50 in of doubled bar becomes 4 in of sleeve, and #14 and #18 cannot be lap spliced at all.
← Back to the calculatorLap splices are prohibited for those bars (25.5.1). Two code grades exist: Type 1 holds 1.25 fy, Type 2 holds the same and has to stretch before it breaks.
The pages that rank for this query are catalogs. They list four families of coupler, show a photo of each, and stop. None of them answer the two questions a person actually has: which one can I use on this bar, and how many do I need.
There is also a number that gets repeated and is wrong: that couplers save steel. They do, and the saving is about four dollars a joint, which does not pay for the sleeve. What you are buying is length, access, and the right to stop a pour at any elevation — that is the honest case, and it is the one this page makes.
Every coupler on the market answers the same question, and it is not “how strong” — they all hold 1.25 fy. The question is can the bar rotate. If it can, a threaded system is the cheapest and the fastest. If it cannot, because it is already cast into a wall or it is 30 ft long and boxed in by other trades, you are in bolted or grouted territory and the price goes up.
| System | Bar end prep | Bar has to rotate? | Where it gets used |
|---|---|---|---|
| Parallel thread | Saw cut, then thread or upset head | Yes — one bar spins into the sleeve | Columns and walls started off a deck |
| Taper thread | Saw cut, tapered thread | Yes | Same, common in precast plants |
| Bolted / shear-screw | Saw cut only, no thread | No | Repairs, existing bars, tie-ins, tight cages |
| Grouted sleeve | None — sleeve is cast into the form | No | Precast, wall-to-wall, starter bars out of a wall |
| Cold swaged | Saw cut, then pressed with a hydraulic tool | No | Shop and precast work |
The second question is bar end preparation, and it is where projects get into trouble. Threaded and swaged systems need a saw cut. A shear cut leaves the end out of square and the thread runs out before it reaches full depth: the joint still torques up, and it carries less than the bar. Every manufacturer says saw cut in the installation manual, and it is the first thing a special inspector checks.
ACI 318-19 does not grade couplers by brand or by family. It grades them by what they have to do, and there are only two grades.
| Type 1 | Type 2 | |
|---|---|---|
| Tension strength | ≥ 1.25 fy | ≥ 1.25 fy |
| Compression | ≥ 1.25 fy | ≥ 1.25 fy |
| Ductility | Not required | Required — the assembly has to stretch before it breaks |
| Where it is enough | Ordinary frames, slabs, walls | Special moment frames and anything the engineer calls ductile |
| Clause | ACI 318-19 25.5.7.1 | ACI 318-19 25.5.7.1, 18.2.7 |
Type 1 is what most building work gets, and it is stronger than the bar it joins: 1.25 fy against a bar that yields at 1.0 fy. Type 2 adds ductility, and that word is doing a lot of work — the splice has to deform before it fails, so a frame can absorb energy instead of snapping at a thread. Special moment frames require it (18.2.7), and so does anything an engineer calls ductile.
The specifying error to avoid is a note that says “mechanical splices permitted” with no type. That sentence gets read as Type 1 by the supplier and as Type 2 by the inspector, and the difference is not visible once the sleeve is on the bar. Write the type, and write the clause.
This is the comparison nobody publishes. A Class B lap is 1.3 ld, and ld depends on cover, spacing and concrete strength; at 4000 psi with ordinary cover the numbers come out as below. The last two columns are the ones that decide anything.
| Bar | Class B lap, 4000 psi | Extra bar in one lap | Bar saved, at $1,000/ton | Lap ÷ coupler |
|---|---|---|---|---|
| #6 | 30 in | 3.76 lb | $1.88 | 7.5× |
| #8 | 50 in | 11.13 lb | $5.56 | 12.5× |
| #10 | 64 in | 22.95 lb | $11.47 | 12.8× |
| #11 | 71 in | 31.44 lb | $15.72 | 14.2× |
Read the fourth column carefully. A #8 lap holds 11.13 lb of extra bar, which at $1,000 a ton is $5.56. A #8 parallel-thread coupler is $12 to $20. So the coupler is two to four times the cost of the steel it saves, and any page claiming couplers “save material cost” is either comparing against a lap that was never going to fit, or not doing the arithmetic.
What the coupler actually buys is the other four columns: 46 in of length on a #8, a bar that stays in one plane, and a joint that can be made at any elevation. On a wall or a column that is the whole argument, because the lap does not only cost steel — it costs the space where the second bar has to sit for over four feet, inside a cage that was detailed without it.
Couplers are counted per bar per splice level, which is not the same as per floor. Take eight columns, four #8 bars each, spliced at three elevations above the foundation:
couplers = columns × bars × splice levels = 8 × 4 × 3 = 96 lap length = 50 in per joint (Class B, #8, 4000 psi) bar in the laps = 96 × 11.13 lb = 1,068 lb = 0.53 ton bar value = 1,068 lb × $0.50/lb = $534 coupler cost = 96 × $15 = $1,440 difference = −$906 — the couplers cost more
That is the honest ledger, and it is why the decision is never made on material. What the $906 buys is 96 joints that are 4 in long instead of 50, which is 400 ft of doubled bar that never has to fit inside the column cage, and three pour breaks that can be set at whatever elevation the crew reaches rather than at the elevation the lap length dictates.
Two things change the arithmetic in favour of the coupler. Large bars: a #11 lap is 71 in and holds 31.44 lb, so the steel saving per joint goes to $15.72, and at that size the lap also starts to govern the bar cut list. And #14 or #18, where there is no legal lap at all, so the comparison is not money but whether the detail can be built.
A steel sleeve that joins two bars end to end and develops the bar’s strength through thread, bolts, swaging or grout instead of through a lap. It is a mechanical splice under ACI 318-19 25.5.7, and it comes in two code grades: Type 1, which has to hold 1.25 fy, and Type 2, which has to hold the same load and also stretch before it breaks.
Stronger, on paper. A Class B lap has to develop 1.3 ld, which lands somewhere around 1.0 to 1.3 fy depending on cover and spacing; a Type 1 or Type 2 coupler has to develop 1.25 fy with no dependence on cover, bar spacing or concrete strength. The catch is that the coupler’s strength is only there if the bar end was cut square and threaded properly — a coupler on a sheared end is weaker than the lap it replaced.
About 4 in for #6 and #8, and 5 to 6 in for #9 to #11 in a standard parallel-thread system. A transition coupler, joining two different bar sizes, is longer because it has two different thread depths. Set that against the lap it replaces: 30 in for a #6 and 50 in for a #8 at 4000 psi, so the coupler is roughly a twelfth of the length.
Bars per member × number of splice levels. Eight columns with four #8 bars each, spliced at three elevations above the foundation, is 8 × 4 × 3 = 96 couplers. Count splice levels from the placing drawings, not from the floor count: a lift that stops at a construction joint needs a splice whether or not there is a floor there.
Roughly $8 to $25 each depending on bar size and system, with the bolted and grouted types at the top of that range. The bar it saves is worth much less than that: a #8 Class B lap contains 11.13 lb of extra bar, which is about $6 at $1,000 a ton. So a coupler costs more than the steel it replaces, and the case for it is length, access and pour sequence, not material cost.
Yes, but they have to be Type 2, and ACI 318-19 18.2.7 says so for special moment frames. That is the whole difference between the two grades: Type 1 holds 1.25 fy and nothing more is asked of it, while Type 2 also has to show ductility, so the assembly stretches rather than snapping at the thread. Specifying “mechanical splices” without the type is how a project ends up with the wrong one.
A torque wrench, usually a click-type, set to the manufacturer’s value for that coupler, and the inspector normally wants a torque log per joint. That log is the difference in paperwork between a coupler and a lap: a lap has no installation record at all, because there is nothing to torque. On a job with a special inspector, the log is the submittal.
Yes — that is a transition coupler, and it is one of the few ways to step a column down in size without a lap. The rules are stricter than for a same-size splice: the coupler is designed for the larger bar, and the development of the smaller bar into it has to be checked. Do not assume a #8-to-#6 transition behaves like a #8-to-#8.
No. Threaded and swaged systems need a saw cut, because a shear leaves the end out of square and the thread runs out before it reaches full depth — the joint then torques up but carries less than the bar. Bolted and grouted sleeves are more forgiving of the cut, which is one reason they get used for repairs on existing bars. Every manufacturer’s installation manual says saw cut, and it is the first thing an inspector looks at.
You have to. ACI 318-19 25.5.1 prohibits lap splices for #14 and #18 bars in tension, so a mechanical splice or end bearing is the only way to continue them. That single clause is why large-bar columns and heavy foundation work are coupler jobs by default rather than by choice.
Not on the invoice — but it is cheaper in exactly one line item, and it is worth knowing how big that line is. Take four #8 bars, three 20 ft lifts, so two splice levels. On laps each bar line is 60 ft of bar plus two 50 in laps = 68.33 ft; on couplers it is 60 ft, because the sleeve adds no bar length. That is 8.33 ft = 22.2 lb less per line, 88.8 lb for the column, about $44 at $1,000 a ton — a 12.2% cut in the bar order. The eight sleeves that bought it cost $96 to $160. So the bar order falls by an eighth and the invoice rises by $52 to $116. The saving is real and it is never the reason to use one.
Only if the bar was bought for it. Welding is a legal splice, but it is not a splice you can make on any bar: ASTM A706 is the weldable grade, with chemistry controlled for it and a single Grade 60, and its welding is done to AWS D1.4 by a qualified welder under a qualified procedure. ASTM A615 — the bar on most residential and commercial jobs — has no chemistry control for welding, so welding it needs a written qualification before anyone strikes an arc. The trade is the same one a coupler makes and the arithmetic is the same: a weld adds no length, so it saves the same 8.33 ft a line. What it also needs is a welder on site, a procedure, and a bar order that specified A706 in the first place. A coupler needs a saw, a torque wrench and a log.