In short. Modular UPS systems build capacity from hot-swappable power modules in a frame. They scale with load and give N+1 redundancy, but true redundancy depends on shared components, battery strings and bypass design, which can be single points of failure.
How it works
Frame with control, static switch and bus → power modules share the load → failed module is removed and replaced while the rest run.
Benefits
- Right-size now, add modules later.
- N+1 within one frame.
- Faster repair by swapping modules.
- Efficient operation when modules share load sensibly.
Redundancy checks
| Item | Question |
|---|---|
| Static switch and control | Single or redundant? |
| Battery | One string or multiple independent strings? |
| Input and output breakers | Common point of failure? |
| Cabling | Separate routes? |
| Maintenance bypass | Can the frame be isolated safely? |
Redundancy needs independent failure paths. See N+1.
Sizing and growth
- Size N for current load, plus the redundant module.
- Reserve frame and cable capacity for growth.
- Match generator capacity to the final build, not just day one.
With generator
Modules ramp together; check walk-in and harmonics at full build. See UPS and DG.
Illustrative example
A small data centre installs a frame for four modules and populates three, one of which is redundant. When one module faults, the others carry the load and service swaps the module with no outage. A single battery string is later flagged as the weaker point, so a second string is added.
Efficiency and load sharing
UPS efficiency typically peaks at moderate to high load and falls at light load. In a modular system, intelligent control can switch off surplus modules so that those running operate at better load, while the redundant module remains on standby. This can improve efficiency compared with several lightly loaded modules, but the redundancy logic must still ensure that a module failure does not drop the load.
Fault domains
| Layer | Question to ask |
|---|---|
| Module | Can one fail without affecting others? |
| Frame bus and controls | Are these redundant? |
| Static switch | Redundant or single? |
| Battery strings | Independent, or one shared bank? |
| Upstream feeders | Same breaker or independent? |
| Downstream distribution | Dual path to loads? |
Capacity planning
- Establish the present and ten-year load.
- Choose a frame that supports the final size.
- Install modules to match near-term needs, plus redundancy.
- Provision cables, breakers and cooling for the final size.
- Review annually.
Service model
Hot-swap modules reduce downtime, but require trained staff and spares. Agree a service plan with response times. Keep at least one spare module where the support contract cannot guarantee delivery quickly.
Testing redundancy
Simulate a module failure by removing one under controlled conditions and confirm the others carry the load. Test battery failure and bypass operation in the same programme. A redundancy never tested is only a claim.
The master engineering question
If any one component fails, does the load stay up, and have we confirmed there is no shared single point?
What to send us
Send load growth plan and room layout. Send your application details.
Questions engineers ask
Is modular always redundant?
No; depends on shared components.
Can I hot-swap?
Many designs allow it; follow the maker’s procedure.
Does it save energy?
Can, by running modules at better load points.
What about batteries?
Plan independent strings where redundancy matters.
Values are illustrative. Use manufacturer data.