IMDHMIControls LibraryShop main store

Selection guide

Sizing a CompactLogix to Replace Four SLC 500 Processors

PLC programming Published 7 min read

Short answer

With all 35 I/O modules staying in the SLC chassis behind four 1747-AENTR adapters, the CompactLogix local-bus module limit never applies. Size the controller on CIP connection count and converted-code memory: the four adapters are four connections, a light load, so memory for the merged programs is the binding constraint. Start with a 1769-L30ER offline, verify the four-adapter configuration, and step up one size only if memory headroom falls short.

On this page
  1. Why the local module limit does not apply
  2. What each number actually limits
  3. The connection budget, with arithmetic
  4. One controller or four
  5. Prove the choice offline, step by step
  6. Acceptance checks before cutover
  7. The decision, case by case
  8. Questions asked

With all 35 I/O modules staying in the SLC chassis behind four 1747-AENTR adapters, the CompactLogix local-bus module limit never applies to this job. The controller sees four EtherNet/IP adapters, so two numbers size it: the CIP connection count and the memory the four converted programs need. Start with a 1769-L30ER as the test candidate, build the four-adapter configuration offline, add every other Ethernet device the cell will use, and let controller verification plus the converted project memory use make the selection. Whether one controller replaces all four processors at all is decided first, by the outage boundary, and on an interlocked continuous line that boundary says one.

Why the local module limit does not apply #

The module counts published against CompactLogix controllers describe modules on the local 1769 bus: I/O banked physically with the controller, sharing its power budget and expansion rules. A 1747-AENTR is not that. It is an EtherNet/IP adapter that owns the SLC chassis; the controller opens one I/O connection to the adapter and reads the whole chassis through it. The 35 modules never appear on the 1769 bus, so the published local figures answer a question this project is not asking.

That is why the knowledgebase reads as contradictory. Some figures describe local 1769 expansion, others describe remote adapter configurations; neither is wrong, they limit different things. Ordering the largest CompactLogix because 35 exceeds a published local count buys capacity this project will not consume, while the two checks that can actually fail, module support at the adapter and controller memory, go untested.

What each number actually limits #

Number you were comparing What it limits Applies to these four racks
Published local I/O module count for the controller Modules on the local 1769 bus No; all 35 modules stay in the SLC chassis
Supported-module table in publication 1747-UM076 Which 1746 modules each 1747-AENTR chassis may carry Yes; walk all 35 catalog numbers against it
Documented CIP connection limit of the controller Adapters, HMIs, MSG targets, programming sessions Yes; four adapters consume four connections
Controller user memory Converted programs, tags, messaging, diagnostics Yes; four merged programs set the size

The hardware gate is the supported-module table in publication 1747-UM076, the 1747-AENTR user manual. Check every retained module by exact catalog number against that list. The adapter promotes multiclass modules to the highest supported class automatically, and the vendor knowledgebase names the 1746-NI16I, 1746-NI16V, 1746-NR8 and 1746-NI8 among the modules handled that way. A module missing from the list is the only item in this inventory that forces a wiring change: replace it while the chassis is already open, not after commissioning.

The connection budget, with arithmetic #

Each 1747-AENTR consumes one CIP I/O connection. The modules behind it add none: the chassis is configured and updated as one unit over that single connection, at the I/O update rate set for the adapter.

Count the finished cell, not just the four racks. A worked example with a realistic mix for a line like this one:

  • 4 × 1747-AENTR adapters = 4 connections
  • 1 × HMI = 1 connection
  • 1 × programming or engineering session = 1 connection
  • 2 × MSG connections to a peer controller and a drive = 2 connections

Total: 4 + 1 + 1 + 2 = 8 connections.

Hold that total below the documented CIP connection limit for the candidate model, with room left for devices added later. Four adapters is a light connection load; what gets tight is the total once HMIs and messaging sit on the same controller, which is why the whole cell is counted. Do not settle the question with arithmetic alone, though: verification in the controller software enforces the documented limit and names the device that crosses it, which is exactly why the offline build is the selection test.

One controller or four #

Ask one question about each of the four processors: can that section of the line keep running while its processor is down for a download, a fault or a replacement? The answer sets the architecture.

Condition on the line Architecture Why
A section keeps producing with its own processor down One controller per section The independent maintenance window is worth keeping
Any processor fault already stops the whole interlocked line One controller for all four racks The outage boundary already is the line
One section is slated for replacement or relocation Keep that section on its own controller The future project must not force a migration of the other three
One controller is chosen for the whole line Write the download policy down first A full-line download becomes a plant decision, not a controls decision

Your line is continuous and interlocked, so the boundary already is the whole line: consolidate on one controller. Take the lifecycle cost with open eyes, because merging racks ties their futures together, the same trade that appears when consolidating CompactLogix remote racks to a ControlLogix CPU. Before hardware is ordered, record who authorizes a full-line download and how a future single-section replacement will be isolated from the shared controller.

Prove the choice offline, step by step #

  1. Create an offline project for the 1769-L30ER. It is a test candidate, not yet the selection.
  2. Add four 1747-AENTR adapters to the EtherNet/IP tree, one per chassis.
  3. Rebuild each physical chassis slot by slot and enter the exact catalog number of every retained module. Placeholders verify nothing.
  4. Add every other Ethernet device the cell will run: the HMI, the MSG targets, any drives. Testing four adapters alone understates the communication load.
  5. Set each adapter I/O update rate to the response the process needs, not the fastest rate the software accepts.
  6. Run configuration verification. Unsupported modules, invalid slot arrangements and connection overruns surface here, before any hardware is ordered or any wiring is touched.
  7. Convert the four programs. Publication 5069-AP001 documents the path: the SLC Migration Wizard, launched from Integrated Architecture Builder. A standalone controller swap is a different and simpler job, covered by migrating a MicroLogix 1500 machine to a current PLC; merging four programs into one is where this project earns its keep.
  8. Merge the converted programs, resolve the four address spaces into one tag database, and read the memory use. Require at least a third of user memory free after the merge, a planning margin for edits and added devices rather than a published limit. If less remains free, take the next controller size up and repeat verification.
  9. Step up one model at a time, so each move buys a named resource, memory or connections, deliberately.

Acceptance checks before cutover #

Bench-test one adapter with representative modules when hardware is available, then extend to the full configuration.

  • Exercise every input and output type on the line and compare raw counts, engineering units and output behavior against the old application.
  • Interrupt one adapter connection and confirm the program detects the lost rack, drives that section interlock permissives false, and reports which rack is gone. This is the check that matters when one controller owns an interlocked line, so never rely on an untested automatic restart. Restore the adapter and verify the recovery behavior you designed.
  • Watch connection, memory and task diagnostics under representative activity.
  • Run the remote I/O commissioning checklist for a modular station against each adapter before the field wiring is handed over.

The decision, case by case #

  • Architecture: one controller. The line is continuous and interlocked, so any processor fault already stops all four sections.
  • First candidate: the 1769-L30ER, with all four adapters and every other Ethernet device in the offline project.
  • Selection rule: verification passes and the merged programs leave at least a third of user memory free. If either fails, take the next size up.
  • Connection rule: count the whole cell, 8 connections in the worked example, against the documented limit; verification enforces it.
  • Module rule: all 35 catalog numbers on the 1747-UM076 supported list. Anything not listed is replaced during the migration, and that is the only wiring change.
  • Exception: keep four controllers only where a section can run with its own processor down.

Questions asked

Does each 1747-AENTR adapter count as one controller connection even with many modules behind it?

Yes. The adapter, not the module count, consumes the CIP I/O connection: four adapters are four connections whether each chassis carries four modules or fourteen. The chassis updates as one unit over that connection at the I/O update rate you set. Count HMIs, MSG targets and programming sessions separately, because they draw from the same documented connection limit.

Is the 1769-L30ER guaranteed to pass verification with four adapters?

It is the starting test candidate, not a guaranteed selection. Four adapters are a light connection load, so the usual failure is elsewhere: a module missing from the supported list in publication 1747-UM076, or four merged programs leaving too little user memory. Verification plus the converted project memory use makes the selection; if either fails, step up one controller size and repeat.

Can the existing 1746 modules and field wiring stay in place?

Wiring stays if every one of the 35 modules appears in the supported-module table in publication 1747-UM076, the 1747-AENTR user manual. The adapter promotes multiclass modules such as the 1746-NI16I, 1746-NI16V, 1746-NR8 and 1746-NI8 to their highest supported class automatically. A module absent from the list gets replaced during the migration, and that replacement is the only part of this project that touches field wiring.

How do the four SLC programs become one project?

Run each of the four programs through the vendor's conversion tooling as covered in the step-by-step section below, then merge the results into one project and resolve the four address spaces into a single tag database. The tooling moves the logic; the merge and the tag correlation remain your work, and the merged result is what the memory estimate is taken from.

How much memory headroom is enough after the merge?

Plan on at least a third of user memory free once the four converted programs, tags and messaging sit in the controller. That margin is a planning rule covering edits and added devices, not a published limit. If the merged project leaves less, take the next controller size up and rerun verification, because memory is the resource that shrinks when four programs share one processor.

Keep reading

  • Selection guide

    Migrating a MicroLogix 1500 machine to a current PLC

    Quote CompactLogix as the baseline for the rebuild: Rockwell's migration reference gives the MicroLogix 1500 two documented destinations, MicroLogix 1400 or CompactLogix 5370 (L1/L2), and a low-volume machine rebuild does not justify the Micro850 route. Take the Micro850 only for high build volume, after the 11 uncovered inputs, serial Modbus, the drive interface and a trial CCW import are verified.

    PLC programming

  • How-to

    Port a PLC-5 filter routine to ControlLogix structured text

    The filter is a first-order lag whose only state is the previous filtered value. Give every filtered signal its own instance of one filter UDT that also holds the exponent and coefficient, translate the rungs to ordered assignments in a periodic task, and preload the state on the first pass. Prove the port with a step: 63.2% of the change after one time constant, about 95% after three.

    PLC programming

  • Selection guide

    Add 4-20 mA analog inputs when the CompactLogix rack is full

    Swap the installed four-channel analog input card for an eight-channel one when channel isolation matches; that adds four inputs against a need of three, with no rack, power-supply or enclosure change. If isolation rules the swap out, reposition the existing power supply to free an end slot. When the cabinet itself is full, add a POINT I/O analog module or IO-Link converters over EtherNet/IP.

    Controllers and I/O