mitsubishi PLC modules.

Omron (CJ2 / CP1). Programmed in Sysmac Studio or CX-One. CP1 is compact; CJ2 is modular. Frequently chosen for packaging, electronics, and semiconductor handling. Browse Omron PLC modules.

Schneider (Modicon M340 / M580). Programmed in EcoStruxure Control Expert. The M340 covers mid-range needs; the M580 is a high-availability ePAC with native Ethernet. Common in process, water and wastewater, and infrastructure. Browse Schneider PLC modules.

ABB (AC500). Programmed in Automation Builder and scalable from compact to high-end. Found in energy, marine, and process industries. Browse ABB PLC modules.

Selection checklist

• Protocol matches your network and field devices
• Program and data memory cover your logic plus headroom
• I/O count covers current needs plus about 20% expansion
• Worst-case scan time sits below your fastest loop requirement
• Lead time and warranty fit your project schedule

Not sure which model fits? Tell us your application, network, and I/O count, and we will match a CPU and quote it. Request a Fast Quote.

Free consultation

5. Replacing or Upgrading a PLC CPU

Choosing a CPU for a new build is one task. Swapping one into an existing system is a different one, with its own traps. This is where most buyer questions come from.

Cross-Compatibility and Part Numbers

The first rule of replacement: confirm the part by its full part number, not by series alone. The part number is printed on the module's label, and a single suffix can change firmware, memory, or supported protocol. Two units that look identical and share most of their part number can still behave differently.

To confirm a valid replacement, match the series, I/O capability, communication protocol, and firmware against the unit you are replacing. Same-brand, same-series replacements are usually straightforward when those four line up. A cross-brand swap is generally not a drop-in (the full answer is in the FAQ below). A common mistake is ordering by series and missing the suffix that quietly changes the firmware baseline. If you are not sure what you have, our Model Library and Advanced Search let you look a part up by number.

Firmware and Version Considerations

Firmware is the part of replacement that catches people off guard. A replacement CPU with the correct part number can still refuse to load your project if its firmware revision does not match what the project expects. The programming software is tied to this too: a project built in one version of Studio 5000 or TIA Portal may need a matching or compatible firmware on the CPU to download cleanly.

Before you swap, note the firmware revision of the existing CPU and the version of the programming tool the project was built in. If the replacement ships with different firmware, plan to upgrade or downgrade it to match, or specify the firmware you need when you order. That firmware revision is one of the details a supplier needs to quote and ship the right unit the first time.

Discontinued CPUs and Migration Paths

PLC platforms have lifecycles. At some point a CPU moves to limited availability, then to obsolete. When yours is discontinued, you have two realistic paths.

The first is to keep the existing system running by sourcing remaining new stock, surplus, or tested refurbished units. This avoids re-engineering and is often the fastest way to get a line back up, which is exactly where a multi-brand stockist helps. The second is to migrate to the successor series, which buys long-term support and new features but costs engineering time for rewiring, program conversion, and testing.

The right choice depends on how long the line needs to run and how much downtime you can absorb. (Check the OEM's lifecycle status for your specific part rather than relying on a general date.) If you need a discontinued or hard-to-find CPU, sourcing it is covered in Section 7.

What to send a supplier for a fast quote

• Brand
• Full part number, including any suffix
• Series
• Firmware version, if known
• Quantity
• Protocol or application
• Condition preference: new, surplus, or refurbished
• Target lead time

Have your part number ready? Send it and get a replacement quote fast.

Request a Quote.

6. Troubleshooting Common PLC CPU Faults

When a CPU faults, the goal is to narrow the cause quickly and decide whether it is a configuration issue, a wiring issue, or a failed module. Always de-energize and follow your site's lockout procedure before handling hardware, and leave live electrical work to qualified personnel. Use the table as a first-step guide, not a repair manual.

If diagnostics point to a failed CPU rather than a configuration or wiring issue, the next decision is how to source the replacement. Need a replacement part? Send your part number for a quote.

7. New, Surplus, or Refurbished: How to Source Your CPU

Once you know which part you need, the last decision is condition.

New vs Surplus vs Refurbished

Each option trades off differently:

• New gives full manufacturer warranty and the latest firmware, at the highest cost and sometimes the longest lead time.
• Surplus (new old stock) can be the only way to get a discontinued part, often at lower cost, though availability varies.
• Refurbished units that have been tested and carry a warranty are a sound choice when budget or lead time matters and the part is no longer made.

Choose new when warranty and the latest revision are non-negotiable, or when the part is current and readily available. Choose surplus or tested refurbished when the part is discontinued, when lead time is critical, or when cost is the deciding factor. What matters in every case is that the unit is tested and warranted, whatever its condition.

How to Make Sure You Buy the Right Part

Before you commit, verify the replacement against the unit you are removing:

• Confirm the full part number, series, firmware, and protocol match
• Ask for test and warranty documentation
• Confirm stock status and lead time before you order

Doing this from the buyer's side, rather than assuming the seller already checked, is what keeps a wrong unit from arriving on a shutdown deadline. You can confirm what is on hand through our in-stock PLC modules and Advanced Search.

FAQ

 

 

What information do you need to quote a PLC CPU module?

Send the brand, full part number (including any suffix), series, firmware version if known, quantity, protocol or application, condition preference, and target lead time. With those, we can confirm stock and quote quickly. Request a Quote.

Should I buy a new or refurbished PLC CPU?

Buy new when you need full warranty and the latest firmware, or when the part is current. Choose tested, warranted refurbished or surplus when the part is discontinued, lead time is tight, or cost matters. See Section 7.

How much memory does a PLC CPU need?

Enough to hold your program and data with room to grow. A practical rule is to size to your current logic plus about 20% headroom, so future edits do not crowd the ceiling. Program size, not a fixed number, drives the requirement.

What does the CPU fault light mean and how do I fix it?

A lit fault LED signals a program fault, configuration mismatch, or hardware failure. Read the fault code in your programming software's diagnostic buffer to identify the cause. The troubleshooting table in Section 6 lists first steps by symptom.

Can I replace a PLC CPU with a different brand or model?

Within the same brand and series, yes, as long as part number, firmware, and protocol match. Across brands, generally no. A different brand uses different programming software, wiring, and protocols, so it means rewriting the program and reworking the installation, not a drop-in swap.

What is the difference between a PLC CPU and a processor?

The processor is the chip that executes instructions; the CPU is the complete module that houses the processor along with memory, ports, and firmware. You buy and replace the CPU module by part number, not the bare chip. See Section 1 for detail.

9. Source Your PLC CPU Module

CHENTUO (Shenzhen Chentuo Technology) stocks PLC CPU modules across Allen-Bradley, Siemens, Mitsubishi, Omron, Schneider, and ABB, with new, surplus, and tested refurbished options shipped to customers in more than 30 countries. Whether you are choosing a CPU for a new line, replacing a faulted unit, or chasing a discontinued part, send us the part number and we will confirm stock, condition, and lead time.

Request a Fast Quote or send your part number, and you get a quote on the right PLC CPU module, in stock and ready to ship.

Free consultation