M39029 Contacts: Decoding the Slash Sheet, the Size Suffix and the Colour Code

Every part number on a circular connector buys a shell, an insert, a coupling ring and a keying arrangement. It does not buy the contacts. Those are separate parts, ordered under a separate specification, with their own part numbers — and they are the only component of the assembly that touches the conductor, carries the current and moves every time the connector is mated. The shell has no opinion about whether the joint inside it is any good.

M39029, now maintained as SAE AS39029, is that specification. It exists because a connector house could otherwise supply contacts to whatever internal standard it liked, and a harness shop in another country would have no way to know whether a replacement was the same part. The specification fixes dimensions, material class, finish, retention and mating life; the slash sheets under it divide those requirements by connector family. Our own catalogue carries 128 M39029 part numbers across twelve slash sheets, and three questions arrive about them repeatedly: which suffix is size 16, why two contacts with the same size take different wire, and whether the non-QPL equivalent can be used. All three are answered by the part number itself — provided you read all four of its fields instead of the two most people look at.

A mated pair carries four part numbers, not one

This is the root of most contact-ordering errors, so it is worth stating plainly. A plug and a receptacle of the same series form a single mated interface, and describing that interface completely takes four part numbers: two shells and two contact sets. The shell numbers and the contact numbers are revised on different schedules, by different documents, and are frequently purchased from different suppliers.

A drawing that calls out D38999/26WB35PN and stops there is not a complete call-out. It names a shell with 13 cavities and says nothing about what goes in them, which is the half of the assembly that the current actually crosses. When a program reports intermittent channels after two years in service, the shell is almost never the finding.

The slash sheet names the connector family, not the contact size

The number after the slash is the part of an M39029 number that people skip, because the three digits after the dash look like they carry all the information. They do not. The slash sheet tells you which connector standard the contact was qualified against, and that in turn fixes the retention geometry, the insert it will fit, and often the extraction tool. Two contacts of identical size and identical wire range, on different slash sheets, are not interchangeable.

Read the family column straight down and one fact stands out: the same contact number serves two different connector standards. The /4 and /5 contacts are the contact set for MIL-DTL-26482 Series II and, without change, for MIL-DTL-83723 Series III — the same thirty part numbers in both cases. A buyer migrating between those two families keeps the contact inventory and changes only the shell and the backshell. That is a real procurement advantage, and it is invisible if you search by connector name instead of by slash sheet.

The other thing the slash sheet encodes is release direction. Front-release contacts are inserted and extracted from the mating face and are retained by a clip inside the insert; rear-release contacts are inserted and extracted from the rear and are retained by a shoulder that a grommet and a tool engage. The retention feature is machined into the contact, so a front-release contact and a rear-release contact of the same size cannot be substituted for one another even though both are “size 16”. This is why MIL-DTL-5015 appears in the table twice: the MS3400 class takes the /44 and /45 front-release contacts, while the MS3450 class takes the rear-release /29 and /30 pair. Order by connector class, not by the size on the purchase requisition.

The same architecture covers the rest of the range, and the extremes are where the size ladder stops being a guide. The power sizes run up to size 0 in the M39029/44 and /45 sheets — a size 0 pin and its size 0 socket are the same part numbers that a MIL-DTL-5015 MS3400 class connector uses, ordered to the same specification as a 22D signal contact. The D-Sub sheets add a size 20 socket and pin for MIL-DTL-24308, and the shielded sheets put a size 12 coaxial pin in the same family table. A buyer who treats M39029 as “the small signal contacts” will end up sourcing the power contacts from a different, uncontrolled route.

Three contacts, all called “size 22”

Here is the second question, and the one that causes the most re-terminations. Within the 38999 slash sheets there are three distinct contacts whose size is written “22”. They are not interchangeable, because they do not accept the same wire.

The overlap is what matters. A 22D takes 22 to 28 AWG, a 22M takes 24 to 28 AWG and a plain 22 takes 22 to 26 AWG. If your wire is 24 AWG, all three will accept it. If your wire is 22 AWG, the 22M will not, even though the cavity is identical and the box says size 22. The letter is a barrel variant, and the published wire range is the only reliable way to tell the variants apart — do not attempt to expand the letters from memory, because the same letters mean different things on different manufacturers’ catalogues.

The same trap exists at size 16, where it is easier to miss because the part numbers look like a sequence. M39029/4-111 and M39029/4-112 are both size 16 pins, but the first accepts 16–20 AWG and the second accepts 20–24 AWG. One is the standard barrel and one is the reduced barrel intended for a thinner wire in a size-16 cavity — a combination that occurs constantly when a bundle is derated for altitude or for a high-temperature zone and the conductor is downsized while the cavity stays. Ordering “size 16 pin” and accepting whatever the distributor ships will work about half the time, and the failure is a crimp that passes pull test on a single strand folded back on itself.

The colour code is the part number, written in bands

M39029 contacts carry a Basic Identification Number, and on small contacts the BIN is applied as three colour bands read from the wire-barrel end. The digit-to-colour mapping is fixed:

Two consequences follow that are worth more than the table itself. First, within the 38999 ladder above the second band separates the genders: sockets read Orange then Green, pins Orange then Blue, and the third band carries the size. On a bench with a loose contact in your hand and no bag, that is enough to tell a pin from a socket before you reach for the part number. Second, the first band identifies the family the contact was binned for, which is why a distributor cross-reference will print two different colour codes against the same contact number. Milnec, for example, lists M39029/4-110 as a single black band for MIL-DTL-26482 Series II and as brown–brown–black for MIL-DTL-83723 Series III. The contact did not change. The binning did, because the colour code exists to let a person confirm a part number in a bin, and the same number appears in two families’ bins.

This is also a procurement fact rather than a curiosity. BIN marking is required on certified M39029 contacts and is applied by stamping, rolling or colour banding depending on contact size — it is one of the process steps the certification pays for, and it is one of the steps a commercial equivalent is allowed to skip. When a lot arrives with no bands and no stamp, that is not automatically a counterfeit; it may be exactly what was ordered. It does mean the part in your hand cannot be identified from the part in your hand, and it must stay in its labelled packaging until it is crimped.

Plating, type and class: the three fields that get dropped

A complete M39029 call-out carries three qualification fields beyond the part number, and they are the ones most often left to the supplier’s discretion because they are not visible on a drawing.

The specification requires gold, and the reason is the one that matters for a connector rather than a wire terminal: two gold surfaces do not form an oxide film, so the contact resistance after 500 cycles resembles the contact resistance at cycle one. The historical requirement was a minimum of 50 µin over the whole outer surface; the specification now permits selective plating, which puts the thickness where the contact actually touches and leaves the barrel lighter. That is a genuine cost saving and a genuine change in what you are buying — two parts both described as “gold plated” can differ by more than a factor of two in the thickness on the mating surface, which is the only place the number does any work.

Note also that the contact class is not the shell class. A 200 °C contact in a 175 °C shell is a 175 °C connector. The assembly rating is the lowest of the insert, the seal, the shell finish and the contact, and the contact is the element most often left unspecified because it is bought separately. See the plating guide for how the finish interacts with the shell, and the contact size guide for the current side of the size decision — which is a different question from the one this article answers.

Thermocouple contacts are an alloy, not a size

The thermocouple slash sheets look like an ordinary size ladder and are not. Each suffix is a specific thermocouple alloy, and the alloy is electrically part of the measurement circuit.

A thermocouple works by generating a voltage across a junction of two dissimilar metals, and it measures temperature only because the voltage is a known function of the temperature difference. Insert a copper-alloy contact into the middle of one leg and you have created two extra junctions of a third metal, each at a different temperature, in series with the pair you meant to measure. The reading shifts by an amount that depends on the temperature gradient along the harness — which is why a mixed-alloy contact set produces an error that looks like drift, calibrates out on the bench and returns in service.

There is a second trap specific to this family: thermocouple contacts carry no colour code at all, because a colour code would be a visual claim about alloy and the marking would have to survive the very temperatures the alloy is chosen for. So the one family where substituting by size is most damaging is also the family where the bands cannot warn you. Both legs must be ordered as the alloy the thermocouple type requires, and the size-16 variants — the /9-514 to –518 and /10-519 to –523 range — follow the same alloy order for 16–20 AWG wire.

QPL, or a commercial equivalent

The specification is deliberately loose in one place and strict in another. It requires the contact be made of a copper alloy — and there are several hundred copper alloys, with different tensile strength, spring properties and crimping behaviour. Bronze, brass, leaded nickel copper and beryllium copper all satisfy the sentence and do not behave alike. The harder alloys need zone annealing before they can be crimped without cracking, which is a real process step with a real cost, and the softer crimpable brasses avoid it. Nothing in the part number tells you which alloy you received.

What the specification does fix is the mating-life floor: contacts must support a minimum of 500 mating cycles, and the number is a statement about the socket spring rather than the pin. Larger contacts have thicker walls and retain their spring more easily, which is why a size-0 power socket can be made from a cheaper alloy than a size-22D signal socket with the same claimed cycle count.

Certified contacts appear on the Qualified Products List, and getting there is a long and expensive government test programme. That is the whole content of the “commercial equivalent” argument: a non-QPL part may be the same design on the same machines with the same plating, or it may not, and the price difference of roughly a quarter to a half is the price of the evidence rather than the price of the metal. Three positions are defensible:

• QPL mandatory — any call-out that cites the military specification by number, any defence or civil-certification programme, and any spares pipeline that will be audited against the drawing.
• Commercial acceptable with evidence — industrial and harsh-environment applications where the supplier can state the alloy, the plating thickness on the mating surface and the mating-life test result. Ask for those three in writing; a datasheet that says “gold plated” without a thickness has not answered.
• Commercial acceptable without evidence — prototype, bench, and non-safety low-cycle internal equipment. Do not let a prototype part quietly become the production part because it passed the prototype.

The one substitution that is not defensible is a mixed set: QPL pins in one half of a pair and commercial sockets in the other. The plating thickness and the alloy now differ across the mating interface, and the wear couple is asymmetric — one side gives up its gold first, and the resulting failure appears on the side that was bought correctly.

What belongs on the RFQ

Ten lines close every ambiguity discussed above. They also make a quote comparable, which is the real reason to send them.

• The connector part number, and the slash sheet of the contact if it is known.
• Contact part number with the full three-digit suffix, not just the size.
• Quantity per cavity, plus the number of spare cavities in the harness — spares are usually the difference between a repair and a re-termination.
• Wire gauge and insulation type actually being crimped, in AWG and in millimetres.
• Plating thickness required on the mating surface, in microinches, and whether selective plating is acceptable.
• Temperature class required, and the class of the shell it is going into.
• QPL required, or commercial equivalent with alloy and plating stated in writing.
• Whether the contacts will be supplied loose or on a reel, since that decides the crimp tool and the inspection method.
• Whether thermocouple contacts are involved, and if so for which thermocouple type.
• Any BIN or lot traceability requirement, and whether the marking must survive a specific cleaning process.

Where the contact meets the tooling

A contact is only as good as the crimp that terminates it, and the tool frame is matched to the contact range rather than to the connector. In the current Daniels/AS22520 family the upper-range frame is the AF8 at M22520/1-01, covering roughly 12 to 26 AWG with an eight-impression crimp and a large library of turret heads and positioners; the miniature frame is the AFM8 at M22520/2-01 for the 20 to 32 AWG range that the 22D contacts live in; and the MH860 at M22520/7-01 covers the 16 to 28 AWG middle ground. Each frame is gaged periodically with an M22520/3-1 GO/NO-GO gage, and the selector position used for gaging is part of the tool’s instruction sheet, not a matter of preference.

The consequence for ordering is that an uncommon contact size can commit you to an uncommon positioner, and a positioner is a longer lead time than the contact. If a design introduces one cavity at a size the shop has never crimped, the tooling question belongs in the same RFQ as the contact. The mechanics of inserting and extracting the finished contact, and of proving the crimp once it is made, are covered in the insertion and extraction tooling guide and the crimp cross-section guide.

When the contact is the problem

Frequently Asked Questions

Can I use a M39029/4 contact in a MIL-DTL-83723 Series III connector?

Yes — M39029/4 and /5 are the contacts for both MIL-DTL-26482 Series II and MIL-DTL-83723 Series III, and the same part numbers serve both families. What changes is the binning and therefore sometimes the colour code printed by a distributor, so the part number is the thing to match, not the band pattern on the contact in your hand.

Is a size 22 contact the same as a 22D contact?

No. Size 22, 22D and 22M share a cavity diameter and differ in wire barrel and published wire range — 22 takes 22–26 AWG, 22D takes 22–28 AWG and 22M takes 24–28 AWG in the 38999 ladder. If your wire is 22 AWG, the 22M is the one that will not work, even though every document in the chain says “size 22”.

Does a thicker gold layer really matter in a connector?

It matters more than in a wire terminal, because the contact is designed to be mated repeatedly. Gold does not form an oxide, so the resistance stays low until the layer is worn through; the thickness on the mating surface is therefore a direct statement of how many cycles remain before the nickel underneath becomes the contact surface. Two parts both described as gold plated can differ by a factor of two or more in that number.

Why is my contact marked with colour bands but the datasheet lists no colour?

Because the colour bands are the BIN — a lot and part identification applied to the contact — and the BIN is mandatory on certified contacts. Datasheets often quote only the part number and omit the binning. If the bands are absent, confirm with the supplier whether the part was supplied as a commercial equivalent, which is not required to carry BIN marking.

Can a commercial equivalent replace a QPL contact?

In non-military harsh-environment applications, yes, and the saving is typically a quarter to a half. In any application whose drawing cites the military specification, no. The defensible middle position is to require the alloy, the plating thickness on the mating surface and the mating-life result in writing — a commercial part with those three stated is a better purchase than a QPL part with none of them documented.

Do I need different contacts for the plug and the receptacle?

You need pins in one half and sockets in the other, and the pin and socket slash sheets are separate numbers, not variants of one. The half that carries pins is a design decision, normally pins in the receptacle for a live interface and pins in the plug for a de-energised one, and it changes both contact part numbers in the assembly.

How many spare contacts should I order?

Enough to cover a re-termination at every cavity you cannot reach easily, plus one per size for the tool setup. A crimped contact cannot be re-crimped, so every repair consumes a contact and a short length of wire; two or three repairs at the same cavity turns a contact change into a splice, which is a different and much more expensive job.