SAE J2064 Explained: A/C Hose Types A-G, Permeation Limits, and What Most Suppliers Get Wrong

Search “SAE J2064 hose types” and you will find three articles giving three different answers. One says the standard defines four types rated by pressure — 350, 500 and 550 psi. Another says Type C is only for the low-pressure suction side. A third quotes a permeation limit of 0.16 kg/m²/year, a number that does not appear anywhere in the standard.

This matters more than an academic argument. When a hose fails on a fleet vehicle, the document that gets pointed at is J2064. If you are buying hose by the meter or specifying assemblies for a customer, you need to know what the standard actually requires — and which of the numbers you are being quoted are real.

What follows is taken from the standard text and from how the testing is actually performed, then translated into the questions you should be asking a supplier.

1. First, the mix-up that causes most of the confusion

SAE publishes two standards for mobile air conditioning hose, and they cover different things:

• SAE J3062 — “Automotive Refrigerant Air-Conditioning Hose Requirements”. This is the bulk hose standard. Hose sizes, structure types, moisture ingression classes and identification markings live here.
• SAE J2064 — “Coupled Automotive Refrigerant Air-Conditioning Hose Assemblies”. This is the assembly standard: hose plus factory or field-coupled fittings and crimping tools.

Here is the part almost every article gets wrong: the seven hose structure types — A through G — are defined in J3062 and referenced by J2064. J2064 does not invent its own type system, and it does not sort hose into pressure tiers. It says, in effect, “the hose in this assembly must already have met J3062”.

Practical consequence: if you buy bulk hose by the meter, the structure type you are buying is a J3062 classification. If you buy a crimped assembly, J2064 adds three tests that do not exist on the bulk hose side — coupling integrity, bursting strength and proof test. A supplier who quotes you “J2064 hose” without being able to say which is which has probably not read either document.

2. The seven structure types

Type letters describe how the hose is built. Nothing more.

A "u" suffix (C_u, D_u, E_u, F_u, G_u) designates the ultra-low permeation limit.

Two things to take from this table:

• Types A and B have no thermoplastic barrier. They are all-elastomer constructions. They still have to pass a permeation test — the limit is simply far more generous.
• Types C through G all contain a thermoplastic layer. Whether it sits between plies (C, F), is veneered to the inner tube (E, F), or is the tube (D, G) is the difference between the letters.

3. Permeation limits: the numbers that actually exist

Permeation limits, SAE J2064 Table 2. These are maximum permitted loss rates, not typical values.

Read that table carefully, because it is the single most useful thing in the standard:

• 15 vs 5. On R134a, a barrier construction is allowed to lose one third of what a non-barrier construction may lose. That is the entire business case for barrier hose, and it applies to R134a — not just to R1234yf.
• R1234yf is harder on A/B hose. The limit for A/B rises to 18 on R1234yf and R152a, while barrier types stay at 5. The gap widens from 3x to 3.6x.
• 1.5 is the “u” class. Ultra-low permeation. If a supplier claims “ultra-low”, the number they should be showing you is at or below 1.5 kg/m²/year, measured to this method.
• These are ceilings, not averages. A compliant hose typically measures well under the limit — which is exactly why a good supplier quotes a measured value, not the limit.

4. Converting the units (so you can compare two quotes)

The standard reports permeation in kg per square metre of internal surface area per year. Suppliers in different markets report g/day, g/m/year, or oz/year. Here are the official conversion factors:

Conversion factors, SAE J2064 Table 1. For sizes not listed: factor = 108.66 / D (D = inside diameter in mm) for kg/m²/year, or 0.877 / D (D in inches) for lb/ft²/year.

And here is the translation most buyers actually want — what those limits mean in grams of refrigerant lost per metre of hose per year:

Grams of refrigerant per metre of hose per year, converted from the kg/m²/year limits above. Figures correspond to the 80 °C accelerated test condition, not to real-world annual loss — see the note below.

Important caveat: these figures correspond to the accelerated test condition (see below), which runs at 80 °C. Real-world loss on a vehicle at ambient temperature is substantially lower. Do not put 127 g/m/year on a quotation and present it as a field figure — use it to compare two hoses against each other, which is what it is good for.

5. How the permeation test is actually run

This is the detail that separates a credible data sheet from a hopeful one. The test procedure is:

1. Cut a specified length of hose, or build an assembly with production-intent fittings.
2. Charge it with refrigerant.
3. Place it in a constant-temperature oven at 80 °C ± 2 °C.
4. Hold it there for 25 days.
5. Measure the refrigerant loss and convert using the factors above.

Two implications you can use immediately:

• A permeation number without a temperature is meaningless. If a supplier quotes “1.6 kg/m²/year” but cannot tell you the test temperature or the refrigerant, you are looking at a number copied from someone else’s catalogue.
• Check the units. It is common to see permeability printed as “kg/m³/year” on Chinese supplier tables. That is a typo for m² — but it tells you the sheet was not written by the person who ran the test.

6. High side vs low side: which type goes where

The standard’s own guidance is straightforward:

• High side (compressor to condenser, and liquid lines): Types C, D, E, F, G — every one of them includes a thermoplastic layer specifically to minimise permeation.
• Low side (evaporator outlet back to compressor): normally Type A or B. These have no thermoplastic barrier, but must still meet their own permeation limit.

So the popular claim that “Type C is a suction-side-only hose” is backwards. It is the non-barrier A/B constructions that belong on the low side.

7. Four claims you should stop believing

8. What the rest of the standard tests — and which three only apply to assemblies

If you crimp in-house and sell the finished assembly, you own the three assembly tests. That is worth knowing before you quote a warranty: coupling integrity and proof testing are where field-crimped assemblies most often fail, and neither is covered by a bulk hose certificate.

9. Where our series sit

Structure class is stated as the closest equivalent. We publish measured inspection values rather than a certification mark we cannot substantiate.

Our measured inspection results, from production batch test reports:

• Burst: 35 MPa measured — more than 2.5x the 13.7 MPa minimum burst figure required for the size tested, and roughly 10x the 3.5 MPa working pressure
• Proof test: 7.0 MPa hold, no leakage or abnormality
• Length change at 3.0 MPa: measured 0%, against a +2% / −4% allowance. A hose that grows under pressure pulls on the ferrule until it weeps.
• Wall eccentricity: 0.15–0.33 mm measured against a 0.4 mm limit. Eccentricity is what makes a crimp seal on one side and leak on the other.
• EA48 permeation: tested with R404a / R410a at +80 °C, max. 9.7 kg/m²/year — inside the 15 kg/m²/year limit that applies to an all-elastomer construction
• EA48 heat aging: +140 °C × 200 h

Test records ship with the order. If a number matters to your customer, ask for the record rather than the catalogue.

10. A spec checklist you can send to any supplier

1. Type letter — and whether they mean the J3062 structure or are using the letter loosely.
2. Refrigerant the permeation figure was measured with. R134a, R1234yf and R404a are not interchangeable.
3. Measured permeation value with the test temperature, not the limit from the standard.
4. Burst and proof pressure on the actual size you are buying.
5. Length change at working pressure.
6. Wall eccentricity — the quiet cause of crimp leaks.
7. Minimum bend radius for the size.
8. Date code. Bulk A/C hose has a shelf life; rubber older than about three years loses flexibility and crimps poorly.

Related reading and products

• Automotive A/C Hose by the Meter: how to choose series, size and barrier class
• EA4 4-layer thin wall A/C hose — standard R134a / R1234yf systems
• CA5 5-layer thin wall A/C hose — extra permeation resistance
• CA9 5-layer thick wall A/C hose — abrasion and heat duty
• EA48 refrigerated truck hose — −40 to +140 °C, R404a / R410a tested
• Custom crimped A/C hose assemblies — built to drawing or OE number

Hose is supplied by the meter from stock, MOQ 100 meters per size, shipping in 3–7 days for stocked sizes. Send us the refrigerant, the I.D., the operating temperature and whether the run is suction or discharge, and we will tell you which structure fits.