Asphalt paving joints and handwork: longitudinal, transverse, and seam density

Joints and handwork, and why the seam is the weak point

A paving joint is a seam where two mats meet. The longitudinal joint runs the length of the road between adjacent passes or lanes. The transverse joint crosses the road where a paving run starts or stops. Handwork is everything the paver cannot reach: the tie-ins around manholes and valves, the edges, the curb line, the spots where a lute and a shovel do what the screed could not.

Drive any old road and look at where it is breaking up first. It is the seam down the middle and the edges, not the field of the mat between them. That is not bad luck. The joint is the one place on the job where the mat has an open, unconfined edge during compaction, so it is the hardest place to drive density, and low density is the start of every asphalt failure. Water gets into the lean seam, the binder oxidizes, and the joint ravels and cracks years before the rest of the surface shows its age.

So the whole guide is about that seam and that handwork. How the longitudinal joint is built hot or cold, how the transverse joint is cut and butted, what the joint density spec wants, and where the rake makes it better or quietly makes it worse. The rolling and the density acceptance behind all of it live in the compaction window guide. This one is about the geometry and the technique of the joint itself.

Why do asphalt joints fail first?

Asphalt joints fail first because the joint is the hardest place to reach density, and low density is where pavement failure starts. In the field of the mat the roller works material that is confined on both sides, so the stones have nowhere to go but down and together. At the joint the mat has an open edge with nothing beside it to push against, so under the roller that edge spreads sideways instead of compacting, and it ends up lean.

Lean means high air voids. Once the air voids at the seam connect into channels, water and air move through the mat instead of running off it. Then the binder strips and oxidizes at the joint, the surface ravels, the raveled groove holds more water, and the cycle feeds itself. A well-built joint runs only a point or two below the surrounding mat. A bad one runs five to ten points low, and that is the joint you watch open up in two or three winters.

The transverse joint fails for a different reason on top of density: ride. A transverse joint built as a bump or a dip is the jolt every driver feels at the same spot, and the impact loading from tires hitting that bump pounds the seam apart on top of whatever density problem it already had. The longitudinal joint is a density problem. The transverse joint is a density problem and a smoothness problem at once.

The longitudinal joint: cold side and hot side

The longitudinal joint is the seam between two paving passes laid side by side, running with traffic. Every road wider than one paver pass has one, and on a multilane road the joints ideally land on the lane lines, not in the wheel paths. The joint has two sides, and they do not compact the same.

The confined side, the hot side, is where fresh hot mix is laid against an already-placed lane. That fresh mat has the cold lane beside it to push against, so the roller can drive it tight, and the hot side of a joint commonly comes out a couple of points denser than the cold side. The unconfined side, the cold side, is the outside edge of the first pass when it was laid with nothing beside it. That edge spread under the roller and ran lean, and now it is cold and stiff when the second pass butts against it.

This is the split that decides joint strategy. If you can keep both sides hot when the joint is rolled, you get the best seam there is. If the second lane goes down a day later against a cold, cured edge, you are building a cold joint and fighting for every point of density across it. The methods that follow are all about confining that open edge, or making both sides hot, or treating the cold edge so the new mat still bonds and densifies across it.

How do you get density at the longitudinal joint?

You get density at the longitudinal joint by confining the open edge and by rolling the joint from the hot side. The overlap is the first move. The second pass laps onto the cold mat by roughly 1 to 1.5 in, leaving a small ridge of new material standing above the joint so there is enough mix to compress down into the seam instead of starving it. Too little overlap and the joint is lean. Too much and you trap a thick, fat strip of luted material that never compacts right either.

Rolling the joint from the hot side is the technique that separates a built joint from a hoped-for one. The breakdown roller works the hot lane first and is run with the drum hanging the joint, biting an inch or two over onto the cold side or holding a few inches off it depending on the spec and the crew's pattern, so it pinches the new material into the seam from the hot mat that can still move. Roll the joint cold and from the wrong side and you iron the lean edge flat without ever closing the voids under it.

An edge restraint or pinch device on the breakdown roller presses the unconfined edge of the first pass while it is still hot, so it cannot spread in the first place. That is the cheapest density at the joint you will ever buy, and it happens on the lane that gets paved first, hours before the joint is even made. The detailed rolling pattern, the pass count, and the gauge-versus-core density check are in the compaction window guide. What matters here is that the joint gets its own deliberate rolling, not whatever the mat rolling happens to leave at the edge.

The notched-wedge joint

The notched-wedge joint is a tapered longitudinal joint shaped by a device on the first pass instead of a vertical edge. Rather than leaving a square, near-vertical drop at the edge of the first lane, a notched-wedge shoe or restrictor on the screed forms a gentle ramp from the top of the mat down to the base, with a small vertical notch at the top and another at the bottom. The common geometry runs a taper about 12 in wide with vertical notches on the order of 0.5 to 0.75 in at the top and bottom, and it is typically called for on lifts thicker than roughly 1.5 to 1.8 in, but the dimensions and the trigger thickness come from the agency spec, so confirm them.

The taper buys two things. It confines the edge of the first pass so it does not spread and run lean, which raises the density of that unconfined edge under the roller. Then the second pass laps up the ramp, and because the joint is sloped rather than a square butt, the new mat seats against more surface and the roller can drive the seam tighter. Research on the notched wedge generally finds it can hold higher joint density than a plain butt joint, largely from the confinement the wedge gives the edge under the drum.

The notch at the top is there so the taper does not feather to a thin, ragged edge that ravels on its own and so there is a clean line for the next pass to match. The bottom notch keeps the thin toe of the wedge from being a zero-thickness feather that crumbles. Pull the wedge too flat and you have a long thin feather that traffic chews off. Make it too steep and you are back to a butt joint with extra steps. The shoe is set to the spec geometry and checked, not eyeballed.

Notched-wedge joint

A longitudinal joint with a tapered edge and small top and bottom notches, formed by a screed shoe to confine and densify the seam

Joint maker / restrictor

The shoe or device on the screed edge that shapes the wedge or pinches the unconfined edge of the first pass

What is echelon paving and the hot joint?

Echelon paving is two pavers running side by side, one staggered slightly behind the other, laying adjacent lanes at nearly the same time so the longitudinal joint is made hot against hot. It builds the best joint there is, because neither side of the seam has cooled and the rollers can pass directly over the joint while both mats are still workable. A true hot joint can come out with no visible seam, the joint region as dense as the mat around it.

The reason it works is the reason the cold joint fails. There is no cold, stiff edge to butt against. The second mat goes down hot beside a first mat that is still hot, the two mix faces knit together under the roller, and the unconfined-edge problem mostly disappears because the edge gets compacted while it can still move. The hot lane laps the still-hot first lane by about an inch, then the roller works the seam from the hot side with material that has not set.

The catch is logistics, which is why you do not see it everywhere. Echelon paving needs two pavers, the mix supply to feed both, and enough closed width to run them together, so it fits new construction and wide reconstruction more than it fits a single-lane overlay under traffic. Agencies that can run it, like several DOTs that have pushed joint density hard, get excellent joints from it. When you cannot pave in echelon, you are building a cold joint, and the rest of the longitudinal-joint methods are how you make a cold joint behave.

The cold longitudinal joint: paving against a cured edge

A cold longitudinal joint is the case where the second lane is paved a day later, or hours later, against a first lane that has already cooled and set. This is the normal case on overlays and lane-at-a-time work under traffic, and it is where joints go bad. The first lane's outside edge is the unconfined edge that ran lean when it was placed, and now it is stiff. The new mat has to bond to that cold vertical or tapered face and still reach density across it.

Tack the cold face before you pave against it. The vertical edge of the cold lane is a cold asphalt surface, and the new mat will not bond to it without tack any more than an overlay bonds to a milled surface without it. A light, complete coat of emulsion on the exposed joint face, brushed or sprayed, glues the new lane to the old one instead of leaving a dry, unbonded seam that water walks straight into. The tack-coat material, residual rate, and break behavior are covered in the mill and overlay guide. At the joint the point is simpler: do not pave a cold lane against a dry edge.

Set the overlap and handle the bump back. The hot lane laps onto the cold mat by about 1 to 1.5 in, leaving a ridge of new material proud of the joint. Some crews knock that ridge back onto the hot side with a lute, the bump back, so the roller compresses material into the seam rather than over the cold lane, while others roll the overlap in place. Both are defensible and the spec or the crew's tested pattern decides. What loses every time is no overlap and no tack, a cold lane jammed dry against a cold edge, rolled flat, and called a joint.

Joint adhesive, void-reducing membrane, and saw-and-seal

There are material fixes for the longitudinal joint, not just technique. Joint adhesive is a hot-applied polymer-modified asphalt run down the vertical face of the cold lane before the next mat is placed, so the new lane is glued to the old edge with a heavy, flexible bond rather than a thin tack. It seals the face and gives the seam something that flexes instead of cracking when the joint works under load and temperature.

A void-reducing asphalt membrane, sold under names like VRAM or a longitudinal joint sealant, is the newer materials approach and it works differently. It is a highly polymer-modified binder placed hot, around 300°F, in a band on the existing or milled surface directly under where the longitudinal joint will be, before paving. As the mat goes down and compacts, the membrane migrates up into the bottom of the joint and fills a large share of the air voids, commonly cited around 50 to 70 percent, cutting permeability at the one place water always finds. It is a way to win density at the joint through the material when the geometry alone will not get there.

Saw-and-seal, sometimes called the cutback joint, takes the opposite approach: instead of preventing the crack, you control where it forms. After paving, a saw cuts a clean groove down the line of the joint and the groove is filled with a flexible sealant, so the inevitable crack opens at the saw line and stays sealed rather than raveling open as a ragged seam. It shows up over jointed concrete and on agency work where reflective or thermal cracking at the joint is the known risk. None of these replaces building the joint right. They are insurance on the seam that is hardest to compact, and the spec decides which one, if any, the project pays for.

Joint adhesive

A hot-applied polymer-modified asphalt run on the cold joint face to bond the new lane to the old edge

VRAM / longitudinal joint sealant

A void-reducing membrane placed under the joint before paving; it migrates up to fill voids and cut permeability

Saw-and-seal / cutback joint

Sawing a groove on the joint line after paving and sealing it, so the crack forms and stays sealed at a controlled line

The transverse joint: butt joint vs tapered joint

The transverse joint is the seam across the mat where a paving run stops at the end of a day or a shift, and where the next run starts against it. It is also the joint where a new overlay meets existing pavement at the project limits. There are two ways to leave and rejoin a mat at a transverse joint, and only one of them rides right.

The butt joint is a clean vertical face. The mat is ended on a full-thickness vertical edge, and the next day's run butts hot mix square against that face. Done right it carries the grade straight across the seam with no bump and no dip, and it is the joint highway work is built around. The catch is that the paver cannot create a clean vertical full-thickness edge on its own at shutdown, so the end of the run has to be formed or cut.

The tapered joint, where the mat is feathered down to nothing in a thin ramp at the end of the run, is the easy way out and it is usually the wrong one. A feathered transverse joint is thin, lean mix that will not compact, so it ravels off, and it almost always rides as a bump or a dip because you cannot hold grade in a feather. You see tapered joints used as a temporary ramp to carry traffic off the end of a run overnight, but that ramp gets cut back to a vertical face before the next mat goes on. A feather left as the permanent joint is the bump every driver remembers.

How do you build a clean transverse joint?

You build a clean transverse joint by ending the mat on a vertical full-thickness edge and butting the next run against it on grade. The classic field method is to run the paver off onto a board or a strip of heavy paper at shutdown so the mat ends on a clean vertical face at full thickness rather than tailing off thin. The mix beyond the board is wasted and removed. The alternative, and the one most specs want for the permanent joint, is to saw-cut the cooled mat back to a straight vertical face and remove everything past the cut, which gives a true full-depth edge to start against.

Tack the cut face before the next run, same as any cold joint. Then the paver starts against the vertical edge, and the screed is set so the new mat comes off even with or slightly above the existing mat at the seam, commonly within about 1/8 in high, never low. A joint that starts low is a dip you cannot fix by rolling, and a joint that starts high gets rolled down flush.

The bump test is how you check it, and the inspector will. Lay a 10 ft straightedge across the joint, perpendicular to the seam and centered on it, and read the gap. A common acceptance is no more than 1/4 in of deviation under the 10 ft straightedge, and if the joint reads worse than that the spec says to correct it, by diamond grinding a high joint or addressing a low one, before it is accepted. Check the joint with the straightedge while the crew is still there, because the bump you find at the straightedge is cheap to fix and the bump the riding public finds is not.

The joint density spec and testing

Many agencies now hold the longitudinal joint to its own density requirement, separate from the mat, because everyone knows the seam runs lean and a mat-only spec lets a bad joint hide. Nearly half of state highway agencies have adopted some form of joint density specification, and the common shape is either a minimum joint density as a percent of theoretical maximum, or a maximum allowed difference between the joint and the adjoining mat.

The numbers cluster, but the spec governs. A well-built longitudinal joint runs about 2 percent below the adjoining mat density, and that 2 percent differential is what several agencies write as the acceptable limit. Minimum joint density requirements commonly land somewhere around 89 to 91 percent of theoretical maximum depending on the agency, with notices of non-compliance or pay reductions below that. Joint cores are typically cut at a few per lot, taken right at the seam, on a notched-wedge joint commonly a set distance such as 6 in from the upper notch, at intervals the spec sets. Confirm the differential, the minimum, and the coring location against the project documents.

The density methods themselves, the percent-of-Gmm framework, the nuclear and non-nuclear gauges, and the way cores settle a dispute, are all covered in the compaction window guide and apply to the joint exactly as they apply to the mat. The difference at the joint is only where the core comes from and the lower number it is allowed to read. The reason the spec exists is money: agencies put a pay line on the joint because the joint is where the pavement they are buying actually fails.

Handwork and raking: where the paver cannot reach

Handwork is the asphalt placed and shaped by hand where the paver and screed cannot go: around manholes and valve boxes, into tight tie-ins, along curbs and against buildings, at the ends of cul-de-sacs, and wherever the geometry beats the machine. The tools are the lute, also called the rake, which is the wide-headed asphalt rake a hand crew uses to move and level mix, the asphalt shovel, and a hand tamp or a plate compactor for the spots a roller cannot reach.

Handwork is also where a good mat gets a bad reputation, because hand-placed asphalt is harder to keep uniform than machine-placed. The screed lays mix at a controlled depth and texture. A rake drags it, and dragging mix is the fastest way to pull the coarse stone out of it. So the rule is to place the mix close to grade first, with a shovel, and then use the lute to true it up with as little dragging as possible, rather than spreading a pile across the area by raking it out.

The good hand crew works the mix while it is hot and works it as little as it can get away with. They drop it where it goes, knock it to grade, lute it true, and get a tamp or a plate on it before it cools. The crew that pushes a cold pile around with a rake for ten minutes is building a segregated, low-density patch that will be the first thing to fail, no matter how good the machine mat beside it is.

What causes segregation, and how do you fix it in handwork?

Segregation is the non-uniform spread of coarse and fine aggregate in the mat, and it comes in two kinds. Physical segregation is the coarse stone separating from the fine, which happens when mix is handled and dropped, in the truck, through the paver, and especially under a rake. Thermal segregation is cold lumps of mix, colder than the mat around them, that will not compact to the same density even though the gradation is fine. Both read the same way in the finished surface: streaks and patches of open, coarse-looking texture that ravel early because they never reached density.

Handwork makes physical segregation worse than almost anything else on the job. When you drag mix with a lute, the fine material and the binder flow and the coarse stone rolls ahead of the rake head, so you end up broadcasting a pocket of clean coarse aggregate, then covering it with fines as the rake passes back over it. The pocket is invisible when you walk away because the fines hide it, and it shows up later as a raveled hole right where the handwork was. Raking operations are a leading cause of segregated coarse pockets for exactly this reason.

The fix is technique, not a product. Place the mix at grade with a shovel instead of raking a pile out across the area, so the stone is not dragged. Keep the lute work to truing the surface, pushing the mix as little as possible and never broadcasting it. If a pocket of coarse stone does pull out, do not just rake fines over it, dig it out and replace it with fresh mix. Work hot and tamp it before it cools, because a cold lump cannot be raked back into a uniform mat. The broadcast move, throwing mix from a shovel to spread it thin, is the classic way to segregate a tie-in, and the good rakers do not do it.

Luting and the do-not-overwork rule

Luting is truing the surface of fresh mix with the lute before it is compacted, and the governing rule is to do as little of it as the work allows. Every pass of the rake over hot mix pulls the surface apart a little, separating fines from stone and dragging the binder-rich top, so the surface you overwork is the surface that segregates and ravels.

The skill is knowing when to stop. A raker who sets the grade in a couple of passes and walks away leaves a tighter, more uniform surface than one who keeps fussing with it. Working the mat too much, chasing a perfect look by raking and re-raking, tears the surface texture open and leaves a fat, fines-rich, low-stone skin that looks smooth and fails fast.

Lute it to grade, check it, and leave it. The roller does the compaction, not the rake. The rake's only job is to get the mix to the right elevation and shape before the roller arrives, and the less it has to drag to do that, the better the patch holds up.

Raking and compacting around manholes and valves

Castings are the most common handwork on a paving job and the most common place a fresh mat fails early. The paver runs around the manhole, valve box, or inlet, and the ring of asphalt the screed could not place gets done by hand, then compacted by hand because the big roller cannot work tight to the frame. That ring is a confined, fussy spot with a hard vertical edge against the casting, and it is exactly the kind of place that ends up low on density.

Tack the casting edge and the cold mat edge around it, place the mix by hand without dragging it into segregation, lute it to match the surrounding grade and the casting elevation, and then compact it. The compaction is the part people shortchange. A hand tamp or, better, a vibratory plate compactor gets density into the ring around the frame that the roller cannot reach, and a small roller can work the area outside the immediate ring. Skip the plate and hand-tamp it lightly and you have a lean, permeable collar around every casting that ravels and lets water down alongside the frame.

The casting also has to come up to the finished grade so the surface runs flush across it, which is set before paving by adjusting or raising the frame, the same as on a mill and overlay. A frame left low is a sunken collar that gets pounded by every tire, and the handwork around a low frame ravels twice as fast from the impact. Get the frame to grade, then do the handwork around it like it matters, because it is the spot the owner will point at.

The unconfined edge and the safety edge

Every mat has an outside edge with nothing beside it, and that unconfined edge is the same density problem as the longitudinal joint, just on the outside of the work instead of the middle. Under the roller the free edge wants to spread sideways and run lean, so it has to be deliberately compacted, commonly by rolling with the drum hanging a few inches over the edge in a pinch pass, or with an edge-restraint device on the first pass, so the edge gets driven down instead of pushed out.

On a curbed road or lot the edge is confined by the curb and the problem is smaller. On an open shoulder or a lot edge with no curb, the edge is fully free and it is both a density problem and a safety problem. A square, vertical drop-off at the pavement edge is a wheel trap that can catch a tire and pull a vehicle off the road.

That is where the safety edge comes in. Instead of a vertical drop, a device on the screed shapes the unconfined edge into a compacted taper, commonly around a 30-degree slope, so a tire that drifts off can climb back on without grabbing. The safety edge gets the edge compacted at the same time it shapes it, so it solves the density and the drop-off together. Whether it is required is an agency call, but on a high-speed unconfined edge it is good practice regardless of who is watching.

Joints and handwork at the curb and gutter

Where the mat meets curb and gutter, the joint is between asphalt and concrete and it is mostly handwork. The paver cannot lay tight to the curb, so the strip along the gutter line gets placed and luted by hand, and it has the same vertical cold edge against the concrete that a casting does. Tack the concrete face, place the mix without dragging it, and tamp or plate the strip, because the roller cannot get tight to the curb either.

The grade at the gutter is the part that has to be right. The asphalt has to meet the gutter line so water still runs to the inlets and does not pond against the curb, which usually means a slight wedge or feather at the very edge to match the existing gutter elevation. Bury the gutter under too much asphalt and you dam the water against the curb and flood the lane. Leave the asphalt low against the gutter and you get a lip that traps water and ravels.

On a resurfacing job the gutter wedge and the tie-in geometry are handled the way the mill and overlay guide lays out, matching the new surface back to the existing concrete and the drainage. New construction is easier because the grades are set, but the handwork along the curb is the same: a hand-placed, hand-compacted strip that has to hold density and hold grade at the same time, and it is a strip that gets shortchanged on both when the crew is in a hurry to chase the paver.

The screed setup and matching the next pass

The longitudinal joint is set up before either lane is rolled, at the screed. When the second pass is paved against the first, the screed has to be set so the new mat comes off at the right elevation to match the cold lane after both are compacted, accounting for the roll-down the new mat will take. Set the screed flush to the cold mat and the new lane finishes low after it compacts, leaving a lip. Set it a touch high, by the amount the mat will roll down, and the two finish flush.

The overlap is a screed-and-handwork detail too. The end gate and the edge of the screed are set to lay the new mat lapping onto the cold lane by about an inch, and a raker rides the joint behind the screed to keep that overlap even and to handle the small ridge of luted material at the seam. A wandering overlap, fat in one spot and starved in the next, makes a joint that is dense in patches and lean in patches.

On a notched-wedge joint the shoe on the screed forms the wedge on the first pass, so the joint geometry is built into the laydown, not added later. Either way the rule holds: the joint is won at the screed and the roller before it is ever a finished seam, and a joint set up wrong at the screed cannot be raked or rolled into a good one afterward.

Planning where the joints fall

Where the joints land is a planning decision, and it pays to make it on purpose. The longitudinal joint should fall on a lane line, not in a wheel path, because the wheel path is where the heaviest, most repeated loading hits, and putting the weakest seam under the heaviest load is asking for early failure. On a normal two-lane road that means the joint sits on the centerline or the lane line, where the tires straddle it rather than track on it.

Stagger the joints between lifts. The longitudinal joint in the surface course should not stack directly over the joint in the lift below it, because two lean seams on top of each other make a full-depth weak line. Offset them by several inches to a foot so the joints in successive lifts do not line up. The same goes for transverse joints between lifts, which get offset by a few feet so the day-joints do not stack into one bump.

Transverse joints get planned around the work, not just dropped wherever the day ends. You want them clear of intersections, crosswalks, and the spots where stopping and turning traffic pounds them hardest, and you want them where the next run can start clean. A little thought about where to stop a run saves a transverse joint in a bad spot, and a bad transverse joint is permanent until somebody mills it out.

What does the inspector check at the joint?

The inspector checks the joint harder than the mat, because the joint is where the job fails. On the longitudinal joint the first thing is density, by core or correlated gauge right at the seam, against whatever joint density requirement the spec carries, the minimum percent or the differential from the mat. Then the look of the joint: a tight, closed seam at uniform texture, not an open, raveling, fat-luted line. An overworked, segregated joint is visible, and a good inspector reads it by eye before the cores confirm it.

On the transverse joint the inspector pulls a 10 ft straightedge across the seam and reads the bump, looking for no more than about 1/4 in of deviation, and checks that the joint was cut to a vertical face and butted rather than feathered. The ride at the transverse joint is an acceptance item on most agency work, so a joint that fails the straightedge gets ground or corrected.

On the handwork the inspector looks at the tie-ins around structures and along curbs for segregation, for the coarse raveling pockets a rake leaves, and for the lean collar around a casting that never got plated. They check that the castings came to grade and that the edge got compacted. The joint and the handwork are the parts of the job a careful inspector spends the most time on, because they are the parts that tell on a crew first, and the crew that knows what the inspector checks builds the joint to pass it the first time.

Commercial lot joints vs highway joints

The physics of the joint is the same on a parking lot and a highway, but what governs it and what gets tested are not. On agency highway work the joint carries a written density spec, joint cores, and a ride tolerance with pay attached, and the crew builds to those numbers because the money rides on them. The joints land on lane lines, the transverse joints get sawed and butted, and the whole thing is documented lot by lot.

On a commercial lot there is usually no joint density spec and no coring, so the joint is only as good as the crew chooses to make it. That cuts both ways. There is no inspector forcing the issue, so a hurried crew leaves cold, lean, feathered joints that ravel in a few seasons and become the callback. The crew that builds the lot joint to highway practice anyway, overlap, tack the cold edge, roll from the hot side, butt the transverse joints, is the crew whose lots are not back open for repair while the competition's are.

The lot also has more handwork per square foot than a highway: more castings, more curb line, more tight corners, more tie-ins to drives and aprons. So on a lot the handwork quality matters even more to how the job ages, and the segregation and structure-collar problems are where lot pavements actually start to go. The owner does not see a density core. They see the raveling around the catch basin and the open seam down the drive lane, and that is the joint and the handwork talking.

What to document

The record on joints and handwork is what defends the work when a seam ravels or a core comes back light, and it is thinner than it should be on most jobs because the joint gets less attention than the mat. Write down the joint type and where it falls, the tack or treatment on the joint face, the density or straightedge result, and what handwork was done where, by station or by area.

Capture the longitudinal joint method, hot echelon, notched wedge, or cold joint, and the tack or joint adhesive or membrane used on the face. Capture the transverse joints, whether they were sawed and butted, and the straightedge bump reading. Capture the joint density cores against the spec joint requirement. And capture the handwork: the castings brought to grade and plated, the curb and gutter tie-ins, and any spot where a coarse pocket was dug out and replaced. A note made at the time about a joint that ran cold or a casting collar that fought density is worth more than a memory at the callback. Holding the joint types, tack, density, and handwork together in a record like FieldOS keeps the as-built tied to the seam it covers.

Common mistakes

• Paving a cold lane against a dry, untacked joint face, so the new mat never bonds to the old edge.
• Skipping the overlap, or running it fat and wandering, so the joint is starved in spots and lean across the seam.
• Rolling the longitudinal joint cold and from the wrong side instead of pinching it from the hot mat.
• Feathering the transverse joint into a thin ramp instead of cutting a vertical face and butting against it.
• Leaving a transverse joint that reads a bump or a dip over a 10 ft straightedge and calling it done.
• Overworking the mat with the lute, dragging the surface open and broadcasting coarse aggregate in the handwork.
• Raking fines over a segregated coarse pocket instead of digging it out and replacing it with fresh mix.
• Hand-tamping the collar around a casting lightly instead of plate-compacting it, leaving a lean, permeable ring.
• Stacking the joints in successive lifts directly over each other, or running the longitudinal joint in the wheel path.
• Leaving an unconfined edge or a square drop-off uncompacted and unshaped on an open shoulder or lot edge.

Field checklist

Standards and references

What governs the joint is the project specification and the state DOT or owner agency. The longitudinal joint density requirement, whether a minimum percent of theoretical maximum or a differential from the mat, the joint coring rate and location, the notched-wedge geometry, the transverse joint and straightedge tolerance, and the tack and structure-adjustment items all live in the agency spec, and nearly half of state agencies now carry a joint density spec. The numbers in this guide, the 2 percent differential, the roughly 89 to 91 percent joint minimum, the 1 to 1.5 in overlap, the 1/4 in straightedge, and the notched-wedge dimensions, are the commonly cited ranges, not the contract values. Confirm them against the project documents and the adopted specification before you build to any single figure.

The Asphalt Institute and the National Asphalt Pavement Association are the trade sources for joint best practice. The Asphalt Institute's MS-22, Construction of Quality Asphalt Pavements, covers placement, joints, and handwork, and FHWA and NAPA have published longitudinal-joint best-practice tech briefs that lay out the notched wedge, echelon paving, the hot-versus-cold joint, and the materials approaches. For the density methods behind the joint cores, the percent-of-Gmm framework, the test methods, and the gauge-versus-core reconciliation are covered in the compaction window guide and rest on the AASHTO and ASTM methods cited there.

Airfield work follows its own joint and density requirements under the FAA advisory circulars rather than the state DOT spec, and the joint density limits there can be tighter. Wherever you are, the controlling document is the one the owner adopted, and the section numbers and editions shift on a cycle, so verify them against the adopted version before citing a clause on a submittal.

Units and terms

The joint gets described across a few names and unit systems, so the same seam reads differently across a plan set, a spec, and a density report. Density at the joint is a percent of Gmm, the theoretical maximum specific gravity, the same reference as the mat, often written as a minimum or as a differential of about 2 percent below the mat. Overlap, notch depth, and lift thickness are in inches, with the notched-wedge taper commonly about 12 in wide. Tack is a residual rate in gallons per square yard. The straightedge tolerance is in inches under a 10 ft straightedge.

The terms carry the meaning. A longitudinal joint runs with traffic between adjacent passes; a transverse joint crosses it where a run starts or stops. The hot side or confined side of a joint is paved against an existing lane; the cold side or unconfined edge was the free edge of the first pass. A hot joint is made hot against hot, as in echelon paving; a cold joint is paved against a cured edge. Handwork is hand-placed asphalt, luting is truing the surface with the lute or rake, and segregation is the coarse and fine aggregate separating, which the rake causes when it drags the mix.

Longitudinal joint

The seam running with traffic between two adjacent paving passes or lanes

Transverse joint

The seam across the mat where a paving run starts or stops, or meets existing pavement

Hot joint / cold joint

A joint made hot against hot (echelon) versus a new lane paved against a cooled, cured edge

Unconfined edge

The free edge of a pass with nothing beside it, which spreads and runs lean under the roller

Handwork / luting

Hand-placed asphalt where the paver cannot reach, and truing its surface with the lute or rake

Segregation

Non-uniform separation of coarse and fine aggregate (physical) or cold lumps in the mat (thermal)

Related tools

Calculators and readiness checks for this work

Compare your options

FAQ

People also ask

• What temperature do you stop rolling asphalt?
• What is a mill and overlay?
• Why does the longitudinal joint fail first?
• Why is asphalt density important?
• Why does asphalt density matter?