Trusses: Method of Joints
Every joint of a truss is a particle in equilibrium.
Before this Lesson
Look up at a steel truss bridge: dozens of bars, and a train rolling across. Finding the force in every bar sounds like dozens of unknowns in dozens of equations, all at once. It isn't. Each bar can only push or pull along its own length, and that one fact breaks the whole bridge into small problems, one joint at a time, each no harder than the ring that held up a sign. Stranger still, under a given load some of those bars often carry exactly nothing, and you can point to them before you pick up a calculator.
A bar that can only push or pull
A truss is a set of straight bars, called members, pinned together at their ends, with the loads applied only at the pins, which are called joints. A real bridge has bolted joints and bars with weight, but this pinned model is the one you solve. The Handbook never lists these assumptions, so carry them in yourself.
Now pull one member out of the truss and ask what touches it. Only its two pins. Each pin pushes on it with some force, in some direction nobody has told you. Two forces, and the bar sits still, so makes them equal and opposite. That isn't quite enough, because equal and opposite forces on offset lines are a couple, and the bar would spin. So take moments about one pin. That pin's force has no arm there, which means the other force can't have an arm either. Its line has to run through the first pin.
That's a two-force member: two forces, equal, opposite, and on the line joining its pins. The Handbook's Concurrent Forces entry says the same of any two-force body. So each member brings just one unknown, how hard it acts along its axis. Pulled outward, it's in tension. Squeezed, it's in compression.
Being straight isn't what earns the name. Having forces at only two points is. A bent bar pinned at both ends, with nothing in between, carries its force along the straight line between its pins, not along its shape. A straight bar with a load at its middle, or a third pin, has three forces on it and isn't a two-force member at all. It bends, and those members wait for Frames and Machines. That's why a truss keeps its loads at the joints.
Every joint is a particle
- From Equilibrium of a Particle: isolate, draw, sum to zero. Forces meeting at a point give two equations and settle at most two unknowns.
- From Beam Reactions: moments about one support give the reaction at the other in one line.
- From Forces and Resultants: a component is the force times the matching leg over the hypotenuse.
Every member's force runs along its axis, so all the forces at a joint pass through its pin. A joint is a ring from the sign problem, nothing more. You get two equations and can settle two unknowns. The Handbook's Plane Truss: Method of Joints entry prints them as and , horizontal and vertical, which are the and used here. It says nothing about where to start or how to keep your signs.
The components are easy, because the member is its own slope triangle: a member that rises 3 over a run of 4 has its force split in the ratios and . It's the ratio that goes in, so a member 5 ft long and one 50 ft long at that slope split the same way. The member's length never appears on its own.
For signs, there's no need to guess. A member in tension pulls on the pins at both of its ends, so draw every unknown member force pulling away from its joint. Then the algebra reports back: a positive answer is tension, and a negative one is compression. When you carry a known member to its next joint, keep the rule. Draw it pulling away from that joint too, and put in the number with its sign.
Here's the whole routine:
- Treat the whole truss as one body and find the reactions.
- Go to a joint with at most two unknown members, usually at a support.
- Draw everything on it: reaction or load, known members, and unknown members pulling away.
- Sum forces in whichever direction sees only one unknown, then the other direction.
- Step to a neighbor that's now down to two unknowns, and repeat.
This is the way to go for a member near a support, or for many members. For one member deep in the middle, walking there is the slow way, and Trusses: Method of Sections cuts straight to it.
A load that pushes sideways
Wind on a roof pushes sideways at a top-chord joint, and two things change. First, the pin has to take that load, because a roller on a level surface can't push sideways. So is no longer zero, and it goes into joint A's horizontal equation alongside the members.
Second, and easier to miss, that load has an arm. A sideways load on a beam runs along its centerline, right through the supports, so it never enters a moment equation. A top-chord joint sits a full truss height above the supports, so the load does enter, and it shifts the vertical reactions too.
Slopes from the rise and span
A small roof truss is two rafters and a tie, and its figure rarely gives you an angle. It gives the span and the rise. With the peak at midspan, each rafter's run is half the span. A 30 ft span with an 8 ft rise means a run of 15 and a length of , so the ratios are and . Move the peak off center, and each rafter gets its own run, and the reactions stop being equal.
Start at the free end
That bracket gave away a shortcut: it never needed its reactions. On a truss built out from a wall, you can't even get them first. Two pins at the wall mean four reaction components, and the whole truss gives only three equations.
So look at the other end. The tip joint, where the load hangs, has just two members, and it knows nothing about the wall. Start there, work inward one joint at a time, and stop when you reach the member you want. You'll never write a reaction.
Members that carry nothing
That wasn't luck. Look at joint C again: three members, two of them in a straight line, and nothing else on it. Sum forces perpendicular to that line. The two collinear members have no part in that direction, so the third is alone in the equation, with nothing to balance it. It has to be zero. That's the collinear rule.
The two-member rule is the same move. Take a joint with two members, not in line, and nothing else on it. Sum forces perpendicular to one of them. Only the other shows up, so it's zero. Then swap, and the first is zero too.
Such a member is a zero-force member, and the Handbook prints both rules under Statically Determinate Truss. Each comes with the condition that matters: no external load and no support reaction at the joint. A load or a reaction gives the perpendicular equation something to balance, and the rule is off.
That failure is useful, though. Hang a load on a joint with two collinear members and a third. The perpendicular equation now holds just two things, the load and the third member, so one equation finds that member. For a vertical hanger above a bottom-chord joint, the force is simply , in tension. You need no reactions and no other joints.
One zero uncovers the next
Here's where the rules get their reach. A zero-force member pulls on nothing, so once you've found one, cross it out and look at the joint at its other end. A joint that had three members may now have two, and a joint that had four may now have three with two in line. A rule that didn't apply a moment ago does now.
So sweep. Apply the rules wherever they fit, cross out what you find, and go around again until nothing changes. Loaded joints and supports stop the sweep, and often whole idle regions go quiet. On a member-force question, sweep first: every zero you cross out is a joint equation you don't have to write.
Move the load, and the zeros move
"Zero-force" describes a member under one loading, and not the member itself. Take the overhanging truss and slide its load around.
With at E, seven members are idle. Move it down to B, and the rule at B switches off, so BE now carries all of in tension, and the other six stay idle. Move it out to D, and BE drops back to zero. The two members at G are still idle, but CD, DF, CF, and EF wake up: three idle.
When a load moves, don't try to update your old list. Run the sweep again from scratch, because the joint the load left can now start a chain and the joint it reached can break one.
That's also why you can't just take these members out. BE looks useless with the load at E, and it's essential with the load at B. Even a member that never carries a calculated force does a job, since it holds a long compression chord in line so it can't bow sideways. A zero-force member is idle under this load, not unneeded.