Solder Cup Termination: Building and Inspecting the Joint

A solder cup looks like the simplest termination in the catalogue. There is a hole, there is a wire, there is solder. Nothing is calibrated, nothing is crimped, nothing needs a positioner or a pull-test fixture. That simplicity is real, and it is also why the joint fails: because a cup asks the operator to solve a heat-transfer problem in four seconds, with a hand tool, inside a connector that is full of elastomer.

The cup is not the problem. It is a small blind hole drilled or formed into the back of a contact, sitting inside an insert, inside a metal shell. Put an iron on it and you are not heating a cup — you are heating a contact, an insert, a shell and whatever fixture holds the connector, and only the first should get hot.

Everything below follows from that: the joint is decided by how much heat arrives, how long it stays, what it does to the materials around it, and whether the metal inside the cup was ever in the right place. It is not decided by how the joint looks — the two commonest defects, solder wicked up under the insulation and a cup half full at the bottom, both photograph well.

Why the Cup Survives — and What It Costs at the Joint

The choice between crimp, solder and PC tail is a selection problem with its own article; this one assumes the cup is already specified and asks what to do about it. It is still worth stating why the answer is often "cup", because a process engineer who treats the cup as a legacy mistake will not build a good one.

Cup termination survives because it needs no tooling per contact size, because it tolerates a wire gauge the cup was not designed around better than a crimp barrel does, and because it can be re-terminated in the field with an iron and a pair of side cutters. In the power inserts of the MIL-DTL-5015 family, in legacy platforms whose drawings were frozen forty years ago, in low-volume builds where a crimp tool and positioner per contact size cannot be justified, and in repairs whose alternative is cutting the harness out, the cup is the only economic answer.

What it costs is control. A crimp is closed by a die, so the finished geometry is set by a tool that can be measured and gated. A cup is closed by a person, so the geometry is set by a dwell time and a feed rate that exist only in that person's hands unless someone writes them down and inspects against them.

Four Interfaces in One Joint

A crimp has one interface that matters: wire strands compressed against barrel metal. A solder joint has four, and they fail for different reasons, which is why a single "does it look soldered" check misses most of them.

Strand to strand

Inside the cup, the individual strands must be bridged by solder. If they are not — because the bundle was twisted tight, because flux never reached the centre, because the wire was inserted dry into a cup that had already solidified — the current path is a few outer strands and the rest of the conductor is mechanical ballast. The joint measures fine at build and runs hot in service.

Strand to barrel wall

This is the interface everyone pictures: solder wetting both wire and cup wall, forming a fillet at the rim. Wetting is chemical, not thermal — it happens only where the surfaces were clean and fluxed at the moment they were hot, which is why a joint made with an exhausted flux core wets on one side and balls up on the other.

Cup bottom to wire end

The wire end must reach the bottom of the cup. A wire that stops halfway leaves a column of solder with no conductor in its lower half: the electrical path is intact, the mechanical grip is halved, and the void at the bottom is invisible from above. This is the defect that a "solder visible in the inspection hole" criterion is actually testing for, and it is also the one that a pre-tinned cup makes impossible to see.

Insulation to cup rim

The fourth is the one nobody lists. Strip far enough that no insulation enters the cup, close enough that no bare conductor is left between cup and jacket. Both errors are thermal: strip too short and the iron melts insulation into the joint; strip too long and unsupported strands fatigue at the rim.

Cup Geometry Decides the Joint Before You Pick Up the Iron

A cup has three dimensions that matter — inside diameter, depth, and the wall thickness that carries heat into it — and one relationship that matters more than any of them: the cup must be sized to the wire. A small wire in a large cup cannot be soldered well by any operator, because the gap it leaves is a reservoir that fills with solder instead of closing around strands. The working rule most trainers apply is that the wire should be inserted to the full depth of the cup and the solder should fill roughly three-quarters of it — treat that as orientation for judging a joint, not as a clause you can quote at an inspector, because the governing document is the one your customer's drawing invokes.

Strip length is derived, not chosen. It is the cup depth plus a small clearance so the strands reach bottom without the insulation being pushed into the cup mouth. If the strip length in the work instruction is "approximately 6 mm", it is not a specification; it is a guess that will be executed differently by each operator and differently again on a Monday.

Cup construction matters too. A drilled cup has a machined wall and a flat or conical bottom; a formed cup has a thinner wall and a radius at the bottom, so it heats faster, wets more easily and distorts more readily. Neither is a defect — but an instruction written for one and applied to the other produces different dwell times for the same wire, which is exactly the drift that surfaces as "some joints look better than others".

Plating is the last variable: silver wets almost instantly and tarnishes in storage, gold wets well and is thin enough that a heavy hand erodes it, tin is the easiest to solder and the least durable under vibration. If the contacts sat in a humid store, the plating is the variable, not the operator.

The Heat Budget: Where the Iron's Energy Actually Goes

Every cup joint is a race between the iron delivering enough energy to melt solder and the rest of the connector conducting it away. Energy leaves through four paths at once, and only the first is useful.

It goes into the wire, a copper rod of large cross-section, so heat runs down the conductor toward the insulation and away from the joint. It goes into the contact body and then the insert, a polymer with a far lower melting point than the solder and no ability to recover. It goes into the shell, the largest thermal mass in the assembly, usually clamped in a fixture that is also a heat sink. And it goes into the air, which is why a cup on a bench solders differently from the same cup inside a shell with eleven neighbours.

The consequence is a rule that sounds like an opinion until you watch it happen: a cup that will not wet within a couple of seconds is telling you the iron, the tip or the heat path is wrong — not that you should wait longer. Waiting longer puts the same energy into the insert, and the insert is the part of the connector that does not come back. A softened or cracked insert loses contact retention, loses the interfacial seal, and produces a connector that passes a continuity check and fails a dielectric or a leak test.

A second-order effect worth naming: the connector gets hotter as the harness progresses. The first cup sees a cold shell; the twentieth sees one that has absorbed everything nineteen rejected joints gave it. Joints at the end of a connector are therefore made in a different thermal environment from those at the start — which is why "the last few positions always look worse" is a process observation, not a coincidence.

Alloy and Flux: Two Materials That Set the Process Window

The alloy decides the temperature you have to reach and what the joint does afterwards. Sn63Pb37 is eutectic: it melts at 183 °C, it passes from solid to liquid at a single temperature with no pasty range, and it solidifies quickly and predictably. That single melting point is the reason it remains the default in high-reliability and defense work even where it is otherwise restricted, and it is the reason a joint made with it is easier to make well than the same joint made lead-free.

Lead-free alloys — SAC305 is the common one — melt roughly 30 to 35 degrees higher, in a pasty range rather than at a point. Higher temperature means more heat into the insert for the same joint. A pasty range means the joint is mushy while it cools, so any movement in that window produces a disturbed, grainy joint that looks cold and behaves worse. Neither alloy is "better"; they demand different dwell times and different handling, and a work instruction that does not say which one is in use cannot be executed consistently.

Flux is the other half. Rosin and rosin-based fluxes are the traditional choice in connector work because their residues are non-corrosive and, in the mildly activated grades, tolerable on a joint that will never be cleaned. No-clean fluxes are formulated so that residue can stay, which is true on a board and less true inside a connector that will be potted, sleeved or sealed, where the residue has nowhere to go. Water-soluble and other highly active fluxes give the best wetting and leave ionic residue that must be removed completely — and "completely" inside a cup, under a sleeve, in a shell, is a much harder promise than it is on a flat board.

The Sequence: Ten Steps and the One That Gets Skipped

Soldering a cup is not difficult. Doing it the same way four hundred times is. The sequence below is deliberately written as a sequence because the failures cluster at the transitions between steps, not inside them.

1. Verify the strip before the joint. Cup depth plus clearance, no nicked or unevenly cut strands. A conductor that lost a third of its strands to the stripper cannot be recovered by better soldering.
2. Check the cup is empty and clean. A cup with residual solder from a previous contact, or with a film of handling oil, will not wet evenly. If it needs cleaning, clean it and let it dry before fluxing.
3. Decide the tinning policy and follow it for the batch. Pre-tinning the wire is standard. Pre-tinning the cup speeds the joint and destroys your ability to see whether the wire reached the bottom.
4. Tin the conductor, not the insulation. Solder should stop at the strip line. If it runs under the jacket, the strands there are solid instead of flexible, and that is where the wire will eventually break.
5. Insert to the bottom. Bottomed out, not "in far enough". This is the step that gets skipped, because it is the one with no visual feedback unless the cup has an inspection hole.
6. Apply heat to the cup wall, feed solder on the far side. The iron heats the cup; the solder melts against hot metal, not against the tip. Feeding onto the tip coats the tip and starves the joint.
7. Remove the solder, then the iron. In that order. The reverse order leaves a spike and a joint that was never fully fed.
8. Hold still until solid. Movement during solidification is the single most common cause of a joint that looks cold without ever having been cold — and with a lead-free alloy the window is wide enough to walk into.
9. Inspect before the next one. Not at the end of the connector — by then the evidence of a mis-set iron is twenty joints back.
10. Clean, then protect. Clean before sleeving, potting or backshell assembly, because after that point the residue is inaccessible by construction.

The Three Things You Cannot See

Visual inspection of a solder cup catches gross defects and misses the three that actually kill connectors in the field. All three have a signature, but the signature is not at the joint.

Wicking under the insulation

Solder follows heat, and heat follows copper. Heat the conductor long enough and solder runs up the strands under the jacket, where it converts a flexible conductor into a rigid one. The failure is not at the cup: it is a fatigue fracture at the boundary between wetted and unwetted strands, a centimetre or two up the wire, after vibration or thermal cycling. The only reliable check is tactile — flex the wire near the termination. A conductor rigid well above the joint is a reject, however good the fillet looks.

Heat damage to the insert and the plating

An insert that has been overheated does not usually announce itself. It discolours slightly, or it does not discolour at all and simply loses retention force, because the polymer relaxed and the contact cavity is now a few hundredths of a millimetre larger than it was. The joint at that position is perfect. The connector fails a contact-retention test, or loses its interfacial seal when mated, or develops an intermittent under vibration that no amount of re-seating fixes. Because the evidence is elsewhere, this defect is attributed to the contact or the mating half for months.

Residue under the protection

Residue on an open joint is a slow cosmetic problem. Residue sealed under a sleeve, a boot or potting is a corrosion cell with a fixed moisture supply and no way to dry out. The symptom is falling insulation resistance or rising leakage current weeks after build, on the positions that were hardest to clean. If the connector will be sealed, cleaning is not optional and its verification belongs in the process, not in the operator's judgement.

Inspection: Reading the Fillet, the Cup and the Wire

A good joint has a concave fillet wetting continuously around the cup rim, a wire visibly entering the solder mass, no exposed base metal inside the cup, no icicles or spikes, no charred flux, and undisturbed insulation at the correct clearance. The fillet looks slightly drawn-in, not domed; a domed joint was usually fed from the top and never wetted the wall.

These criteria are not aesthetic. They map to the four interfaces above, and the disposition differs by interface — which is why a defect table that says what an appearance actually means beats a flat list of what to look for.

Note what the table does not contain: a number for how full the cup should be. A fill percentage is meaningful only against a stated cup depth and a stated wire gauge, and it is not measurable on an assembled connector without a section. Where a customer's drawing or a workmanship standard states a fill criterion, that criterion governs; where nothing states one, the fill rule you apply is an internal decision and should be written down as one.

Rework: The Second Joint Is the Riskier One

Every rework puts a second heat cycle into the same insert, a second flux dose into the same cavity, and a shorter conductor into the same cup. The insert has already been softened once and the second cycle takes it further past its glass transition in the very cavity that has to hold the contact. Residue from the first attempt is still there unless it was cleaned, and reheating drives it deeper. And the wire is shorter, so either the strip is marginal or the service loop the design assumed is gone.

That is why good work instructions cap rework rather than encouraging it: one re-termination per conductor is a common internal limit, because the third attempt is not a repair, it is a new joint made from degraded materials with a shortened wire. It is also why the correct response to a bad joint is a fresh strip and a fresh contact where the contact is replaceable, not a longer dwell and more solder on the joint that already failed.

Rework has to be recorded: a connector with three re-terminated positions is a different article from the one on the drawing, and that matters to whoever investigates a failure two years later. For what is repairable in the field at all, see field repair and re-termination.

Making the Joint Stop Carrying the Load

A solder joint is an electrical connection. It is not a fastener and not a strain relief. The loads in a connector are real — cable weight, vibration, mating force, the thermal expansion of the harness — and if any of it arrives at the cup, the joint eventually cracks, not because it was badly soldered but because the design asked solder to do a job it cannot do.

The load path is a chain and every element in it has to be continuous: wire supported by jacket, jacket by a boot or backshell clamp, clamp by backshell, backshell by the shell coupling. A service loop between clamp and contacts absorbs movement without putting it into the joint. Bend radius at the exit follows the cable's own minimum, not a rule of thumb, and the clamp grips the jacket, never the conductors.

This is where M85049 backshells and accessories and the various strain-relief adapters earn their place: not as cosmetic finishing, but as the mechanical element that converts a cable load into a shell load so the solder joint only ever sees current. A connector built with the cups correctly soldered and no strain relief will fail at the cups, and the failure will be blamed on the solder.

Which Workmanship Document Governs Your Cup Joint

Connector soldering sits between the electronics workmanship standards and the cable/harness standards, and which applies depends on where the joint lives rather than what it is. That matters when an inspector arrives with one document and your process was written against another.

The three IPC documents routinely disagree on how much solder is enough at a cup, and the resolution is never "whichever is convenient" — it is whatever the purchase order invokes. A supplier that cannot state which document its work instruction was written against cannot defend a joint that fails an incoming inspection it did not know was coming.

What to Put in the Work Instruction

Because the cup gives the operator so much discretion, the work instruction is the process control. If it is thin, the control does not exist. Ten lines that make it a real specification:

• The alloy and its form (wire diameter, flux core or not), named explicitly — not "solder".
• The flux class, and whether cleaning is required, permitted, or forbidden.
• Strip length with a tolerance, derived from cup depth, plus a no-nicked-strand criterion.
• Whether the cup is pre-tinned, and if so how the wire-to-bottom condition is verified.
• Iron tip geometry and set temperature, and the maximum dwell per joint.
• The acceptance criteria for the fillet, stated as a shape rather than a percentage.
• The inspection class and the inspection point (during, not after, the connector).
• The rework limit per conductor, and what happens to the contact at the limit.
• The cleaning method and how its effectiveness is verified.
• The electrical test after assembly — continuity alone is not enough; low-level contact resistance and a dielectric withstand test are what find the joints that conduct and should not.

Solder-Cup Lines We Build

Airoadcon supplies solder-cup variants across the MIL-DTL-5015 power family — Series II and Series III, plus the older MS5015 insert arrangements — and across the signal families where cup termination is still specified: J599 (MIL-DTL-38999 Series III), ZH83723 (MIL-DTL-83723) and the legacy MIL-DTL-26500 bayonet range. Contacts are stocked as M39029-type solder cups in the common contact sizes, and the mechanical half of the joint — backshells, clamps and strain relief — is in M85049 accessories. Where a program needs the harness built rather than the connector supplied, that sits under custom cable assemblies and wire harnesses.

If the joint is being specified rather than bought, send the wire gauge, insert arrangement and workmanship standard; contact selection follows from those three, not from the shell.

Frequently Asked Questions

Is a solder cup joint weaker than a crimp?

In a room-temperature pull test, often yes: a correctly set crimp is repeatable and a hand-soldered cup varies with the operator. In fatigue the comparison depends on strain relief, where a cup with a proper service loop and a clamped jacket beats a crimp with none. A crimp is more predictable; a cup depends more on the design around it.

Should I pre-tin the cup before inserting the wire?

It shortens the dwell and eases wetting, and it removes your ability to confirm the wire reached the bottom. If the plating has tarnished in storage, pre-tinning is the pragmatic choice — but then insert-to-bottom has to be controlled by the work instruction and verified on the first article by section, not by eye.

How much solder should be in the cup?

Enough that the cup is substantially filled around a wire that reaches its bottom, without a dome standing above the rim. Where a drawing or a workmanship standard states a fill criterion, that governs. Where none does, the common working interpretation is about three-quarters of the cup with the wire at full depth — treat it as an orientation figure, write whatever you adopt into the work instruction, and verify it destructively on the first article rather than arguing about it at inspection.

Why do the last contacts in a connector always look worse?

Because the shell and insert have absorbed the heat earlier joints rejected, so the last positions are soldered in a hotter assembly with a larger thermal load behind them. Spread the sequence, allow cool-down between adjacent positions, and treat a hot shell as a signal to stop rather than to work faster.

Can I solder a cup without removing the contact from the connector?

Usually yes, and that is the normal practice for a repair. It also means the insert is in the heat path for every joint. If the contact can be removed with the correct extraction tool, doing so moves the heat problem to the bench, where it is easier to control — see contact insertion and extraction tooling for the tooling side of that decision.

How do I know the joint will survive vibration?

Not by looking at it. Solder joints fail in vibration at the wetted-to-unwetted strand boundary and at the cup rim, both behind the visible surface. The evidence comes from the design — service loop, clamped jacket, respected bend radius — plus a sample-level vibration test. A connector that passes a pull test and has no strain relief is qualified for pulling, not for vibration.