I've been doing research on the electrical system in my home, for my own knowledge and also to verify things are code compliant.

I have a Mitsubishi mini-split which was installed 3 years ago by an HVAC company. The outdoor unit model is MXZ-3C24NAHZ2. The manual specs Max Fuse Size: 40A, and Min. Circuit Ampacity 29.9. There is a 40A breaker in my panel for this. However, they ran 10ga Romex from there to the disconnect (which is 8' away), instead of 8ga, which is what I thought was required.

  1. If MCA 29.9 can run on 10ga, why does the manufacturer specify a 40A instead of 30A?
  2. Did the contractor size things correctly?

The outdoor unit provides both heat and cooling, and everything has run fine since installation.

  • #2 is a yes if the answer to imaginary question #3 is also a yes; that it is 90c THHN copper? Romex is a no. That's cable not wire.
    – Mazura
    Feb 14 at 2:26

6 Answers 6


Having both a minimum (in your case 29.9A, which effectively means 30A) and a maximum (40A) is quite common. Note however, that if the actual load is 29.9A then it may need a continuous use derate, which brings it to 29.9A x 1.25 = 37.375A, which means a 40A breaker anyway.

There are a bunch of possible reasons why you might prefer the maximum rather than the minimum (when there really is a difference):

  • Add-on equipment powered from the same circuit
  • Better allowance for higher load during motor startup
  • Allow for future larger size equipment

But whatever you do, the wire size must be large enough for the breaker. The only way you get to 40A on 10 AWG wire is 90 C rating on wires, which (a) is not normally applicable even with the right type of wires and (b) you probably have cable rather than individual wires anyway.

You haven't had a problem because:

  • The actual load is probably close to 30A. That matches the basic rating of 30A for the cable anyway.
  • It is probably rare that it runs for really long times continuously.
  • Unless the cable is inside an insulated wall, it probably dissipates heat at a rate that safely allows more current (within reason).

It sounds like the contractor sized the circuit correctly (40A) but got lazy (reused existing 10 AWG cable already in place) and/or cheap when it came to the wiring.

That being said, replacing the 10 AWG cable with 8 AWG cable (or 8 AWG wires in conduit) is the right thing to do.

  • 4
    The 8 feet isn’t going to cost much at all.
    – Jon Custer
    Feb 13 at 2:20

MCA of 29.9 cannot use #10. An Air Conditioner isn't 125% because it's continuous, it's 125% because it's required in NEC 440.32(2), the Air Conditioning section. 29.9 x 1.25 = 37.375A. The best available termination rating limits even the best wire to the 75° column of 310.16, and would allow up to 35A on #10. But "Romex" is NM cable and is limited to 60°, so is limited to 30A.

Understanding UL labelling of Max Fuse Size is a little more complicated, UL considers what temperature the copper will be damaged and time curves for insulation damage, and other factors directly relating to equipment components, and results in numbers not defined in the NEC.

  • Because your question didn't give full details I didn't address other implied issues, but agree with other answers that fuse means fuse required and NM cable shouldn't penetrate into wet or damp location. Feb 13 at 18:22
  • 90° THHN #10 is good to 40 amps. But OP has romex though so that's right out. ... cannot use #10 romex.
    – Mazura
    Feb 14 at 2:23
  • 1
    @Mazura I was trying to not go too deeply into the weeds, i did say "best available wire and termination". The terminations are the key, the breakers and disconnect will be 75°, which brings the wire rating down to the termination rating. 90° rating is basically only good for doing some de-rating calculations. Feb 14 at 3:34

I think what’s gotten my attention is the part where you said they ran 10AWG ROMEX to the disconnect. ROMEX isn’t suitable for outdoor use, unsupported use, etc.

Replace that with an 8AWG whip of the required length. There are limits on unsupported whip length, attach it to something every 6 feet or less. There’s a 6 foot limits on the length of the flexible metal type when relying on the spiral metal tape for fault grounding. There’s no limit when there’s a green wire other than the support spacing. If you can’t find a pre-made whip you can buy liquid tight and 8AWG Type THHN wire and make your own, but it’s usually both easier and cheaper to buy than to make.

  • 1
    You may be right. But since the 8 feet of Romex is to the disconnect, it is possible that the disconnect is on the outside wall with the cable running through the wall into the back of the disconnect box. There seems to be some controversy as to whether that is allowed or not, but my understanding is that in many places it is allowed (including I am pretty sure my electrician (with inspection) wired up my generator inlet that way. Definitely needs to be 8 AWG. Feb 13 at 4:08
  • The Romex comes out of my garage wall just below the load center and goes into a 90 degree conduit outlet body. It then runs along the house in conduit for 2' and goes into the disconnect. From there, a whip goes out to the unit.
    – DaveinPA
    Feb 13 at 16:02
  • 1
    runs along the house in conduit for 2' If that means, as I think it does, on the outside wall of the house, then that is definitely against code. Romex (and similar non-metallic cable) is not normally rated for use in wet areas, and outside == wet, even though it is in conduit. Feb 14 at 15:11

The X-factor you're missing is that most motors have overload protection onboard the motor. And that works better because it can sense actual motor temperature instead of guesswork from afar.

Normally UL wants the breaker to be close to the running amps of the machine so it can provide overcurrent protection to the machine. However, with overcurrent protection covered, the breaker is only concerned with short-circuit (and bolted fault; similar to short-circuit) protection. As such, breaker sizing can focus on another common problem of motors: nuisance trips from normal startup surges.

And that brings us to Article 430, which allows us to upsize a motor breaker as much as 225%. A 40A breaker is going to do about the same job as a 30A breaker at detecting short circuits. A 300A short will simply be 750% of rating instead of 1000% of rating, making little difference in trip time.


The manufacturer gives you a range.

MCA of 29.9 means a 30A breaker is fine. So you could leave the wire alone, swap the breaker to 30A, done with the wire you have. This is the manufacturer telling you EXACTLY how much capacity is required to run the thing under worst case conditions, so exceeding it is neither required, nor particularly beneficial. Most of the time it will use less than that.

Max Fuse (or "Overcurrent protection") of 40A means you can use a 35 or 40A breaker if you want to, and that's also fine. If you do, you need suitable wire size for 35 or 40A, which you don't have now. If you were swapping out an old unit with 40A breaker and wiring in place, great, no need to change anything. If you choose to stay with the 40A breaker you don't need, you need to change the wiring to suit it.

Mitsubishi mini-splits are inverter drive, so they don't have huge start-up surges - they start slow and ramp up to the required speed. Knee-jerking to an oversized breaker to run them is outdated thinking.

I'm fairly certain that being specified as Minimum Circuit Ampacity (rather than just "Amps") the manufacturer has already performed the derate calculation; derating twice is not required.

Example case: I run a pair of them with an MCA of 10.9A (on 15A breakers, 240V) They are the major load on my power in winter. Actual typical hourly power use overnight, when it's basially them and the fridge, and the motion lights if a deer walks past, for the whole building recently was around 2.4KWh/h in cold weather and closer to 0.6 KWH/h in the late January/early February thaw we've just had. So 10A for the pair of them (and the fridge, but it's like 1 KWh/day) in colder wather, 2.5A for the pair of them in the thaw. Nowhere close to 21.8A for the pair.

If I go back to last year when the power company claims it was -20°F and my recollection is it was more like -16F the worst hour of the day (just before dawn) was 3.8KWh, or 15.83A for the pair and the fridge. Almost but not quite 8A actual, each.

  • 1
    I neglected to mention that Mitsu spec'd "Time Delay" for the fuse. Would a standard trip 30A trip unnecessarily?
    – DaveinPA
    Feb 13 at 2:55
  • 1
    If they spec a FUSE, you have to use a fuse (somewhere in the circuit.) Typically that would be in a fused disconnect, since we don't tend to have fuse panels any more. Given the trip curve of a typical breaker, it would be unlikely to trip without a fault. If the spec is for Max Overcurrent Protection Device a fuse is not needed, but if a fuse was used and they spec time delay for a fuse, you'd need that. Regukar fuses blow faster than breakers, typically.
    – Ecnerwal
    Feb 13 at 3:10

When the labeling on a complex machine with its own internal fuses and breakers specifies a maximum fuse on its input, it sometimes means that if the fuse doesn't stop a short circuit as high as what that size fuse would usually be required to stop, the short circuit current that makes it through a larger fuse might dangerously destroy the internal breakers in the machine . In IEC standards for power control boxes ("fuse panels") there is a technical requirement that the box shall have label stating the max worst case short circuit that the incoming supply can produce for a brief moment of horror, usually something like 4000 A or 8000 A, with many of the breakers and fuse holders needing to be more expensive models to handle the higher worst case rating. What this large number means is that if the main breaker or fuse has to, it can stop such a huge current safely at least once before catching fire, and that the installing electrical contractor is supposed to measure (in some clever but standard way) how much the incoming feed from the grid or local power source might supply if some gremlin suddenly put a huge copper bolt across the wires entering the box.

So if the makers of this air conditioner wrote "max 40 A fuse" it probably means that their factory engineers looked at the reaction times of common household fuses and concluded that a 50A fuse would let through enough juice to fry the thermostat relay contact points if something inside the unit shorted out out, but a 40A fuse would be safe enough for their product liability insurance . So running two of these units on a 60A fuse would be dangerous even though on a normal day they only draw 30A each.

Other answers are correct that the wires after a fuse must be rated to whatever that fuse is allowed to let through, so a 40A fuse must be followed by 40A wiring to not cause a fire if someone plus a 40A space heater into the wire instead of the device currently using those wires.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.