How to Reduce Output Points in PLC Control System Design (2026 Update)
Share
How-to · PLC Output Wiring · 2026 Update
By the KOEED Engineering Desk · 2026-06-30 · 9 min read · How-to
Output-point shortages still trip up PLC retrofits in 2026. A typical compact CPU ships with 16 to 32 transistor outputs, and a single machine can need 80 to 120. The fix is rarely "buy a bigger PLC" — it is almost always a smarter wiring strategy. This how-to walks through six field-proven methods to reduce output points in a PLC control-system design, with concrete examples on Allen-Bradley, Siemens, Mitsubishi, and Omron parts we keep in stock for these exact scenarios.
AI Summary — Key Takeaways
- Count honest outputs first — mutually exclusive loads collapse the count by 30% before any wiring trick.
- Group control (one output drives many loads via contactors) is the single biggest 2026 saving on motor-heavy panels.
- Diode matrixing and relay multiplexing remain valid for indicator stacks and BCD displays; high-density modules make them optional.
- Across 50+ discrete outputs, remote I/O (POINT I/O, ET 200SP, CC-Link IE TSN) replaces home-run wiring and bends the panel-build labour curve.
A 2026 procedural guide to cutting PLC output-point count using group control, multiplexing, diode matrixing, BCD coding, high-density modules, and remote I/O — with cross-references to Allen-Bradley, Siemens, Mitsubishi, and Omron parts KOEED keeps in active stock.
What changed since 2024
Three things changed in the PLC output landscape between the original 2024 article and this 2026 rewrite:
- CC-Link IE TSN is now mainstream. Mitsubishi, ABB, and several European vendors ship TSN-capable remote I/O at near-EtherNet/IP pricing. This changes the calculus for any panel over 50 outputs — TSN replaces home-run shielded cable with a single Cat 6A drop, and the per-point installed cost dropped roughly 40% between 2024 and 2026.
- High-density modules are denser. A 64-point module (e.g. Mitsubishi QY42P or Siemens 6ES7 322-1BL00) is now the default for tight panels, not the exception. The per-point cost is roughly half of what it was in 2020.
- IEC 61131-3 4th edition (2025) clarified function-block output semantics. "Once-cyclic" output coils and "qualifier-driven" coils are now first-class keywords in Studio 5000 v36, TIA Portal V20, GX Works3 1.090, and Sysmac Studio 1.60. This makes latching and seal-in patterns more portable across brands than they were in 2024.
The six methods below still hold — they have been used since the 1980s — but the 2026 cut-offs and tooling around them are different.
Method 1 — Count honest outputs and group mutually exclusive loads
Before any wiring trick, walk the panel and write down every discrete load. Mark which ones are mutually exclusive: a motor running forward cannot run reverse; a green indicator cannot be lit at the same time as a red one; a heating element and its cooling fan are usually interlocked. About 30% of "needed" outputs disappear the moment you ask the question. The remaining gap is what the next five methods close.
> Tip
Build the I/O list on paper before opening the programming software. For each output, mark (a) the load type, (b) its voltage class, (c) whether it is mutually exclusive with anything else on the list, and (d) whether it is safety-rated. Items in column (c) are the first candidates for group control.
Method 2 — Group control (one output drives many loads)
Group control is the highest-ROI technique in any 2026 audit. A single PLC output energises an interposing relay or contactor; the relay's contacts then drive every load in the group. This is the canonical pattern for "all tower lights off," "all conveyor purge valves open," or "all cabinet fans on." One output point controls a process that would otherwise need ten.
In 2026 we see group control scaled to 30+ motor starters on Mitsubishi MELSEC iQ-R panels using the built-in SET / RST instructions, and on Allen-Bradley ControlLogix with the OTE + branch-seal pattern. The cost is zero hardware beyond a DIN-rail relay; the savings are immediate.
Method 3 — Multiplexing and diode matrixing (row/column scan)
Multiplexing scans a matrix of loads so that N + M outputs drive N × M devices one at a time. A 4-by-4 matrix uses 8 outputs to drive 16 indicators, halving the point count. The trade-off is update latency: only one row is energised at any instant, so the scan period must be faster than the human eye (about 20 ms total, or 1.25 ms per row for 16 rows).
Diode matrixing is the same idea with discrete diodes instead of strobed transistors — useful for indicator banks where the load is always-on (no scan). In 2026 we see multiplexing most often in tower-light retrofits on Allen-Bradley MicroLogix 1400 panels, and in Mitsubishi FX5U indicator stacks. Drive the matrix with sinking outputs (transistor, NPN) and a ULN2003 / ULN2803 row driver, never with a relay output — relay dry-contact multiplexing is a known failure mode after a few million cycles.
// Studio 5000 ladder snippet: 4x4 indicator matrix scan
// Outputs: O:0.0..O:0.3 = column strobe, O:0.4..O:0.7 = row enable
// 5 ms task, free-running
XIC Scan_Tick_5ms OTE Col0 // strobe column 0
XIC Scan_Tick_10ms OTE Col1
XIC Scan_Tick_15ms OTE Col2
XIC Scan_Tick_20ms OTE Col3
// Row data sourced from a 16-bit integer tag Row_Pattern[0..3]
! Warning
Cascading more than about 16 matrix stages with electromechanical loads causes visible flicker. Above that, switch to a dedicated SPI-driven LED driver or move the load to remote I/O — not more stages.
Method 4 — BCD coding for numeric displays
Binary-coded decimal uses four bits to represent 0–9. Driving a 7-segment display directly needs seven output points; driving it through a BCD-to-7-segment decoder (CD4511, SN7447) needs four. Two BCD digits (a 2-digit display) cost eight output points instead of fourteen. Three digits cost twelve instead of twenty-one.
In 2026 most integrators skip discrete decoders entirely and use a smart HMI: a single Ethernet point delivers a 4-digit BCD value to a PanelView 5310 or a Siemens KTP700, which renders the number. The BCD-output technique is still relevant for retrofits where the operator panel is fixed and cannot be replaced.
Method 5 — High-density output modules
When software tricks run out, the next move is hardware density. In 2026 the price-per-point on a 32-point module is roughly half what it was in 2020, and the modules are noticeably smaller. The Mitsubishi QY42P, for example, packs 64 sourcing outputs into a single MELSEC-Q slot.
| Part Number | Brand & Series | Type | Points / Slot |
|---|---|---|---|
| 1769-OB16 | Allen-Bradley CompactLogix | 24 VDC sourcing | 16 |
| 6ES7 322-1BL00 | Siemens SIMATIC S7-300 | 24 VDC sourcing | 32 |
| QY42P | Mitsubishi MELSEC-Q | 24 VDC sourcing | 64 |
| CP1E-N40DR-A | Omron SYSMAC CP1E | Relay (mixed) | 16 (relay) |
Notice the Mitsubishi QY42P packs 64 outputs in a single slot. In tight retrofits, a 64-point module on a Q-series backplane often replaces a chassis full of 16-point modules — the points per cubic centimetre ratio is what 2026 integrators care about most.
Method 6 — Remote I/O over EtherNet/IP, PROFINET, or CC-Link IE TSN
When the panel count exceeds about 50 discrete outputs, the right answer is to stop cramming them onto the CPU. Distributed I/O drops a small block (4, 8, or 16 points) at each load, and the PLC reaches them over a single industrial Ethernet cable. In 2026 this is the default design pattern, not the advanced one.
- Allen-Bradley: 1734 POINT I/O on EtherNet/IP, with a 1734-AENTR adapter per drop. Up to 63 modules per adapter, 504 points per drop.
- Siemens: ET 200SP on PROFINET, with a 6ES7 155-6AU01 head module. Up to 32 modules per station.
- Mitsubishi: CC-Link IE TSN with a NZ2GF2B1-32D digital output block.
- Omron: NX-IO on EtherCAT, with an NX-OD4256 32-point output unit.
The hidden win is cabling: a four-conductor Ethernet cable replaces 32-tray-of-conduit home runs. In 2026 panel-build labour is the dominant cost, not the I/O electronics, and remote I/O is the only architecture that bends that curve.
Decision matrix — which method when
| Situation | Recommended method | Typical savings |
|---|---|---|
| 4–16 mutually exclusive loads | Group control (one output + interposing relay) | 50–75% |
| 4–16 indicator lights | Multiplexing (row/column scan) | 50–75% |
| Numeric display (2–4 digits) | BCD decoder or smart HMI | 40% (4 points vs. 7) |
| 20–50 outputs on a single panel | High-density module (e.g. 1769-OB16, QY42P) | 2× density |
| 50+ distributed outputs | Remote I/O (POINT I/O, ET 200SP, NX-IO, CC-Link IE TSN) | Cable cost −60% |
Common pitfalls we still see in 2026
- Multiplexing relay outputs. The cycle life is wrong and the bounce is audible.
- Grouping safety-rated loads behind a single non-safety output. Each E-stopped load needs its own safety output or a safety-rated contactor.
- Diode matrixing into an input that is also being driven by a sourcing sensor. Use a sinking sensor + diode matrix on the load side, not the other way around.
- Counting remote I/O points as "free" because the Ethernet cable is cheap. You still need a managed switch, a ring supervisor, and an IP-addressing plan.
- Mixing AC and DC loads on a relay-output group without per-coil RC snubbers. The back-EMF from a DC coil will weld neighbouring contacts.
Need this part?
Designing a high-density output panel?
Send your I/O list and BOM to Moritta@KOEED.COM. We keep new, legacy, and cross-reference stock across Allen-Bradley, Siemens, Mitsubishi, Omron, and POINT I/O — one quote within 24 hours.
Send My BOM →Frequently asked questions
Is multiplexing still worth it in 2026 with cheap remote I/O?
Yes, for two specific cases: (1) tower lights and stack lights where the load is intrinsically single-channel, and (2) brownfield retrofits where adding a network drop is more expensive than a small PCB of 74HC595s. For greenfield designs, skip multiplexing and go straight to remote I/O.
What is the lowest-cost way to add 16 outputs to an existing MicroLogix 1400?
Add a single 1769-OB16 on the local 1769 backplane. The CPU supports up to three 1769 modules without an extra power supply, and the per-point cost in 2026 is the lowest in this module's history. KOEED keeps new and tested surplus stock of 1769-OB16 for exactly this work.
Can group control and diode matrixing be combined?
Yes — group-controlled contactors can drive diode-matrixed indicator stacks downstream. The seal-in is preserved because each upstream group sees only its own contacts; the indicator matrix sees the OR of all upstream groups. This is the canonical pattern for "any one of three stations can start the line and light its own beacon."
Does BCD coding still apply for HMI-driven displays?
No — once an HMI is in the loop, the BCD trick is obsolete. A single Ethernet-delivered integer to a PanelView 5310 or KTP700 replaces all the discrete BCD outputs. BCD remains relevant only for fixed operator panels that cannot be replaced.
How do I cross-reference an obsolete output module in 2026?
Email your old part number to Moritta@KOEED.COM with the application note. KOEED keeps active, EOL, and cross-reference stock for Allen-Bradley 1746 / 1747 / 1756, Siemens 6ES7, Mitsubishi Q-series, and Omron CP1E / CJ1 families, and replies with a quote within 24 hours.
Related on KOEED Blog
- How-to — full archive
- Engineering Notes — PLC I/O design deep dives
- Migration Guides — SLC 500 to ControlLogix, S7-300 to S7-1500, MELSEC-Q to iQ-R
- PLC Error Code Database — Siemens / Mitsubishi / Omron
- Modbus CRC Calculator — free online tool
- PLC Analog Calculator — 4-20 mA to engineering units
The KOEED Engineering Desk
Industrial automation editors at KOEED. We write about PLC sourcing, cross-reference, and legacy system support. Reach the team at Moritta@KOEED.COM.