An MCCB protecting a feeder that sits downstream of a fuse is a common arrangement in retrofit work, in mixed-brand panels, and in assemblies where the incoming device is a fuse-switch and the outgoing ways are breakers. It is also one of the few coordination problems where a single ratio, quoted confidently in a lot of panel-building folklore, decides whether the design is actually selective or only looks that way on paper.
The ratio is real, but it is conditional. It holds when the fuse is the upstream device and the breaker is downstream, it describes a narrow band of fault currents, and it collapses the moment the two devices are swapped round or a current-limiting fuse enters the picture. MCCB fuse selectivity is therefore not a rule you apply — it is a claim you verify against a published coordination table and a fault level you have actually calculated.
This article sets out what selectivity between the two device types means, why the direction of the pair changes which method applies, the ratio bands that appear in published tables and what they are really a proxy for, and the six situations where the ratio method gives the wrong answer. It closes with the fields a specification should carry so the panel builder is not left guessing.
What Selectivity Between an MCCB and a Fuse Actually Means
Selectivity, or discrimination, is the property that a fault is cleared by the protective device nearest to it and by no other device. When a fault appears on an outgoing feeder, the outgoing breaker should trip and the incomer should stay closed. Everything downstream of the fault loses supply; everything upstream keeps running.
Two different claims get made under the same word, and they are not equivalent:
| Term | What it guarantees | How it is proven |
|---|---|---|
| Total selectivity | The downstream device clears every fault up to its own breaking capacity without the upstream device operating | The published selectivity limit current Is is at least the prospective fault current at that point |
| Partial selectivity | The downstream device clears faults up to a stated current Is; above Is both devices may operate | Is is published as a number; the real fault level must be compared against it |
| Cascading (back-up protection) | Nothing about who clears the fault — the upstream device is allowed to help break the fault current | A series rating table, which is a different table with a different purpose |
The number that matters is Is. A coordination table that says “selective” without a current is not telling you anything you can design with. Manufacturer tables for fuse-and-breaker pairs usually print Is either as a current in kA or as a footnote saying the pair is selective up to the breaking capacity of the downstream device, which is the strongest statement available.
One more distinction belongs here. Selectivity is verified at a stated voltage and for a stated pair of products. A table entry for a 250 A breaker behind a 250 A fuse at 415 V says nothing about the same breaker behind the same fuse at 690 V, and nothing about a different breaker of the same rating from another manufacturer.
Why the Order of the Two Devices Changes the Rule
A fuse and an MCCB clear faults by different mechanisms, and that asymmetry is the whole reason the ratio method only works in one direction.
A fuse clears by melting a calibrated element. Its behaviour is described by an I²t characteristic: the energy let through depends on the current and the pre-arcing and arcing times, and for current-limiting fuses the total clearing I²t is published and remarkably consistent. It has no settings, no adjustability, and no short-time delay. It cannot be told to wait.
An MCCB clears by a mechanism that has at least two stages. The long-time element handles overloads over seconds or minutes. The instantaneous element operates in milliseconds above a threshold. On a thermal-magnetic device that threshold is fixed by the frame and the rating. On an electronic trip unit it becomes Ii, and often an adjustable short-time stage with a delay tsd sits between the two.
That difference produces two very different coordination problems:
- Fuse upstream, MCCB downstream. The breaker is the faster, more controllable device. The question is only whether the fuse’s I²t is large enough that it will not melt while the breaker is clearing. A ratio of ratings is a workable proxy for that energy margin.
- MCCB upstream, fuse downstream. The breaker is the upstream device and must be prevented from operating. If its instantaneous element has a fixed threshold, that threshold will be reached during a downstream short circuit no matter what the fuse does. A ratio of ratings cannot express that problem, because the limiting factor is a current threshold, not an energy margin.
Much of the confusion in panel shops comes from applying a rule learned in the first case to the second. The tables do not transfer.

Fuse Upstream, MCCB Downstream: The Ratio Method
This is the arrangement the familiar ratio rule describes: an upstream fuse-switch or fuse incomer, with MCCB outgoing ways. The reasoning runs in two separate current regions.
In the overload region, selectivity is almost automatic. The downstream breaker’s long-time element operates in seconds at a modest multiple of its rating. The upstream fuse, sized well above that current, is far below its melting I²t at the same current and simply does not respond. Providing the fuse rating is meaningfully above the breaker rating, overload selectivity follows.
In the short-circuit region both devices are fast, and the comparison becomes an energy comparison. The breaker’s clearing I²t in its instantaneous range must be lower than the fuse’s total clearing I²t at the same prospective current. Because both quantities scale roughly with device rating, a ratio of rated currents is a serviceable stand-in for that comparison — which is why the rule exists at all.
The bands below are the kind of figures that appear in published coordination tables for gG fuses upstream of thermal-magnetic breakers. They are illustrative, not universal: the same pair of ratings from two manufacturers can produce different Is values, and the only number you can design with is the one in the table for the products you bought.
| Upstream fuse / downstream MCCB rating ratio | What is commonly published | Practical reading |
|---|---|---|
| Below about 1.6 : 1 | Rarely selective in the short-circuit region; may still be selective in overload | Treat as non-selective until a table proves otherwise |
| About 1.6 : 1 to 2 : 1 | Frequently quoted as the minimum band for total selectivity on gG fuses | Plausible, but check the actual table entry |
| Above about 2 : 1 | Usually selective, often up to the downstream breaking capacity | Comfortable margin; still verify at the design voltage |
Two conditions travel with that table and are frequently dropped when the rule is repeated. The fuse must be a general-purpose gG type, not a semiconductor or motor-rated fuse. And the ratio refers to the fuse’s rated current against the breaker’s rated current In — not against its frame size, and not against its long-time setting Ir. A 250 A frame carrying 160 A, which is a normal and correct choice, needs the ratio computed from 160 A.
MCCB Upstream, Fuse Downstream: Why Ratios Stop Working
Reverse the pair and the physics changes. A short circuit on the fuse-protected feeder draws a current that the upstream breaker also sees. If that current exceeds the breaker’s instantaneous threshold, the breaker’s trip decision is already made within a few milliseconds, and the fuse — even though it is the device nearest the fault — never gets the chance to finish clearing.
What determines selectivity here is not a rating ratio but the relationship between two numbers: the maximum fault current the downstream fuse will pass before it clears, and the upstream breaker’s instantaneous pick-up. If the breaker’s threshold sits above that current, it holds. If it sits below, it trips.
On a thermal-magnetic MCCB the instantaneous threshold is fixed at manufacture, typically at a multiple of the rating that is not adjustable. That leaves almost no design freedom: either the numbers happen to work out or the arrangement is only partially selective. Ratio tables for this direction are therefore scarce, and where they exist they are tied to specific device pairs.
An electronic trip unit changes the picture, because it introduces a short-time stage. Setting a short-time pick-up Isd above the downstream fault current and a delay tsd long enough for the fuse to clear lets an upstream breaker hold through a downstream short circuit — the breaker waits deliberately rather than by luck. That is a design decision with a cost: while the breaker is holding, it is carrying fault current, and its short-time withstand rating has to be adequate for the delay chosen. The relevant settings and their limits are covered in the guide to electronic trip unit settings.
The practical consequence for a panel builder is blunt. An MCCB incomer feeding fuse outgoing ways is a coordination problem that needs a table, not a ratio, and if the incomer is thermal-magnetic the answer is often “partially selective up to X kA” — which may be perfectly acceptable, as long as somebody wrote X down.
Where the Ratio Method Breaks Down
Six situations turn a ratio that looks fine into a design that is not selective. Each one has caused a real panel to lose its incomer on a downstream fault.
- Current-limiting and semiconductor fuses. aR and gR fuses are built to clear very fast and let through very little energy. Their clearing I²t is a fraction of a gG fuse of the same rating, so the upstream device barely sees the fault — but the ratio rule, which assumes a gG characteristic, no longer describes the device. A ratio that would be generous with a gG fuse can be marginal or wrong with an aR fuse.
- A breaker with no short-time delay. As above: the fixed instantaneous threshold, not the rating ratio, decides. This is the single most common failure of the rule.
- High prospective fault current with a low X/R ratio. The asymmetric peak and the DC component shift the I²t that each device actually experiences. A pair verified at 25 kA is not automatically verified at 50 kA, and the tables are usually stated per fault level band for exactly this reason.
- Mixed standards regimes. An IEC gG fuse, a UL Class J fuse and a UL 489 breaker are characterised under different standards with different test circuits and different definitions of rating. A ratio derived from IEC 60269 data does not carry over to a UL combination, and vice versa. This matters most on export panels where the incoming device and the outgoing devices come from different standards worlds.
- DC systems. Arc extinction on direct current has no natural zero crossing, and neither the fuse’s nor the breaker’s AC rating transfers. DC selectivity is a separate verification with separate tables.
- Fuse derating inside a hot enclosure. A fuse selected on its nominal rating but operated at elevated ambient carries less than its nameplate suggests, and its melting characteristic shifts. If the fuse was already close to the load current, the margin assumed by the ratio quietly disappears. The same enclosure temperature effect on breakers is covered in the article on MCCB derating for temperature and altitude .
Cascading Is Not Selectivity
This is the confusion that costs the most, because both concepts involve an upstream device helping out and both appear in the same catalogues.
Cascading, also called back-up protection or series rating, is a technique for installing a downstream breaker whose own breaking capacity is lower than the prospective fault current, on the basis that the upstream device will share the duty and the pair has been tested together. It is a legitimate and economical technique. It says nothing about which device operates first.
Selectivity is the opposite claim: the downstream device operates first, every time, and the upstream device does not operate at all.
A panel can be cascaded and non-selective at the same time, and that combination is extremely common in cost-optimised assemblies. The consequence is that a fault on one outgoing way trips the incomer and takes the whole board down — the exact outcome selectivity was supposed to prevent. When a coordination table entry is being read, the first question is which of the two claims it is making. A table headed “back-up protection” or “series rating” answers a breaking-capacity question. A table headed “selectivity” or “discrimination” answers the one you want.
How to Verify Selectivity With Real Data
Verification is a short sequence, and every step produces a number that goes into the design file.
- Establish the prospective fault current at the point of interest. Not at the transformer, and not the board’s headline figure — the value at the downstream device’s terminals, after the impedance of cables and busbars between them. The method is set out in the guide to calculating the prospective short-circuit current .
- Get the coordination table for the actual products. Manufacturer name, series, rating, fuse type and standard, system voltage. A table for a different series is a different table.
- Read off Is. If the table gives total selectivity up to the downstream breaking capacity, and the downstream breaking capacity is at least the prospective fault current, the pair is fully selective. If the table gives a number, compare it against step 1.
- Check the voltage and the fuse type in the table header. A 415 V entry is not evidence for a 690 V system. A gG entry is not evidence for an aR fuse.
- Sanity-check the two curves visually. Plot both time-current characteristics and confirm they do not cross below the fault level. Curves that touch at 3 kA with a 20 kA fault level are irrelevant to the design, and curves that cross at 18 kA with a 20 kA fault level are a marginal design. The mechanics of reading the two time-current curves together are worth the ten minutes.
- Record the result, including partial selectivity limits. “Selective to 12 kA” is a usable statement. “Selective” on its own is not.
What to Write in the Specification
Most selectivity failures on site are documentation failures. The panel builder receives a single-line diagram with device ratings and nothing else, chooses devices on price and availability, and the coordination question is never asked. Six fields close that gap.
| Specification field | Why it is needed |
|---|---|
| Prospective fault current at each distribution level | Without it, no selectivity statement can be evaluated |
| Required selectivity level (total or partial, and to what current) | Distinguishes a design requirement from an aspiration |
| Fuse type and standard for each fuse position (gG, aM, aR, Class J, Class T) | The ratio and the table both depend on the characteristic, not just the rating |
| Breaker type — thermal-magnetic or electronic, and whether a short-time delay is available | Decides whether the ratio method applies at all |
| System voltage and earthing arrangement | Table entries are stated per voltage; pole configuration follows the earthing system |
| Who verifies coordination, and against which table | Assigns the check to a named party instead of to nobody |
For export projects, add the standard regime to the list. A board designed to IEC 61439 and a board designed to UL 508A are coordinated with different tables, and a buyer comparing quotations will not see the difference until the panel is inspected.
Selectivity Is a Number, Not a Habit
The ratio rule is a useful first estimate and a poor final answer. It works in one direction of the pair, for one family of fuse, over a band of fault currents, at one voltage, under one standard. Every one of those conditions is a place where a real installation can fall outside it.
The defensible approach is unglamorous: calculate the fault level where it matters, take the manufacturer’s coordination table for the products actually specified, read the selectivity limit current, and write both numbers into the design file. If the table says partial selectivity to a current below the prospective fault, that is a design choice to be made deliberately — accept it, change a device, or add a short-time delay — not something to be discovered when the incomer trips.
If you are coordinating a mixed board and the upstream device is already fixed, send the fault level and the existing device details through the contact form and we will work through which outgoing combination gives the widest selective range.
Perguntas frequentes
Is a 2:1 fuse-to-breaker ratio enough for selectivity?
It is a reasonable starting assumption for a gG fuse upstream of a thermal-magnetic breaker, and it is the band most often quoted. It is not a guarantee. The same 2:1 pair can be published as selective to the downstream breaking capacity by one manufacturer and as selective only to 10 kA by another, because the outcome depends on the actual clearing I²t of both devices. Treat 2:1 as the point at which you go and find the table, not the point at which you stop checking.
Can I get selectivity with the MCCB on the incoming side?
Yes, but not by ratio. If the incomer is an electronic trip unit with an adjustable short-time stage, you set the short-time pick-up above the maximum fault current the downstream fuse will pass and the delay long enough for the fuse to clear — provided the breaker’s short-time withstand rating covers the delay. With a thermal-magnetic incomer the instantaneous threshold is fixed, and the result is usually partial selectivity. That may be entirely acceptable for a small board; the point is to state the limit rather than assume it does not exist.
Does a bigger upstream fuse always improve selectivity?
In the fuse-upstream case, yes, up to a point: a larger fuse has a larger clearing I²t and gives the downstream breaker more room. Two constraints limit how far that can be pushed. The fuse must still protect the cable and busbar between the two devices, and it must not be so large that it fails to clear a fault in that section. Selectivity achieved by oversizing the incomer fuse past the cable’s capacity trades one protection problem for another.
What is the difference between a selectivity table and a cascading table?
A selectivity table states the fault current up to which only the downstream device operates. A cascading, or series rating, table states that the pair has been tested together so that a downstream device with insufficient breaking capacity can be used behind a specified upstream device. The second says nothing about which device trips first. A panel can be fully cascaded and completely non-selective.
Do coordination tables apply at 690 V if they were published for 415 V?
No. Breaking capacity, clearance and the tested combination all change with voltage, and manufacturers publish separate entries or separate tables per voltage. This is one of the reasons 690 V boards need their coordination verified independently rather than inherited from a 415 V design. The wider question of what changes above 415 V is covered in the article on stepping up to an air circuit breaker and in the Icu and Ics ratings of the devices you are pairing.
