Build a two-rail load matrix, test cross-regulation under uneven loads, and decide when one dual-output supply is not the right architecture.
Direct answer. Do not approve a dual-output power supply from total watts alone. Define the minimum, typical and maximum current on each rail, then test every credible pair—especially when one rail is heavily loaded and the other is light or idle. Cross-regulation is the change on one output caused by a load change on another. If either rail leaves the equipment’s voltage window, choose separate supplies or post-regulation: a downstream regulator for the affected rail.
AI-generated conceptual visualization of a two-axis rail-load matrix. It is not a circuit diagram, WEHO product, factory, customer installation or measured result.
The purchasing error: one wattage number hides two load axes
OEM teams often begin with a neat arithmetic check: add the 5 V and 24 V loads, compare the result with the supply’s headline wattage, and move on. That check is necessary, but it is not sufficient for a coupled multi-output supply.
The real requirement is a two-dimensional operating envelope:
- the current demanded from rail A;
- the current demanded from rail B;
- the combinations that occur during boot, normal operation, standby, service and fault recovery;
- the voltage window tolerated by each downstream load at its own terminals.
Jakiś Analog Devices design note on multiple-output flyback converters explains the underlying risk: one output is commonly regulated more directly, while other outputs can be affected by transformer coupling, rectifier drops and winding resistance. That reference describes converter design, not the internal topology of a particular WEHO unit. The procurement lesson is narrower and useful: do not assume two outputs behave like two independent supplies unless the product documentation and validation say so.
Build a rail-load envelope from equipment states
Start with states, not averages. A useful load ledger has one row per state and one column per rail.
| Equipment state | Rail A current | Rail B current | State duration | Allowed rail-A voltage | Allowed rail-B voltage | Evidence source |
|---|---|---|---|---|---|---|
| Power applied, controllers held in reset | ___ A | ___ A | ___ ms/s | ___ to ___ V | ___ to ___ V | Component data |
| Boot and peripheral enumeration | ___ A | ___ A | ___ ms/s | ___ to ___ V | ___ to ___ V | Measurement / data |
| Normal minimum activity | ___ A | ___ A | continuous | ___ to ___ V | ___ to ___ V | Load list |
| Normal maximum activity | ___ A | ___ A | ___ | ___ to ___ V | ___ to ___ V | Sequence review |
| Standby or sleep | ___ A | ___ A | continuous | ___ to ___ V | ___ to ___ V | Measurement / data |
| Restart after a protective event | ___ A | ___ A | ___ ms/s | ___ to ___ V | ___ to ___ V | Control specification |
Do not invent a minimum load merely to complete the table. Use each downstream device’s current specification or a controlled measurement. Include tolerances and simultaneous events. A logic rail may be at maximum while actuators are disabled; later, the actuator rail may step up while the logic rail falls. Those are precisely the off-diagonal conditions that a single “full load” test misses.
Convert the state table into a test matrix
At minimum, evaluate these static corners:
- both rails at their credible minimum loads;
- rail A maximum, rail B minimum;
- rail A minimum, rail B maximum;
- both rails at their credible maximum loads;
- each normal operating state that is not represented by those corners.
Then add transitions between credible states. For example, move the actuator rail from idle to its largest permitted step while the control rail remains at its normal load. Repeat the reverse transition. The purpose is not to create arbitrary abuse conditions; it is to reproduce the equipment’s documented operating sequence.
Run the matrix at the input-voltage and ambient conditions required by the project, following the power-supply documentation. Record voltage at both the supply terminals and the downstream load. Cable, connector and protection-device drop is a system result, not cross-regulation inside the supply, but both affect whether the load receives an acceptable voltage.
Specify acceptance limits before requesting a sample
“Stable” is not an acceptance criterion. For each rail, record:
- steady-state minimum and maximum voltage at the load;
- permitted transient deviation and its duration;
- maximum ripple/noise under the measurement method defined for the project;
- startup order, monotonicity or delay requirements, if the loads require them;
- behavior required when either rail is unloaded, overloaded or disconnected;
- recovery behavior after the fault is removed.
Use the downstream component limits to establish these values. Do not copy a generic percentage from another design. If a processor, sensor, relay or interface module has an undervoltage threshold, the margin should be assessed at that device—not only at the power-supply terminals.
The separate WEHO guide on multi-output power-supply sequencing addresses startup and shutdown order. This article addresses a different question: whether each rail remains within its required voltage window as the other rail’s load changes. Teams deciding whether they need one or several supplies can also consult Single Output vs Multi Output Enclosed Power Supply.
A worked requirement example—not a product performance claim
Assume an OEM’s preliminary load study produces this envelope:
- a 5 V logic rail varies from 0.45 A in standby to 2.20 A during processing;
- a 24 V peripheral rail varies from 0.10 A at idle to 1.40 A during an operating step;
- the logic devices must remain inside their documented input window during the peripheral step;
- the peripheral voltage must also remain inside its own load limits at the far connector.
These are hypothetical requirement values, not measurements of a WEHO supply. The useful test points are not just 2.20 A plus 1.40 A. They include (0.45 A, 0.10 A), (2.20 A, 0.10 A), (0.45 A, 1.40 A) and (2.20 A, 1.40 A), plus the actual transitions between states. If the 5 V rail moves outside its allowed window when the 24 V load changes, the proposed supply fails this project even when its total output power appears adequate.
That result may be solved by changing the supply, adding a properly engineered post-regulator, changing the load architecture or separating the rails. The correct response depends on measured behavior, thermal conditions, efficiency and the downstream tolerance—not on headline watts.
What the published WEHO D-series pages establish
The current WEHO pages provide useful shortlist data, but they do not replace the application matrix.
The WEHO D-120 dual-output power-supply page lists two variants:
- D-120A: 5 V and 12 V outputs;
- D-120B: 5 V and 24 V outputs.
Its published table lists different current ranges, adjustment ranges and load-regulation figures for CH1 and CH2. In particular, the page lists CH1 load regulation as 0.5%, while CH2 is listed as 5% for D-120A and 6% for D-120B. These are model-page values, not an independent laboratory verification and not a prediction for a customer’s wiring or transient loads.
| Published variant | CH1 | CH2 | Published current ranges | Published load-regulation rows |
|---|---|---|---|---|
| D-120A | 5 V | 12 V | 0–12 A / 0–5 A | 0.5% / 5% |
| D-120B | 5 V | 24 V | 0–6 A / 0–4 A | 0.5% / 6% |
The page does not publish the conditions for allocating the headline 120 W between rails. Do not infer that both listed rail maxima may be combined arbitrarily; ask for the applicable combined-load limits for the exact variant.
The WEHO D-60 page is an additional family reference. Its table lists D-60A (5 V and 12 V) and D-60B (5 V and 24 V), output-current ranges with non-zero lower bounds, rail-specific load-stability values, and voltage adjustment on CH1. The page does not define whether those lower bounds are mandatory minimum-load requirements. Treat them as a question for model-specific confirmation, not as permission to infer that a preload is required.
Official WEHO D-120 product image from the published product catalogue. Confirm the exact variant and current revision before design release.
The pages establish that relevant dual-voltage variants exist and identify several electrical limits to review. They do nie publish a complete customer-specific cross-regulation surface, dynamic interaction for the proposed loads, downstream cable drop, or evidence that every load combination in your machine will pass. Request a model-level review and agree a sample-validation plan with WEHO where the documented limits alone do not close the decision.
When one multi-output unit is the wrong architecture
A dual-output supply can reduce component count and wiring, but separate regulation is often the cleaner choice when:
- one rail must stay tightly controlled while the other has large steps;
- one rail operates near zero load and the proposed supply specifies a minimum load;
- the loads require independent enable, shutdown or fault containment;
- different grounding or isolation domains are required;
- one rail’s noise budget cannot tolerate interaction from the other load;
- future options make the combined envelope uncertain.
This is not a universal rule in favor of separate supplies. It is a prompt to compare the cost of validation and mitigation with the cost, space and efficiency of another converter.
Send an RFQ that can be technically answered
For a useful model recommendation, attach this evidence to the enquiry:
- required output voltages and their tolerance at the loads;
- minimum, typical and maximum current for every rail;
- the state table and the largest credible load steps;
- input-voltage range, frequency and any source constraints;
- ambient range, enclosure, airflow and mounting orientation;
- startup/shutdown order and fault-recovery requirements;
- isolation, grounding, ripple/noise and compliance requirements;
- available installation envelope and expected quantity.
Use the WEHO contact page to paste the key rail-load values into the enquiry, or ask for the approved route for transferring the completed table. The primary next step is a model-level fit review, not an immediate claim that D-60 or D-120 is suitable.
Często zadawane pytania
Is total output wattage enough to size a dual-output power supply?
No. Total power is only one boundary. Each rail’s current range, combined limits, voltage tolerance and behavior under uneven loads must also match the application.
What is cross-regulation in a multi-output power supply?
It is the change in one output voltage caused by a load change on another output. The effect depends on converter design and operating conditions, so it must be assessed from model documentation and relevant tests.
Can I leave one output unloaded?
Only if the selected model’s documentation permits it and the remaining rail stays within the project’s limits. The D-60 product page lists non-zero lower bounds in its output-current range, but does not define them as mandatory minimum loads; ask WEHO how the field applies to the exact variant and revision.
Should I add a dummy load to improve regulation?
Do not add one by default. A dummy load wastes power and adds heat. First confirm the model’s requirements and measured behavior; then evaluate post-regulation, a defined preload or separate supplies as an engineered system decision.
Decision summary
- GO: documented rail and combined-load limits cover the requirement, and the two-axis test keeps both loads inside their voltage windows.
- HOLD: the real light-load state, combined-load rule, cross-regulation behavior or acceptance evidence remains unknown.
- CHANGE ARCHITECTURE: a credible load pair violates a rail limit or voltage window; evaluate post-regulation or separate supplies.
