Good I/O allocation decides whether one failed card trips a whole unit or only raises an alarm. This page walks through the rules engineers use to spread redundant signals, group by unit, size spares and count cards with a calculator.
DCS IO allocation is the job of mapping every field signal in the IO list to a controller, rack, card and channel. Learn the separation rules for redundant transmitters, grouping by unit and signal type, spare sizing and a card count calculator.
What Is DCS IO Allocation?
DCS IO allocation is the design step where every signal in the IO list is assigned to a specific controller, rack, I/O card and channel. It follows fixed rules so that redundant signals sit on different hardware, each process unit stays together and enough spare capacity remains for future loops.
- Put duplex and triplex transmitters on different cards, and preferably different racks and power supplies
- Group signals by process unit so one controller failure affects only one unit
- Keep AI, AO, DI, DO, HART and intrinsically safe circuits on their own card types
- Plan 20 to 30 % installed spare channels plus empty slots for new cards
- Count cards from points times spare factor, divided by channels per card, rounded up
Allocation starts only after the instrument IO list is frozen, because every row in that list becomes one channel. If you need a refresher on how the system itself is built, read our guide to the distributed control system before going further.
Control Global explains that in traditional marshalling the field wiring lands on terminal blocks and each pair is then cross wired by hand to the right I/O card and controller. Every one of those cross wires is a decision made during allocation, so a bad allocation becomes expensive rewiring later.
DCS IO Allocation Rules for Cards and Racks
Most EPC contractors follow a similar rule set for DCS IO allocation, applied in the order below.
Zero Instrument notes that signals for one piece of equipment, such as a pump start, stop, run feedback and trip status, should stay in the same controller or cabinet. This keeps interlock logic local and makes loop checks and DCS commissioning far simpler.
Emerson states in its DeltaV IS electronic marshalling data sheet that one CHARM I/O card handles 96 channels across 8 baseplates of 12 CHARMs. The same data sheet says each channel can have its own individually defined signal type.
How to Separate Redundant Transmitters
The main purpose of DCS IO allocation for redundant instruments is simple: no single card, rack, power supply or cable should be able to take out more than one leg of the measurement. If two pressure transmitters on the same header share one AI card, a card failure removes both and the redundancy you paid for is gone.
| Arrangement | Minimum Rule | Better Practice |
|---|---|---|
| Duplex AI (1oo2 or 2oo2) | A and B on different cards | Different racks and power supplies |
| Triplex AI (2oo3) | A, B and C on three cards | Three racks or three I/O nodes |
| Redundant control valve output | Two AO channels on two cards | Redundant AO card pair |
Our article on 2oo3 voting logic shows why a single shared card turns a triplex trip into a common cause failure. The same thinking applies to the broader family of voting architectures, where the logic is only as good as the independence of its inputs.
When you separate A and B transmitters, also separate their multicore cables and junction boxes. A single damaged cable tray can otherwise defeat perfect card allocation.
Signal Type Separation: AI, AO, DI, DO, HART and IS
Card types are built for one family of signals, so DCS IO allocation always groups points by type before anything else. A quick review of AI, AO, DI and DO signals helps when you sort the IO list into these groups.
4 to 20 mA transmitter inputs with HART diagnostics.
4 to 20 mA outputs to positioners and drives.
Dry contacts and NAMUR sensors for status.
Powered or relay outputs for solenoids and motors.
Low level temperature inputs with linearisation.
Energy limited channels for Zone 0 and Zone 1 devices.
Zero Instrument also advises keeping signals with different power types, such as 24 V DC and 230 V AC, on separate modules to avoid interference. HART capable cards matter if you want asset management, and our guide on how the HART protocol works explains what that extra channel carries.
How Much Spare I/O Should a DCS Have?
There is no single standard number, and engineers in a Control Design survey gave answers from 5 to 25 %. A Siemens engineer quoted there called 20 % the generally recommended spare, with 10 % as the lowest he would accept.
Many project teams settle on 20 to 30 % for DCS IO allocation, counting installed spare channels plus empty slots for future cards. Always follow the client design basis first, since spare philosophy is a commercial decision as well as a technical one.
- Easy to add loops during commissioning
- Absorbs late IO list changes
- Leaves room for plant debottlenecking
- Lower hardware cost at purchase
- Smaller cabinets and footprint
- Expensive new racks for any expansion
Card Count Calculation With Spare Percentage
The calculator below does the DCS IO allocation card count for one signal type. Enter the used points, channels per card, spare percentage and 1 for simplex or 2 for redundant card pairs.
Required channels = 180 × 1.25 = 225
Cards = 225 ÷ 16 = 14.06, round up
Cards needed = 15
Installed channels = 15 × 16 = 240
Real spare = 240 minus 180 = 60 channels
Real spare = 33.3 % of used points
Rounding up to whole cards gives more spare than you asked for, so the real spare here is 33.3 % and not 25 %. Run the calculation once per signal type.
For a quick check on transmitter loops on those cards, our HART loop voltage budget calculator confirms that each 4 to 20 mA loop still has enough voltage after barriers and cable resistance.
Marshalling Cabinet vs Electronic Marshalling
In a classic system, field multicores land in a marshalling cabinet and jumper wires connect each terminal to the allocated card channel in the system cabinet. With electronic marshalling, the field pair lands on any free terminal and the channel is assigned to a controller in software.
- Allocation must be fixed before cabinet build
- Late IO list changes mean rewiring
- Separate IS barriers and marshalling panels
- Any channel can serve any controller
- Signal type set by a plug in module
- Fewer cabinets and no cross wires
Emerson states in its DeltaV electronic marshalling data sheet that a CHARM I/O card can serve up to 4 controllers and that no cross wiring is required.
Electronic marshalling does not cancel DCS IO allocation separation rules, because two CHARMs on one baseplate still share a carrier. Keep duplex and triplex inputs on different CHARM I/O cards, and read how remote and distributed I/O changes the cabinet layout.
Before you lock the allocation, it helps to understand where each controller sits in the overall DCS architecture of operator stations, engineering stations and control networks.
IO List Columns Needed for Allocation
A DCS IO allocation sheet is just the IO list with hardware columns added. The table shows the minimum set that most projects use, which also feeds the instrument loop diagram and the cable schedule.
| Column | Example | Why It Matters |
|---|---|---|
| Tag number | PT 1101A | Links to instrument index |
| Service | Reactor feed pressure A | Shows redundant partner |
| Signal type | AI HART 2 wire | Selects card type |
| System | BPCS or SIS | Keeps safety separate |
| Hazardous area | Zone 1, IS | Selects IS channel |
| Controller / Rack | CTL 01 / R2 | Unit grouping check |
| Card / Channel | Slot 5 / Ch 07 | Wiring and loop check |
| JB and multicore | JB 11 / MC 11 03 | Cable separation check |
| Redundancy group | PT 1101 A, B, C | Card separation check |
The tag column must match the instrument index exactly, otherwise the allocation and wiring documents drift apart during revisions.
Common DCS IO Allocation Mistakes
| Symptom in Plant | Likely Allocation Cause | Fix |
|---|---|---|
| Both level legs fail together | A and B on one AI card | Move B to another rack |
| Whole unit trips on one card fault | Units mixed on shared cards | Regroup by unit at next shutdown |
| Slow loop response | PV and output in different controllers | Keep loop in one controller |
| No room for new pump | No installed spare or slots | Add remote I/O node |
| Noise on AI channels | AC and DC signals on shared cable | Separate cables and cards |
Print the allocation sorted by card and highlight every redundancy group in a different colour. Any card with two highlights of the same colour is a separation error.
Emerson describes each CHARM as having current limiting and a fail open design, so a fault on one channel is isolated from its neighbours.
- BPCS and SIS points on separate hardware
- Redundant legs on different cards, racks and power supplies
- Each process unit grouped in its own controller or card group
- Control loop PV and output in the same controller
- IS circuits on IS channels with segregated wiring
- 20 to 30 % installed spare and empty rack slots
- Controller CPU load below 50 % at FAT
- Allocation sheet tags match the instrument index
Use this DCS IO allocation checklist again during the DCS FAT and SAT, because late IO list changes often break the original separation.
DCS IO Allocation FAQ
DCS IO allocation is the process of assigning every field signal from the IO list to a controller, rack, card and channel. It is done after the IO list is frozen and before cabinet drawings are released.
The allocation follows rules for redundancy, unit grouping, signal type and spare capacity. A good allocation sheet becomes the base document for wiring, loop diagrams and loop checks.
Most projects keep 20 % installed spare channels and extra empty slots for future cards. Guidance from vendors and experts ranges from about 10 % to 30 % depending on project size.
Zero Instrument suggests 15 to 20 % for analog channels and 10 to 15 % for discrete channels. Always follow the client design basis, because spare capacity affects both cost and future expansion.
If two redundant transmitters share one card, a single card failure removes both readings at once. That turns a redundant measurement into a single point of failure for control or trip logic.
Putting each leg on its own card, rack and power supply keeps at least one healthy reading. The same rule applies to 2oo3 voting where three legs need three separate cards.
A marshalling cabinet receives field multicores and provides terminals, barriers and cross wiring to sort the signals. The system cabinet holds the controllers, I/O cards and power supplies that process those signals.
Cross wiring between the two cabinets follows the allocation sheet channel by channel. Electronic marshalling designs reduce or remove this cross wiring because channels are assigned in software.
Yes, the measured value and the valve output of one control loop should sit in the same controller. This avoids delays and extra failure points from peer to peer communication between controllers.
The input and output can be on different cards inside that controller without any problem. Only cascade or plant wide strategies normally need signals shared between controllers over the network.
Related Articles
- Instrument IO List for DCS and PLC
- Marshalling Cabinet in DCS Wiring
- DCS IO Card Types Explained
- Voting Architectures in Safety Systems
- DCS Controller Redundancy Types
External References
- Emerson DeltaV IS Electronic Marshalling Product Data Sheet
- Zero Instrument: I/O Allocation Principles and I/O Point Calculation for DCS
- Wikipedia: Distributed control system
What We Learn Today
- How dcs io allocation assigns each IO list signal to a controller, rack, card and channel using clear and repeatable design rules.
- Why duplex and triplex transmitters, 1oo2 and 2oo3 inputs must sit on separate cards, racks, power supplies and cables to avoid common failures.
- How to size spare channels, count cards with a calculator, and compare traditional cross wiring with electronic marshalling in modern control systems.
