SIMATIC S7-1500 Power Budgeting (24V Load Planning): How Engineers Stop “Mystery Resets” Before Commissioning
In automation, 24 V DC is the quiet backbone of everything you care about: PLC, remote I/O, sensors, network switches, relays, IO-Link masters, and all the little “aux” loads that never make it into the first spreadsheet. When a 24 V system is under-designed, the failures look random: a CPU that reboots when a valve fires, ET 200 stations that drop off the network, analog values that wander after a contactor event, and troubleshooting that turns into folklore.
This guide is decision-oriented. It follows the workflow most experienced panel builders use: define what must stay alive → build a load schedule (steady + peak) → size PSU with realistic margin → design distribution & branch protection → add buffering/UPS where it matters → verify voltage at the far end under worst-case load.
24 V distribution still has real hazards (fire risk from overheated conductors, arc energy on DC faults, machine safety implications, and code requirements). Use this guide to avoid common design mistakes, and follow your jurisdiction’s standards for final design and installation.
Step 1 — Decide what “must stay alive” when 24 V gets stressed
Before you add up numbers, decide what “failure” means in your project. Some machines can tolerate a brief drop and recover; others need controlled shutdown or safety logic that must never brown out.
- Must stay alive: CPU, safety controller (if used), critical network switches, remote I/O feeding safety/critical sensors
- Can drop briefly: non-critical valves, indicator lights, convenience relays, nonessential auxiliaries
- Should be isolated: “dirty” loads (solenoids/contactor coils) from “clean” electronics (PLC/analog/comms)
If your design allows a solenoid or contactor coil to pull down the same 24 V rail feeding the PLC and network gear, you’re basically budgeting for “intermittent problems.”
Step 2 — Build a load schedule that includes steady current and peak events
Most “PSU sizing” mistakes come from treating everything as a steady load. Real panels have event-driven peaks: coils energize, IO-Link devices boot, sensors warm up, and remote stations come online together after an E-stop reset.
| Load group | Examples | Steady current | Peak / inrush | Notes that matter |
|---|---|---|---|---|
| Control electronics (clean) | S7-1500 CPU, central I/O, comms modules | From datasheet/manual | Inrush exists | Often modest current, but intolerant to dips |
| Distributed I/O (clean-ish) | ET 200SP / ET 200MP stations | Module sum | Startup + group loads | Watch per-group limits and terminal current capacity |
| Sensors | prox, photoeyes, pressure transmitters | 10–200 mA each (typ.) | Usually small | Don’t forget encoder supply, IO-Link loads |
| Actuators (dirty) | solenoids, contactor coils, relays | Varies | Peaks can be big | Suppression + branch protection changes everything |
| Accessories | switches, IO-Link masters, small HMIs | Often ignored | Boot peaks | These loads explain many “why now?” reboots |
Total 24 V current (steady) ≈ Σ(Isteady)
Worst-case event current ≈ Σ(Isteady) + Σ(Ipeak events)
(Then decide whether your protection/buffering strategy needs to handle the event without dipping the clean rail.)
Step 3 — Use real device data for the “clean core” (CPU + comms + remote I/O)
With S7-1500, you can pull meaningful numbers from equipment manuals/datasheets: you’ll see rated current, max current, and inrush current for CPUs. Treat these as the non-negotiable baseline—because if the control core browns out, everything else becomes noise.
Control electronics usually don’t “draw a lot,” but they’re the first to complain about voltage dips. Your job is to prevent dips—by separating dirty loads, using selective protection, and adding buffering where needed.
If you’re using ET 200SP, pay attention to electrical limits that don’t show up in simple “point counts.” For example, terminal current carrying capacity and per potential-group limits can cap what you can safely feed through a station, especially when motor starters or other higher loads are involved.
Step 4 — Size the PSU like a grown-up: margin is not optional, but it must be justified
The clean way to size a 24 V PSU is to separate three ideas:
- Steady current (the machine running in normal state)
- Event current (coils firing, restarts, device boot, simultaneous loads)
- Design margin (aging, temperature, future adds, “reality tax”)
- If your loads are mostly electronics and sensors: +20–30% margin is usually sane.
- If you have a lot of coils/actuators on the same PSU: either separate rails or design for the worst event (often +40%+), or add buffering/selectivity.
- If you expect expansion: leave headroom intentionally, not accidentally.
A quick sizing example (how you explain it in a design review)
| Group | Steady current | Worst event add | Design note |
|---|---|---|---|
| Control core (CPU + comms + central I/O) | 1.2 A | 0.3 A | Keep on clean rail |
| ET 200SP stations + IO-Link | 1.8 A | 0.6 A | Watch per-group limits and wiring |
| Sensors | 2.0 A | 0.2 A | Often underestimated |
| Actuators (solenoids/relays) | 2.5 A | 4.0 A | Dirty rail, suppression and branch protection |
| Total | 7.5 A | 5.1 A | Worst event ≈ 12.6 A (before margin) |
In that scenario, a “10 A PSU” looks fine on steady current and fails on events. The better solution is usually either: (1) separate a dirty actuator PSU, or (2) keep one PSU but add selectivity/buffering so peaks don’t drag down the core.
Step 5 — Distribution & branch protection: why selectivity beats “one big fuse”
This is where field experience shows up. If your 24 V distribution is a single undifferentiated rail, faults propagate: one short or overloaded branch can collapse the whole bus and reset the controller—turning a small wiring fault into a system outage.
The modern approach is electronic branch protection / selectivity: brief peaks are permitted, but sustained overloads are isolated to the failing branch, and the rest of the system stays alive.
Selectivity is not “extra cost.” It’s how you keep the PLC alive while you troubleshoot a single faulty sensor cable instead of rebooting the whole line and losing time arguing about what happened.
| Design choice | What happens on a short | Commissioning & maintenance reality |
|---|---|---|
| Single rail + basic fuse | Often drops the whole rail | Fault looks “random” (PLC resets, stations drop) |
| Branch protection / selectivity module | Faulty branch is isolated | You keep control power and diagnose fast |
| Clean vs dirty rails | Dirty events don’t pull down PLC rail | Stops the classic “valve fires → CPU reboots” pattern |
Step 6 — Buffering / UPS: the difference between “dip” and “reset”
Two real-world situations cause 24 V dips even when your PSU is “big enough”:
- Upstream disturbances: AC supply flickers, transformer taps, contactor events, or facility issues
- Event piles: many loads start together (restart after E-stop, station boot + coil firing)
Buffering and DC UPS modules exist for a reason: they bridge short gaps and keep the control core stable long enough to ride through disturbances or shut down cleanly.
If your CPU/network gear must not reboot, give them a “ride-through” strategy: buffer module for short dips or DC UPS for longer events / controlled shutdown.
Step 7 — Voltage drop: why “24 V at the PSU” is meaningless at the load
Voltage drop is the silent killer on long 24 V runs (remote sensors, field boxes, distributed stations). The 24 V rail can be perfectly sized and still fail because the far-end voltage falls below what electronics tolerate during peaks.
ΔV ≈ I · R(total loop)
(Remember: loop length is “there and back.” A modest resistance becomes painful at higher current.)
Measuring “no-load voltage” at the far end proves almost nothing. Measure under the worst event (coils firing / restart scenario), because that’s when voltage drop matters.
Design-freeze checklist (print this for the power review)
- Load schedule exists and includes steady current, peak events, and which loads are clean vs dirty.
- Control core protected: PLC/comms/critical I/O are isolated from dirty coil events (separate rail and/or selectivity).
- Branch protection defined: a fault in one branch doesn’t collapse the whole 24 V system.
- Distributed I/O limits checked: station current/potential-group constraints and terminal current capacity are respected.
- Buffer/UPS strategy decided for ride-through and controlled shutdown needs.
- Voltage drop verified at the far end under worst-case load events.
- Expansion headroom intentional: margin documented (not accidental).
Key takeaways
A good 24 V plan isn’t “pick a bigger PSU.” It’s a system decision: what must stay alive, how faults are isolated, how peaks are handled, and whether the far-end still gets enough voltage when real machines do real machine things.