Quick answer: The JST SDN connector is a 3.96mm pitch board-in crimp connector — the housing itself is soldered into the PCB and the crimped contact is pressed into it from above, so there is no header and no wafer. JST rates it at 7A AC/DC at AWG #18, 250V AC/DC, −25 to +85°C, over 2 to 12 circuits, for wire AWG #22 to #18 with insulation φ1.6 to 3.0mm, on a 1.6mm board. The KONNRA KR3963 is the cross-reference equivalent, and this is the most exact match in this comparison series so far: seven published rows line up, including the two that are hardest to line up — the current rating’s gauge basis, and the housing’s flame-retardant grade and colour. Our 7A is at AWG #18 exactly as the original’s is; our housing is PA66 UL94 V-0 in white exactly as the original’s is; the voltage, the withstanding voltage, the insulation resistance and the wire range all agree. The one line that does not match is the metal: JST specifies the contact as phosphor bronze with a reflowed tin finish, and our specification says brass tin plated over nickel — while our own product page claims phosphor bronze for the same part number. That contradiction is the first thing to settle, and it is the subject of the fourth section below.
I work on connector and harness programmes at KONNRA, so treat the disclosure as read. Everything below comes from manufacturer documents on both sides, and where they disagree I have said so rather than averaged. This round the disagreements are almost all internal rather than between the two companies — our own specification and our own product page name two different base metals for one part number — so the defects section is longer than usual and the comparison section is shorter.
KONNRA KR3963 series SDN 3.96 board-in crimp connector
👉 KR3963 product page · KR3963 housing · KR3963 terminal · Board-in connector range
What the JST SDN Actually Is
JST’s own description is unusually specific about the architecture, and it is the sentence that decides whether this family fits your board at all: “The SDN connector is a 3.96mm pitch board-in crimp style connector capable of connecting a wide variety of circuits including signal and power supply circuits. The connector features a slim, low-profile design. In addition, it has an exclusive lance construction to reduce insertion force, thus substantially improving the working efficiency in mounting the contact.”
| Attribute | JST published value |
|---|---|
| Series | SDN connector |
| Architecture | Board-in crimp style — no header, no wafer |
| Pitch | 3.96mm |
| Current rating | 7A AC/DC (AWG #18) |
| Voltage rating | 250V AC/DC |
| Temperature range | −25°C to +85°C, including temperature rise in applying electrical current |
| Withstanding voltage | 1,500V AC/minute |
| Insulation resistance | 1,000MΩ min. |
| Applicable wire | AWG #22 to #18 |
| Applicable PC board thickness | 1.6mm |
| Circuits | 2 to 12 |
| Housing material | PA 66, UL94V-0, natural (white) |
| Contact material and finish | Phosphor bronze, tin-plated (reflow treatment) |
| Standards | Recognized E60389 · Certified LR20812 · R50100763 |
| Compliance | RoHS |
Four features are named on the original’s side, and three of them are about assembly rather than electricity: a slim, low-profile design; easy and secure contact mounting; improved solderability; and a polarizing boss. The one that carries engineering weight is the exclusive lance construction to reduce insertion force — a lance geometry designed to make pushing the crimped contact into the housing easier, which is the operation that a board-in connector performs once per contact on the line. And the polarizing boss is a documented ordering option rather than a fixed feature. JST ships each circuit count twice: 2P-SDN through 12P-SDN with the polarizing boss, and 2P-SDN-A through 12P-SDN-A without it. A customer who has designed around one of the two cannot substitute the other, and the difference is a single suffix letter.
Board-in: what it changes about the bill of materials
The architecture is the first thing to match, and here it matches exactly.
| Wire-to-board with a header (for example JST VH) | Board-in (JST SDN, KONNRA KR3963) | |
|---|---|---|
| Board-side part | A separate header (wafer) soldered to the PCB | The housing itself is soldered to the PCB |
| Wire-side parts | Housing plus crimp contact | Crimp contact only |
| Parts in the mated system | Three: wafer, housing, contact | Two: board-mounted housing, contact |
| Mating operations over the product’s life | Repeated plug and unplug of the wire half | The contact is pressed in once, at assembly |
| Overall height above the board | Header plus housing stack | One housing height |
| Field rework | Unplug the wire half | Release and re-press the individual contact |
Two consequences follow, and they are the reason this family exists. The first is a lower profile and a shorter board footprint, because the housing sits directly on the board instead of on top of a header. JST names the low-profile design as the family’s headline feature and gives a slim, low-profile claim rather than a mounting height — which is itself worth noting, since the VH family publishes a mounting height of 16.5mm and the high-box VH publishes 21.3mm. For SDN, JST publishes no overall height at all. The second is that the standard insertion/withdrawal force ladder that a wire-to-board family would carry is architecturally absent, and its absence is correct rather than a gap. A board-in contact is pushed into its housing once during assembly and is not repeatedly mated afterwards, so a table of mating force against circuit count has nothing to describe. Our own specification reflects that correctly: it publishes terminal insertion force and terminal-to-housing retention force at the contact level, and it publishes no per-circuit mating-force table at all — and it also marks the entire wafer block as N/A, in both the straight and the right-angle variants, for base material, contact and solder tab. That is a form answering a question that does not apply, and it is the right way to fill it in.
The part-number map on the original’s side
| Function | JST part numbers | Published detail |
|---|---|---|
| Crimp contact | SDN-21T-P1.5 |
0.33 to 0.83mm²; AWG #22 to #18; insulation O.D. 1.6 to 3.0mm; 5,000 per reel; phosphor bronze, tin-plated (reflow treatment) |
| Housing, with polarizing boss | 2P-SDN … 12P-SDN |
2 to 12 circuits; PA 66 UL94V-0 natural (white); 1,000 per bag for 2 to 5 circuits, 500 per bag for 6 to 12 |
| Housing, without polarizing boss | 2P-SDN-A … 12P-SDN-A |
Same materials and quantities; the -A suffix removes the boss |
| Crimp tooling | AP-K2N machine, MKS-L applicator, MK/SDN-21-15 die, APLMK SDN21-15 applicator with dies |
“Contact JST for fully automatic crimping applicator” |
Two things in that table are unusual for a JST family, and both matter commercially. Every circuit count has a published dimension. The SDN catalogue gives an A and a B dimension for each of the eleven housing variants — 3.96mm and 8.96mm at 2 circuits, rising to 43.56mm and 48.56mm at 12 — so a footprint can be taken directly from the original’s document without a drawing request. We publish no housing dimensions and the KR3963 housing drawing is a vector PDF with no text layer, so a footprint has to be requested from us. And the bag quantity falls at 6 circuits. JST ships 1,000 per bag from 2 to 5 circuits and 500 per bag from 6 to 12 — a packaging break that follows the part’s size and is worth knowing before planning a line that consumes whole bags. The board layout carries two conditions that are easy to miss. JST publishes the layout “viewed from soldering side” with the note that “tolerances are non-cumulative: ±0.05mm for all centers. For connectors with 5 or more circuits the hole dimensions will differ from the above values. Contact JST for details.” So the published hole dimensions are not valid for 5 circuits and above without asking — a rare case where the original’s own document points the reader away from its own table.
Two KONNRA 3.96mm series, and why they are not interchangeable
This is the section I would read first if you arrived here from a quotation rather than from the original’s datasheet, because our catalogue contains two 3.96mm families whose published ratings are almost identical and whose products are not substitutes for each other.
| KR3961 | KR3963 | |
|---|---|---|
| Original family | JST VH | JST SDN |
| Architecture | Wire-to-board with header: straight and right-angle wafers plus housing | Board-in: housing soldered to the board, no wafer |
| Parts required | Wafer + housing + contact | Housing + contact |
| Published voltage | 250V AC/DC | 250V AC/DC |
| Published current | 7A at AWG #18 | 7A at AWG #18 |
| Published temperature | −40 to +105°C | −40 to +105°C |
| Contact resistance | 20mΩ max | 20mΩ max |
| Insulation resistance | 1,000MΩ min at 500VDC | 1,000MΩ min at 500VDC |
| Insulation O.D. | 2.8mm max | 2.50mm max |
| Circuit range published on the website | 2 to 11 (housing) / 2 to 10 (wafer) | 2 to 15 |
| Terminal part numbers | T39610PT0101A, T39610BT0101A |
T39630BT0101A |
Six of the nine rows are identical, and that is exactly the problem. Voltage, current, current basis, temperature, contact resistance and insulation resistance all read the same on both series. What differs is the architecture, the circuit count, the wire window and the terminal — and the first of those is the one that decides whether the part can be used at all. A designer comparing our two product pages sees the same numbers and would reasonably conclude either part will do; a designer comparing the two original datasheets sees a wire-to-board system with a header against a board-in housing with no header, and no ambiguity. So the practical instruction is: pick the series by the original part number, never by the rating. If the part being replaced starts with B and ends in -VH or -VH-FB-B, it is a header and the answer is KR3961. If it is a nP-SDN housing, it is board-in and the answer is KR3963.
The Ratings Side by Side
| Figure | JST SDN catalogue | KONNRA spec §4.0 | KONNRA product page | KONNRA component pages |
|---|---|---|---|---|
| Architectures | Board-in crimp style | Board-in, wafer marked None | Board In | Category: Board In |
| Pitch | 3.96mm | 3.96mm | 3.96mm | 3.96mm |
| Voltage rating | 250V AC/DC | 250V AC/DC | 250V | 250V |
| Current rating | 7A AC/DC (AWG #18) | 7A (18AWG) | 7A | 7A |
| Applicable wire | AWG #22 to #18 | AWG 18# to 22# | — | 18#–22# |
| Insulation O.D. | 1.6 to 3.0mm | 2.50mm Max (no minimum) | 2.8mm Max | 2.80mm(MAX) |
| Temperature range | −25°C to +85°C | −40 to +105°C | −40℃~+105℃ | −40℃~105℃ |
| Withstanding voltage | 1,500V AC/minute | 1,500V AC, 1 minute | 1500V AC/minute | — |
| Insulation resistance | 1,000MΩ min. | 1,000MΩ Min at 500VDC | 1000MΩ Min | 1000MΩ Min |
| Contact resistance | not published | 20mΩ Max (dry circuit, initial) | 20mΩ Max | 20mΩ Max |
| Contact resistance after test | not published | 40mΩ Max after each named test | — | — |
| Durability | not published | no clause published | — | — |
| Applicable PC board thickness | 1.6mm | not published | — | — |
| Circuit range | 2 to 12, with published A and B dimensions for every count | not published anywhere in the specification | 2–15pin | 2P~15P |
| Housing material | PA 66, UL94V-0, natural (white) | PA66 UL94 V-0 | PA66/UL94/Phosphor Bronze | PA66 UL94 V-0 |
| Contact material | Phosphor bronze, tin-plated (reflow treatment) | Brass, tin plated over nickel | (claimed as Phosphor Bronze) | not published |
| Agency standards | Recognized E60389 · Certified LR20812 · R50100763 | — | — | UL E482542 |
| Board layout | Published, viewed from the soldering side, with a ±0.05mm non-cumulative hole-pitch tolerance | not published | — | — |
Two columns in that table are worth reading against each other before anything else. Our product page publishes an insulation diameter of 2.8mm where our specification says 2.50mm, and our product page publishes “Phosphor Bronze” where our specification says brass. Both are covered in the defects section; both are internal to us rather than differences with the original.
Seven Rows That Match Exactly
A comparison guide that leads with the failures is no use to an engineer who needs to know whether the parts can be swapped, so the matches come first — and this family produces more of them than any other in this series.
| Row | JST SDN | KONNRA KR3963 | Verdict |
|---|---|---|---|
| 1. Architecture | Board-in crimp style: housing soldered to the PCB, crimped contact pressed in from above, no header and no wafer | Board-in, wafer column reads “None”, parts list is housing plus terminal only | Match — and it is the row that matters most |
| 2. Current rating | 7A AC/DC (AWG #18) | 7A (18AWG) | Match, including the gauge basis |
| 3. Voltage rating | 250V AC/DC | 250V AC/DC | Match |
| 4. Withstanding voltage | 1,500V AC/minute | 1,500V AC, 1 minute | Match |
| 5. Insulation resistance | 1,000MΩ min. | 1,000MΩ min at 500VDC, adjacent contacts, per EIA-364-21B | Match, including the test voltage |
| 6. Applicable wire | AWG #22 to #18 | AWG 18# to 22# | Match |
| 7. Housing material | PA 66, UL94V-0, natural (white) | PA66 UL94 V-0, colour white | Match, down to the grade and the colour |
The current row matches including its gauge basis, which is rare
In the previous two rounds of this comparison series, the current rating was the row that did not match, and both times the problem was the conductor rather than the ampere figure. Against the KR3000, the original published an eight-row derating table where our single figure covered three of seven conductor sizes. Against the KR3961, the original’s 10A rode on AWG #16 while our wire window stopped at #18, and our published 7A turned out to be the rating of a different JST 3.96mm product — the high-box VH and the VR series. Here both sides publish 7A and both sides attach it to AWG #18. JST’s catalogue states the rating as “7A AC, DC (AWG #18)” and the applicable wire range in the same block as “AWG #22 to #18”; our specification §4.0 states “Rated Current (Max.) 7A(18AWG)” and “AWG 18#~22#”. The rating and its basis line up on both sides of the comparison, and no derating table is needed to know what we are claiming — because both companies are claiming the same thing about the same conductor. One caveat belongs on the row, and it is ours rather than the original’s. JST publishes no current derating table for SDN at all — the catalogue gives the single 7A figure and the wire range. So neither side tells a customer what the connector carries at 20 or 22 AWG, even though both accept those conductors. A design running 6A on 22 AWG has no published answer on either datasheet, which is the one place where the agreement between the two documents is agreement to say nothing.
And the housing grade matches down to the UL94 V-0
The housing row is where this round quietly beats every previous one, and it is worth saying why. JST states the SDN housing as “PA 66, UL94V-0, natural (white)” — material, flammability grade and colour in one line. Our specification §3.0 states “Housing PA66 UL94 V-0” and our component page adds the colour, white. Neither company is guessing at the other’s material here, and neither is relying on a reader to look up a grade. In earlier rounds this row was consistently the weakest: against the KR2541 the original’s housing was glass-filled polyester and ours was PA66; against the KR3000 the original published “Nylon” and referred elsewhere for the flame-retardant grade while our wire-to-wire female housing was UL94 V-2 rather than V-0; and against the KR3961 the original’s VH catalogue published only “PA, natural (white)” with the grade referred to a separate standards list. Here the two documents say the same thing in the same units, and a customer comparing them needs no third document. What the seventh row buys you in practice is a two-part bill of materials. The original’s board-in construction means the mated system is a housing soldered to the board plus a crimped contact — one fewer part than the wafer-housing-contact stack of a wire-to-board family, one fewer solder joint to qualify, and one fewer line on the drawing. Our KR3963 mirrors that exactly: the specification’s §2.0 parts list contains housing H396301**0101A, terminal T39630BT0101A, and wafer “None”, and the product page’s component selector offers only two components — Housing and Terminal. There is no wafer in the series because there is no wafer in the original’s design, and that is a match of architecture rather than of numbers.
KONNRA KR3963 series SDN 3.96 board-in housing, soldered directly to the PCB
The One Line That Does Not Match Is the Metal
Every family in this comparison series has had a row where the two documents diverge. Until now that row has always been a number — a temperature range, a wire window, a resistance limit, a force value. This round it is a material, and the divergence has a direction: JST specifies phosphor bronze and we specify brass.
| JST SDN | KONNRA KR3963 | |
|---|---|---|
| Contact base metal | Phosphor bronze | Brass (per specification §3.0) |
| Contact finish | Tin-plated (reflow treatment) | Tin plated over nickel |
| Contact part number | SDN-21T-P1.5 |
T39630BT0101A |
| Wire range of the contact | 0.33 to 0.83mm² / AWG #22 to #18 | AWG 18# to 22# |
| Insulation O.D. of the contact | 1.6 to 3.0mm | 2.50mm max |
| Reel or bag quantity | 5,000 per reel | not published |
Two different answers to two different problems, and neither should be described as better without measurement. The base metal decides how much spring the contact can hold and how much current it can carry through a given cross-section. Phosphor bronze is the higher-strength, higher-conductivity-per-unit-spring-force choice; brass is cheaper and easier to form, and it is the common choice where the contact is not asked to act as a spring at elevated temperature. JST’s own documents treat the distinction as consequential: its VH high-box catalogue states that if high current must be branched in parallel, “be sure to use contacts made of phosphor bronze” — naming phosphor bronze as the requirement for that duty. So the original’s written position is that for at least one operating condition, phosphor bronze is not optional. Our contact on this series is brass, and the question a customer should ask is whether the duty they are designing for is one the original would have required phosphor bronze for. The finish is a separate and more interesting difference, because the two approaches solve opposite problems. JST’s contact is “tin-plated (reflow treatment)” — the tin is melted after plating, which fuses the coating, removes the as-plated oxide and produces the smooth, low-whisker-risk surface that reflow tin is chosen for. Ours is tin plated over nickel, where the nickel layer is a barrier between the tin and the copper alloy beneath it, which is the standard defence against the copper-tin intermetallic growth that drives contact resistance up under thermal ageing. So the original’s answer addresses the surface; ours addresses the interface. Both are legitimate, and neither document states a plating thickness — JST gives the process and not the micrometres, we give the layer system and not the micrometres. A customer who needs to compare the two finishes quantitatively has to request thicknesses from both suppliers.
And our own documents disagree about what we ship
This is the finding of the round, and it is entirely internal.
| Source | What it says the terminal is made of |
|---|---|
| PS-KR3963-01 §3.0 (product specification) | “Terminal Brass Tin plated over nickel” |
| PS-KR3963-01 §4.0 | rated current at 18 AWG, no material statement |
| Terminal component page | no material published at all |
| Product page, General Specification | “Material: PA66/UL94/Phosphor Bronze” |
| Terminal part number | T39630BT0101A |
Our specification says brass. Our product page says phosphor bronze. The component page says nothing. And the part number carries the BT code. That code can now be decoded rather than guessed at, and the decoding is worth setting out because a previous article in this series could only offer it as an inference. Our wire-to-board 3.96 specification lists two terminals, T39610PT0101A and T39610BT0101A, and I recorded at the time that the letter very probably selected the material without being able to prove it. Two independent pieces of evidence now settle it. First, the KR3963 specification publishes the material for its single terminal — T39630BT0101A, stated as brass — so in our own part numbering the B position reads brass. Second, JST’s own model-number allocation for this very family prints the rule: “Surface finish: T···Tin-plated (reflow treatment)” and “Material: P···Phosphor bronze.” So P is phosphor bronze and B is brass in both companies’ grammars, and T39630BT0101A is our brass, tin-plated terminal. Put the two facts together and the consequence is uncomfortable. Our product page claims phosphor bronze for a part number that our own specification and the decoded part-number grammar both say is brass — and phosphor bronze is precisely the material the original specifies for this contact. A buyer reading only our product page would conclude we match the original’s material. A buyer reading our specification would conclude we do not. One of the two documents is wrong, and neither is marked as provisional. Three questions follow, and they are the first three in the enquiry list at the end of this guide:
- Is the KR3963 terminal brass or phosphor bronze? The specification says brass; the product page says phosphor bronze; the part-number grammar says brass.
- If it is brass, is that acceptable for the duty the original requires phosphor bronze for — the parallel high-current branching case that JST’s own VH document calls out?
- If a phosphor bronze option exists, is there a part number for it, and will the product page be corrected?
I am not going to resolve the contradiction by picking the document I prefer. The specification is the engineering-controlled document and the one a purchase order references — PS-KR3963-01 revision A1 — and it says brass. The product page is the customer-facing document and it says phosphor bronze. Asked which we ship, the honest answer available from the documents is: the specification says brass, and that must be confirmed in writing before a material declaration is issued from either page.
KONNRA KR3963 series SDN 3.96 crimp terminal, brass tin plated over nickel per specification
The Wire Window: Tighter at the Top, Open at the Bottom
This family produces the sharpest insulation-window contrast of any round so far, because the original publishes a corridor and we publish a ceiling — and this time the gap at the top is half a millimetre rather than a few hundredths.
| JST SDN | KONNRA KR3963 (specification) | KONNRA product page | |
|---|---|---|---|
| Insulation O.D., contact | φ1.6mm to φ3.0mm | 2.50mm maximum, no minimum | 2.8mm maximum |
| Conductor | AWG #22 to #18 | AWG 18# to 22# | 18#–22# |
Three consequences follow, and they do not all point the same way.
- Against the specification, we are 0.5mm tighter at the top. A wire with 2.6mm to 3.0mm insulation is inside the original’s published range and outside ours. That is not a hairline exclusion — half a millimetre of insulation is the difference between a standard 18 AWG PVC wall and a thicker jacket, and it will reject wires the original accepts.
- Against the product page, we are only 0.2mm tighter. The same wire at 2.9mm passes the page and fails the specification. Which of our two figures applies is therefore not a detail but a design decision.
- At the bottom end neither of our figures imposes a floor, where the original requires 1.6mm. A thin-wall wire of 1.3mm insulation would be rejected by the original and accepted by both of our documents. This is the same asymmetry the previous round found against the KR3961, and it is worth naming as a family-wide pattern rather than a one-off: where our specification publishes only a maximum, it silently removes a constraint the original states.
Ask for the insulation diameter as a number, and ask which document governs it. The gap between 2.50mm and 2.8mm is a real order-line risk when a customer has selected a wire against one page and the purchase order cites the other.
Temperature: the House Figure Comes Back
Two rounds ago I wrote that the temperature row had finally matched, and one round ago I confirmed it a second time. This round it does not match again — and the reason is worth stating precisely, because it corrects a conclusion a reader might have drawn from the two rounds in between.
| Round | Original family | Original’s published range | Our specification |
|---|---|---|---|
| 17 (KR3000) | Molex Micro-Fit 3.0 | −40 to +105°C (tin) | −40 to +105°C |
| 18 (KR3961) | JST VH | −40 to +105°C | −40 to +105°C |
| 19 (KR3963) | JST SDN | −25°C to +85°C | −40 to +105°C |
JST’s VH family reaches −40 to +105°C and its SDN family does not. The two families share a pitch and share a manufacturer and their temperature ranges differ by 15°C at the cold end and 20°C at the hot end. Our specification publishes −40 to +105°C for both of them. What that means for the two rounds is different, and the difference matters. Against VH, our figure was exact. For the KR3961 the original’s own crimp VH catalogue publishes −40 to +105°C, so our range matched its counterpart and the match stood. Against SDN, our figure is wider than the original’s on both ends. Our cold limit is 15°C below the original’s floor and our hot limit is 20°C above its ceiling. The direction of that claim is the one that needs care: we are publishing a broader operating range for the cross-reference than the part it replaces is rated for. For a customer running at −30°C or at 95°C, the limiting figure is the original’s −25 to +85°C, not ours — and a design already qualified at −25 to +85°C does not become qualified to −40 to +105°C because a substitute datasheet says so. So the correct summary across the three rounds is not that our temperature figure tracks the original’s, and not that it ignores it either. It is that our house figure is fixed and the original’s varies by family — which means the temperature row must be checked family by family and can never be assumed. Two consecutive matches were a property of those two families rather than evidence of a trend, and this round is the correction. The same warning applies to the SDN family itself, on the original’s side. JST’s three 3.96mm products do not share a temperature range: the standard crimp VH is −40 to +105°C, the VH high-box type is −25 to +85°C, the VH quick-connect type is −25 to +105°C, and SDN is −25 to +85°C. A cross-reference that quotes “the JST 3.96mm temperature range” is quoting something that does not exist.
What a Board-in Part Does Not Need
A reader who has compared this series’ earlier guides will expect a section on the force ladder, because every wire-to-board round has had one. This family does not have one, and the reason is architectural rather than a gap in our documentation. A board-in contact is pressed into its housing once, during assembly, and is not repeatedly mated afterwards. There is no plug to insert and withdraw against a header, so a table of mating force against circuit count has nothing to describe. Our specification handles this correctly. It publishes no per-circuit insertion/withdrawal table at all — §7 carries no force ladder and §8.0 is the Remark section rather than a force table — and it marks the entire wafer block as N/A, for base material, contact and solder tab, in both the straight and the right-angle variants. That is a form being answered rather than skipped, and it is the correct treatment. What matters for a board-in part is the contact-level pair, and that we do publish:
| Clause | KONNRA KR3963 | JST SDN |
|---|---|---|
| Terminal insertion force | 1.5 kgf (14.7 N) Max. — the crimped terminal inserted into the housing | not published |
| Terminal-to-housing retention force | 3.0 kgf (29.4 N) Min. — axial pull-out on the assembled terminal at 25.4 ± 3 mm per minute | not published |
| Crimp geometry | Width 1.80 ± 0.1mm; conductor crimp height 1.30 / 1.10 / 0.90 ± 0.05mm at 18 / 20 / 22 AWG; insulation crimp height 2.20 / 2.10 / 2.00mm max | not published |
| Crimp strength | 9.07 / 6.80 / 4.54 kgf Min. at 18 / 20 / 22 AWG | not published |
| Stripping length | 3.5 to 4.2mm | not published |
| Durability | no clause in the specification | not published |
Two observations belong on that table, and the first is a credit. The insertion-force maximum of 14.7N and the retention minimum of 29.4N are the two numbers that describe the only mechanical operation a board-in contact performs, and both are published with the test rate and the standard — EIA-364-13D at 25.4 ± 3 mm per minute — which is the same method family used across this comparison series. A customer can therefore reproduce both tests without asking us how they were done. The original publishes neither figure for SDN, so there is nothing to compare against; these are capabilities we document and JST’s SDN catalogue does not. The second observation is that the stripping length is the longest in this series. At 3.5 to 4.2mm the KR3963 requires a longer conductor exposure than the KR3961’s 2.5 to 3.0mm or the KR3000’s 2.5 to 3.0mm. A harness shop running both series on one bench needs two settings, and a crimp that has been set up for one series will not transfer.
The numbering defect our board-in specifications share
Our KR3963 specification numbers its vibration clause §7.32. The sequence runs §7.1 Temperature Rise, §7.32 Vibration, §7.3 Shock, §7.4 Heat Resistance — so the vibration clause should read §7.2 and does not, and every cross-reference to it by number in any downstream document will point at a clause that does not exist. This is the second time this exact defect has appeared in our own board-in documentation, and it is worth naming as a pattern rather than a slip: the same §7.32 appeared in the specification for our other board-in series. The practical instruction is unchanged from that round and bears repeating: cite our clauses by subject rather than by number, because the numbering in this family is not reliable. One more layout oddity in the same document. §7.10, the solder-resistance clause, is printed after §8.0, the Remark section, on the final page. The clause itself is complete — 250°C for 3 to 5 seconds, measured 0.8mm from the terminal tip, per EIA-364-71B, appearance no damage — but a reader who stops at the Remark will not find it, and a reader who indexes the document by section number will look for §7.10 in section 7. And a third, smaller one: the specification publishes no durability clause where our own 3.96mm wire-to-board specification does. PS-KR3961-01 carries a §7.1 durability requirement of 30 mating cycles at 10 cycles per minute, verified by a contact resistance limit of 40mΩ afterwards. PS-KR3963-01 has no equivalent. For a board-in part that is defensible — the contact is mated once, so a mating-cycle count measures something the product does not do. But it means the series publishes no quantified life figure at all, and the original’s SDN catalogue publishes none either, so neither side can tell a customer how many insertion cycles the housing tolerates before retention falls below 29.4N. That is the one genuinely missing number in this comparison, and it is missing on both sides.
What the Original Publishes That We Do Not
Every item below is a figure or document JST publishes for SDN and ours does not. The correct response to each is a document request, not an assertion. The applicable PC board thickness. JST publishes 1.6mm. Our specification publishes no board thickness at all, which for a connector that is soldered directly into the board is the more conspicuous omission — the housing is the board-mounted part, so the board thickness sets the solder-joint geometry with nothing in between. The board layout, with its two conditions. JST publishes a layout drawing “viewed from soldering side”, with the note that “tolerances are non-cumulative: ±0.05mm for all centers” and — importantly — that “for connectors with 5 or more circuits the hole dimensions will differ from the above values. Contact JST for details.” It adds that “hole dimensions differ according to the type of PC board and piercing method” and that the published dimensions “should serve as a guideline.” We publish no layout and no hole-pitch tolerance, so a footprint has to be requested. The housing dimensions, for every circuit count. The SDN catalogue publishes an A and a B dimension for all eleven housings — 3.96mm and 8.96mm at two circuits rising to 43.56mm and 48.56mm at twelve. We publish none, and the KR3963 series drawing is a vector PDF with no text layer, so no dimension in this guide was taken from a drawing. The contact’s material and finish, stated as a material and a finish. JST prints “Phosphor bronze, tin-plated (reflow treatment)” against the contact part number. Our terminal page publishes no material at all, and our specification’s statement of it is the one that disagrees with our product page. This is the item on this list that matters most. The contact’s wire window in square millimetres as well as AWG. JST publishes 0.33 to 0.83mm² alongside AWG #22 to #18, which lets a metric-market designer check the conductor without a conversion. We publish AWG only. Reel and bag quantities. 5,000 contacts per reel; housings at 1,000 per bag for 2 to 5 circuits and 500 per bag for 6 to 12. We publish none of these, and our package specification for this series is a PDF with no extractable text layer. The crimp tooling, named. JST publishes the AP-K2N crimping machine, the MKS-L applicator, the MK/SDN-21-15 die and the APLMK SDN21-15 crimp applicator with dies, and notes that a fully automatic applicator is available on request. We publish no tooling identification. The agency registrations, with file numbers. JST prints Recognized E60389, Certified LR20812 and R50100763 on the catalogue page — a UL recognition, a CSA certification and a third registration, each with its number. Our component pages publish UL E482542 only. A customer whose approval file references a CSA registration has a gap to close with us even though the UL file exists, and the original’s file numbers make that a one-line comparison rather than an argument. And the ordering grammar. JST publishes the model-number allocation for both the contact and the housing: the contact code carries the wire range, the surface finish (T for tin-plated reflow treatment), the base material (P for phosphor bronze) and the terminal size, and the housing code carries the circuit count and the shape option where no suffix means with the polarizing boss and A means without it. Our own ordering code for the housing, H396301**0101A, is not decoded in the public documents, so a customer needing a non-standard circuit count or the without-boss variant has no route from our pages to a part number — and note that the original’s -A option, the housing without the polarizing boss, has no published counterpart in our cross-reference at all.
Where Our Own Documents Disagree With Each Other
This round’s disagreements are almost entirely internal, and two of them are the kind that reach a customer’s material declaration. All are reported rather than smoothed over. 1. Our specification and our product page name two different base metals for the same terminal. The specification says brass; the product page says Phosphor Bronze; the component page says nothing; the part number carries the B that our own numbering uses for brass. The full analysis is in the metal section above — the first question to put to us about this series. 2. The insulation diameter has two values, and this is the fourth time this field has been wrong in this series of reviews.
| Source | Insulation O.D. |
|---|---|
| PS-KR3963-01 §4.0 | 2.50mm Max |
| Product page, General Specification | 2.8mm Max |
| Terminal component page | 2.80mm(MAX) |
The website’s figure is 0.3mm looser than the specification’s, and it is the looser figure that a buyer reads. A wire with 2.7mm insulation passes both pages and fails the specification — and the failure appears at the crimp station rather than at the quotation stage. This same field has now been found holding a foreign value in the 2.5mm board-in series (2.50mm, where the correct value was 1.60mm), the 2.54mm series (2.54mm, the pitch, where the specification says 1.63mm), the 3.00mm series (0.4 to 0.8mm, against a specification that says 1.3 to 1.8mm) and now here two figures that disagree with each other. Four rounds, four defects, one field. 3. The circuit range on the website has no backing in the specification, and it is wider than the original’s. Our product page and housing component page publish 2 to 15 positions; the original publishes 2 to 12 with a dimension table for every count; and our specification publishes no circuit range anywhere — there is no circuit list in §2.0, no dimension table and no force table to bound it. This is not an error in the way the previous two items are, and it deserves stating in our favour: if we tool 13, 14 and 15 positions, that is coverage the original does not offer, and it is a reason to buy from us rather than a defect. What is missing is the supporting document — a customer asked to design a 15-way footprint has nothing from us to design against, and cannot tell from our pages whether 13 to 15 exist as standard parts or as tooling on request. The claim and the documentation need to arrive together. 4. Our contact-resistance figures have no counterpart in the original’s document, and our pages do not say which basis they use. JST’s SDN catalogue publishes no contact resistance at all for this family. Our specification publishes 20mΩ maximum initially by dry circuit at 20mV and 100mA per EIA-364-23C, and 40mΩ maximum after each of the environmental tests — a two-level structure like the one in our 3.96mm wire-to-board specification. Our product page and terminal page publish “20mΩ Max” with no statement that it is the initial, dry-circuit value. Since the original says nothing on this row, a customer is comparing our requirement against nothing — which is fine, provided it is read as a requirement rather than a measurement. 5. The “Material” row on the product page lists a flammability standard among the materials. It reads “Material: PA66/UL94/Phosphor Bronze” — UL94 is not a material, it is the flammability test whose grade our housing meets, and the row also omits any statement that the terminal is a different alloy from the housing’s polymer. The third occurrence of this same row shape in this series of reviews, after the 2.54mm and 3.00mm families. 6. The product page’s advantages include two claims with nothing behind them, and one empty heading.
| Claim on the product page | What our documents support |
|---|---|
| “Long Life and Durability — utilizing high quality materials and advanced production processes, the connectors have a long service life and excellent durability” | Our specification publishes no durability clause and no cycle count, and the original publishes none for SDN. The claim is unquantified on both sides. |
| “Wide range of compatibility — suitable for a wide range of different sizes and types of PCB boards” | We publish no applicable board thickness, while the original publishes 1.6mm. A claim about board compatibility without a board thickness is not actionable. |
| A heading reading “Application” with no content beneath it | — |
7. The three markets named in our Overview are ours, not the original’s. The page lists data centres, communication equipment and industrial automation. JST’s published scope for SDN is “a wide variety of circuits including signal and power supply circuits” — a statement about circuit type, not about industries. Ours is a positioning choice rather than a quotation, and it is worth saying so, because an engineer checking our application claim against the original’s catalogue will not find it there. 8. The component pages carry no series name in their URL or their heading. They resolve at /components/3963-t/ and /components/3963-h/ with page titles reading simply “Terminal” and “Housing”, and a Compatible field reading “SDN Series” — accurate, but with no series number, no part number and no circuit range, while the product page resolves the same two parts with full names. The two-component selector is correct for a board-in series — there is no wafer to list — but a customer searching for the series they were quoted will not find it in these two URLs.
KONNRA KR3963 series SDN 3.96 single row board-in housing
The Cross-Reference Map
| KONNRA component | JST SDN | KONNRA part number | Notes |
|---|---|---|---|
| Housing, with polarizing boss | 2P-SDN … 12P-SDN |
H396301**0101A |
2 to 12 circuits on the original’s side; PA66 UL94 V-0, white |
| Housing, without polarizing boss | 2P-SDN-A … 12P-SDN-A |
no separate part number published | The original’s -A option has no mapped counterpart |
| Terminal | SDN-21T-P1.5 |
T39630BT0101A |
Base metal differs: original phosphor bronze, ours brass per the specification |
| Wafer | — (board-in: the original has none) | None |
Correct — the architecture has no wafer |
| Crimp tooling | AP-K2N, MKS-L, MK/SDN-21-15, APLMK SDN21-15 |
not published | — |
| Board layout | Published, with a ±0.05mm non-cumulative hole tolerance and a 5-circuit caveat | not published | — |
Every JST number in that table is a real SDN number, and the mapping is correct — including the row that says the original has no wafer, because board-in genuinely has none. The uncovered items are the without-boss housing, the tooling and the board layout, and for a customer using any of them the answer is a quotation or a drawing request rather than a substitute. One practical instruction for anyone moving a design across. Our housing ordering code H396301**0101A is not decoded in any public document, and our product page offers no variant for the original’s without-polarizing-boss housing. If your board has no clearance for the boss, that is not a substitution question — it is a new part request, and it should be raised before the board is released rather than after.
Applications Where the SDN Pattern Is Used
JST’s published scope for SDN is a single clause, and it is deliberately broad: the connector is “capable of connecting a wide variety of circuits including signal and power supply circuits.” There is no market or application matrix for this family on the original’s side — no appliance, automotive, industrial or data-centre list — in the way that other manufacturers publish for comparable families. What the original does publish is a set of design intentions, and they are more specific than the scope sentence:
| Published feature | What it implies about the application |
|---|---|
| “Slim, low-profile design” | Designs where board height or enclosure depth is constrained, and where the housing sitting directly on the board is an advantage rather than a saving |
| “Exclusive lance construction to reduce insertion force” | Manual or semi-automatic assembly, where the contact-pressing operation is a throughput constraint |
| “Improved solderability” | A through-hole soldering process on the housing, which is what board-in requires |
| “Polarizing boss” | An orientation-critical assembly, where the boss prevents the wire half being fitted the wrong way |
| Signal and power supply circuits in the same family | Mixed boards where a designer does not want two connector systems for a small number of power lines |
Two implications follow that a cross-reference should carry. The first is that board-in suits a captive wire rather than a serviceable one. Because the housing is soldered down and the contacts are pressed in, field replacement means releasing and re-pressing an individual contact rather than unplugging a wire half. For an appliance or an industrial control that is assembled once and serviced rarely, that is a reduction in parts and height. For a harness that is plugged and unplugged by the end user, it is the wrong architecture — and it is where the KR3961, our wire-to-board 3.96mm series, belongs instead. The second is that the polarizing boss is a design intent, not an accessory. JST ships every circuit count with and without it, and the choice belongs to the board layout. A customer who has designed against a housing without the boss, and receives one with it, has a clearance problem rather than a connector problem — which is the strongest argument for confirming that specific option in writing.
Harness and assembly options
We supply the KR3963 as components, as crimped housing assemblies, and as finished harnesses. Because the series is crimp-terminated on the wire side and board-soldered on the board side, the assembly questions are the crimp geometry and the soldering process rather than a mating interface. Two inputs the original publishes for a harness shop, and we do not: the crimp tooling — the AP-K2N machine, the MKS-L applicator and the MK/SDN-21-15 die — and the package quantities, at 5,000 contacts per reel and 500 to 1,000 housings per bag. Both decide whether an existing bench can run the part, and both are the kind of figure a production engineer asks for before a quotation is signed rather than after. And one input we publish that the original does not: the crimp strength minimum of 9.07 / 6.80 / 4.54 kgf at 18 / 20 / 22 AWG and the stripping length of 3.5 to 4.2mm, both in the specification. A harness shop quoting from an SDN drawing alone will not have these numbers from the original’s side, and they are what the pull-test acceptance is written against.
Sourcing: What Procurement Teams Ask
A cross-reference enquiry fails for commercial reasons as often as technical ones, and this family’s commercial questions are unusually specific because two of them are about which of our own documents governs. Which document governs the material and the wire window? Our specification says brass and 2.50mm insulation; our product page says phosphor bronze and 2.8mm. The purchase order should reference the specification revision — PS-KR3963-01 revision A1 — and name the terminal material explicitly, because the website and the controlled document do not agree and nothing marks either as provisional. Which half of the pair am I buying? This is a board-in series, so there are exactly two components — the housing and the terminal — and no wafer. If your quotation includes a wafer, it is for a different series. Our board-in range contains five families — 1.25mm, 1.5mm, 2.0mm, 2.5mm and this 3.96mm — and the order code decides the family, not the pitch. Is the wire I already use inside the window? AWG 18# to 22#, with an insulation ceiling that is 2.50mm in the specification and 2.8mm on the website. Two exclusions catch people: wires with 2.6mm to 3.0mm insulation, which the original accepts, and any wire below 1.6mm insulation, which the original rejects and our documents do not. Check the gauge and the diameter separately. What is the minimum order quantity and the lead time? Connector production lead time is typically 2 to 4 weeks and wiring harness lead time typically 3 to 4 weeks, with complete connector set samples within 45 days. For this family samples are the right first step, because the two questions most likely to change your design — the base metal of the terminal and the true insulation window — are answered faster with a sample and a written statement than by reconciling our two pages. What documentation arrives with the parts? Ask for the product specification, the package specification, the series drawing and the housing drawing, then for the four documents we do not publish: the applicable board thickness, the board layout with the hole-pitch tolerance, the housing dimensions for your circuit count, and the material declaration for the terminal. What about the second source? The honest answer is that this is the strongest cross-reference in the comparison series on architecture and ratings — board-in matching board-in, seven rows level including the current’s gauge basis and the housing’s grade and colour — and it rests on two documents that do not agree about the contact metal. A design at 250V and 7A on AWG #18, with the board designed for a board-in housing, is the case where the equivalence is straightforward. A design where the contact’s spring properties or the plating finish is the governing factor is the case where the material statement has to be closed first.
The Five Questions I Would Ask Us, In This Order
1. “Is the KR3963 terminal brass or phosphor bronze?” Our specification says brass; our product page says Phosphor Bronze; the terminal component page says nothing; and the part number carries the B that our own numbering uses for brass. The original specifies phosphor bronze for this contact. This is the first question because it is the one our own documents answer two different ways, and because the answer decides whether a material declaration can be issued at all. 2. “Is brass acceptable for my duty, given that the original requires phosphor bronze for parallel high-current branching?” JST’s VH high-box catalogue states that where high current must be branched in parallel, phosphor bronze contacts must be used. If your design branches a 7A-rated circuit in parallel, this is a specification question, not a preference. 3. “Is the insulation ceiling 2.50mm or 2.8mm?” The specification says 2.50mm; the product page and the terminal page say 2.8mm. A wire between the two figures passes our website and fails our crimp process. Ask which document the quotation references. 4. “What is the applicable PC board thickness, and where is the layout?” The original publishes 1.6mm and a full layout with a ±0.05mm non-cumulative hole-pitch tolerance, plus a warning that the published hole dimensions do not apply at 5 circuits and above. We publish no board thickness and no layout, and for a part that is soldered directly into the board that is the first document a mechanical engineer needs. 5. “How many circuits can you actually supply, and is the without-boss housing available?” Our website publishes 2 to 15 positions and the specification publishes no circuit range at all; the original publishes 2 to 12 with a dimension table for every count. And the original’s housing without the polarizing boss, 2P-SDN-A through 12P-SDN-A, has no counterpart in our cross-reference — if your board has no clearance for the boss, that is a new part number rather than a substitution.
Engineer’s Pre-Release Checklist
Nine checks, in the order the failures actually happen.
- Board thickness. Confirm it against our documents rather than assuming — we publish no figure, and the original publishes 1.6mm. For a housing soldered directly into the board, this is the check that precedes everything else.
- Footprint and the 5-circuit caveat. The original’s layout is “viewed from soldering side” with a ±0.05mm non-cumulative hole-pitch tolerance, and its published hole dimensions do not apply from 5 circuits upward without asking. Get the layout for your specific circuit count from us.
- Polarizing boss. Confirm whether you need the housing with or without the boss. The original ships both; our cross-reference maps only the with-boss housing, so the without-boss variant is a part request.
- Terminal material. Confirm brass or phosphor bronze in writing — our specification and our product page disagree — and check the answer against JST’s statement that phosphor bronze is required where high current is branched in parallel.
- Wire gauge and insulation diameter, separately. AWG 18# to 22# on both sides; the insulation ceiling is 2.50mm in our specification and 2.8mm on our website, against the original’s φ1.6 to 3.0mm corridor. Check both ends against the original’s range, not against ours alone.
- Temperature. Ours is −40 to +105°C; the original’s is −25 to +85°C. The limiting figure is the original’s, and a design qualified at −25 to +85°C is not extended by a substitute’s wider claim.
- Current at your conductor. Both sides publish 7A at AWG #18 — a genuine match — but neither publishes a derating table, so at 20 or 22 AWG there is no published answer on either datasheet. Ask for it.
- Crimp settings. Conductor crimp height 1.30 / 1.10 / 0.90 ± 0.05mm and stripping length 3.5 to 4.2mm at 18 / 20 / 22 AWG. The stripping length is the longest in our 3.96mm range, so a bench set up for the wire-to-board series needs a second setting.
- Life and retention. Confirm the 3.0kgf (29.4N) minimum terminal-to-housing retention and the 1.5kgf (14.7N) maximum insertion force, both reproducible per EIA-364-13D at 25.4 ± 3 mm per minute — and note that neither company publishes a durability cycle count for this family.
Frequently Asked Questions
What is a JST SDN connector?
It is a 3.96mm pitch board-in crimp connector — JST’s own description is “a 3.96mm pitch board-in crimp style connector capable of connecting a wide variety of circuits including signal and power supply circuits” — rated 7A AC/DC at AWG #18, 250V AC/DC, −25 to +85°C, withstanding 1,500V AC/minute, insulation resistance 1,000MΩ minimum, for wire AWG #22 to #18 with insulation φ1.6 to 3.0mm, on a 1.6mm board, in 2 to 12 circuits. It is slim and low-profile, uses an exclusive lance construction to reduce insertion force, offers improved solderability and a polarizing boss, and JST lists it as Recognized E60389, Certified LR20812 and R50100763.
Is the SDN connector wire-to-board or wire-to-wire?
Neither, in the usual sense — it is board-in. The housing is soldered directly into the PCB and the crimped contact is pressed into the housing from above, so the wire terminates in a contact that plugs into a board-mounted housing. There is no header and no wafer, and the mated system is two parts rather than three.
What is board-in, and how does it differ from a header-based connector?
In a header-based wire-to-board family the board carries a header, the wire carries a housing and a contact, and the housing plugs onto the header — three parts and a repeated mating operation. In a board-in family the housing is the board-mounted part: it is soldered down, the contact is pressed in once at assembly, and there is no repeated mating. The result is fewer parts, one less solder joint to qualify, and a lower profile — at the cost of a wire half that is not designed to be unplugged by the end user.
Is JST SDN the same as JST VH?
No — they share a pitch of 3.96mm and nothing else that matters. VH is a wire-to-board crimp family with headers (top entry and side entry), housings, contacts and an optional retainer, rated 10A AC/DC at AWG #16 on the standard header with a range of 2 to 11 circuits and −40 to +105°C. SDN is a board-in crimp family with no header at all, rated 7A AC/DC at AWG #18 over 2 to 12 circuits and −25 to +85°C, for a narrower wire range of AWG #22 to #18. JST also sells a high-box VH and a quick-connect VH, and a separate VR series at the same pitch. Pitch is not a family.
What is the KONNRA equivalent of the JST SDN?
The KR3963 series, a 3.96mm board-in system covering the housing H396301**0101A and the terminal T39630BT0101A, with no wafer — matching the original’s architecture part for part. Our published ratings are 250V AC/DC, 7A at AWG #18, −40 to +105°C, AWG 18# to 22#, insulation 2.50mm maximum, contact resistance 20mΩ maximum, insulation resistance 1,000MΩ minimum at 500VDC and 1,500V AC for one minute, in PA66 UL94 V-0 white.
What current and voltage does the SDN carry, and what does the KR3963 carry?
Both 7A AC/DC at AWG #18, and both 250V AC/DC. This is the row that matched most closely in this comparison series, because the gauge basis matches too — JST states the rating as 7A at AWG #18 and our specification states it as 7A(18AWG). The caveat is that neither side publishes a current derating table, so at 20 or 22 AWG — both of which both sides accept — neither datasheet answers what the connector carries.
What wire gauge and insulation diameter does the SDN accept?
Wire: AWG #22 to #18 (0.33 to 0.83mm²). Insulation: φ1.6mm to φ3.0mm. Our wire range matches exactly. Our insulation window does not: the specification says 2.50mm maximum with no minimum, the product page says 2.8mm maximum, against the original’s corridor. So we are 0.5mm tighter at the top than the original and impose no floor where the original requires 1.6mm.
What is the contact material of the SDN, and what is ours?
JST specifies the contact as phosphor bronze with a tin-plated (reflow treatment) finish. Our specification says brass, tin plated over nickel, for the part number T39630BT0101A — whose B our own numbering uses for brass — while our product page claims Phosphor Bronze for the same part number. JST’s model-number grammar independently prints “P···Phosphor bronze” and “T···Tin-plated (reflow treatment)”, so the code is decodable and points to brass on our side. This is the one line in the comparison that does not match, and it is the question to settle first.
What is the contact resistance of the SDN?
JST publishes none for this family — the SDN catalogue gives current, voltage, temperature, withstanding voltage, insulation resistance, wire range and board thickness, and no contact resistance figure. Our specification publishes 20mΩ maximum initially by dry circuit at 20mV and 100mA per EIA-364-23C, and 40mΩ maximum after each of the environmental tests. Our product and terminal pages publish “20mΩ Max” without stating that it is the initial, dry-circuit value.
What are the insulation resistance and withstanding voltage?
Both 1,000MΩ minimum and 1,500V AC for one minute — an exact match. Ours specifies the insulation-resistance test conditions as 500VDC applied for one minute between adjacent contacts per EIA-364-21B, which is what a 1,000MΩ figure means something against. The one departure is post-humidity insulation resistance: our §7.6 requires 100MΩ minimum where the original’s general requirement is 1,000MΩ.
What is the operating temperature range?
JST: −25°C to +85°C, including temperature rise from applied current. KONNRA: −40 to +105°C — wider than the original at both ends. That is the safe-for-us, unsafe-for-you direction: the original’s range governs a design qualified against it, and our broader figure does not extend that qualification. Note also that JST’s own 3.96mm products do not share one temperature range — VH is −40 to +105°C, high-box VH and SDN are −25 to +85°C, quick-connect VH is −25 to +105°C.
How many circuits does the SDN come in?
JST: 2 to 12, with a published A and B dimension for every count, and every count available with the polarizing boss (nP-SDN) or without it (nP-SDN-A). Our website publishes 2 to 15 positions and our specification publishes no circuit range at all. If 13 to 15 positions exist in our tooling that is coverage the original does not offer, but the supporting document does not exist, so confirm the range against a specific part number.
Does the KR3963 support the housing without the polarizing boss?
Not in our cross-reference. The original ships 2P-SDN-A through 12P-SDN-A, the same housing without the boss, at the same materials and quantities. Our cross-reference maps only the with-boss housing, and our ordering code is not decoded in the public documents, so a board with no clearance for the boss needs a part request rather than a substitution.
What UL, CSA and TÜV files does the SDN carry?
JST prints Recognized E60389, Certified LR20812 and R50100763 on the SDN catalogue page — each with its number. Our component pages publish UL E482542 only. If your approval file references a CSA registration, that is a gap to close with us.
Is the KR3963 a drop-in replacement for the JST SDN?
On architecture and on the electrical ratings, this is the closest cross-reference in this series. The board-in construction matches part for part with no wafer on either side; seven published rows are level, including the current rating’s gauge basis (7A at AWG #18) and the housing’s material, flammability grade and colour (PA66, UL94 V-0, white). On three points it is not: the contact base metal differs — the original specifies phosphor bronze and our specification says brass; our insulation ceiling is tighter at the top and open at the bottom against the original’s φ1.6 to 3.0mm; and our published temperature range is wider than the original’s on both ends. Treat it as a drop-in for a ≤250V, 7A, 18–22 AWG board-in requirement once the terminal material is confirmed in writing.
How long does it take to get samples?
Complete connector set samples within 45 days. Connector production lead time is typically 2 to 4 weeks and wiring harness lead time typically 3 to 4 weeks. For this family, say which housing variant you need — with or without the polarizing boss — and include the circuit count, the conductor, the insulation diameter and the board thickness in the enquiry, because those four fields are where this cross-reference is decided.
Start Your Cross-Reference Check
KONNRA supplies the KR3963 series as individual components, crimped housing assemblies or complete wire harnesses, with customisation available for application-specific requirements.
- Request a quote — KR3963 pricing, MOQ and configuration for your circuit count and wire
- Request a sample — complete connector set samples within 45 days
- Request cross-reference verification — confirm KR3963-to-SDN equivalence against your specific original part number
- Request the terminal material in writing — brass or phosphor bronze, with the answer reconciled against the product page’s Phosphor Bronze claim and against JST’s requirement of phosphor bronze for parallel high-current branching
- Request a material declaration — issued from the controlling document rather than from the product page
- Request the insulation window reconciled — 2.50mm per the specification against 2.8mm per the product page and the terminal page
- Request the applicable PC board thickness — JST publishes 1.6mm and we publish nothing
- Request the board layout for your circuit count — including the hole-pitch tolerance and the original’s caveat that its published hole dimensions do not apply from 5 circuits upward
- Request the housing dimensions — JST publishes A and B for all eleven counts and we publish none
- Request the without-polarizing-boss housing —
2P-SDN-Athrough12P-SDN-Aon the original’s side has no counterpart in our cross-reference - Request the circuit range for the exact part number — our website publishes 2 to 15 and the specification publishes none
- Request the current derating answer — neither side publishes a table, and both accept 20 and 22 AWG
- Request the crimp tooling identification — applicator and die references for the conductor you are using
- Request the reel and bag quantities — 5,000 contacts per reel and 500 to 1,000 housings per bag on the original’s side; we publish none
- Request the crimp strength and stripping data — 9.07 / 6.80 / 4.54 kgf minimum and 3.5 to 4.2mm, which we publish and the original does not
- Request the durability answer — neither company publishes a cycle count for this family
- Request the agency statement — our UL E482542 file and a statement on CSA coverage against JST’s Certified LR20812
- Request the plating thicknesses — the original specifies a reflowed tin finish and we specify tin over nickel, and neither document gives micrometres
- Request the contact resistance basis — our 20mΩ is the initial, dry-circuit value per §5.1 and our pages do not say so
- Request the specification revision — PS-KR3963-01 revision A1, with the terminal material and the circuit count named in the purchase order
- Request a vendor qualification pack — certificates, test capability summary, RoHS documentation, quality system records
- Submit a drawing for review — we will flag any specification mismatch before you commit tooling or a board respin
Contact KONNRA Electronics
- Phone: (86)-769-85449875
- Email: info@konnra.com
- Address: No.6 Nanchang South Road, Chijiao, Wangniudun, Dongguan, Guangdong, China
- Contact us
👉 KR3963 product page · KR3963 housing · KR3963 terminal · Board-in connector range · KR2508 Series SCN Board In Connector · KR3961 Series VH 3.96 board guide · JST XA 2.5 Connector Complete Guide (KR2516) · JST SCN 2.5 Connector Complete Guide (KR2508) · Molex Micro-Fit 3.0 Connector Complete Guide (KR3000) · Wire-to-board connector range · Wiring harness range
Sources and method. Every figure here is taken from a manufacturer document, and where two manufacturers — or two documents from the same manufacturer, or two documents from KONNRA — disagree, the difference is stated rather than averaged or resolved. JST figures come from the SDN connector catalogue eSDN.pdf, which is the family’s product specification document. From it: the scope statement (“The SDN connector is a 3.96mm pitch board-in crimp style connector capable of connecting a wide variety of circuits including signal and power supply circuits. The connector features a slim, low-profile design. In addition, it has an exclusive lance construction to reduce insertion force, thus substantially improving the working efficiency in mounting the contact.”); the four named features (slim, low-profile design; easy and secure contact mounting; improved solderability; polarizing boss); the ratings (current 7A AC, DC (AWG #18); voltage 250V AC, DC; temperature range −25°C to +85°C including temperature rise in applying electrical current; withstanding voltage 1,500V AC/minute; insulation resistance 1,000MΩ min.; applicable wire AWG #22 to #18; applicable PC board thickness 1.6mm); the standards block (Recognized E60389, Certified LR20812, R50100763); the RoHS statement; the board-layout notes (“viewed from soldering side”, “Tolerances are non-cumulative: ± 0.05mm for all centers. For connectors with 5 or more circuits the hole dimensions will differ from the above values. Contact JST for details.”, “Hole dimensions differ according to the type of PC board and piercing method. The dimensions above should serve as a guideline.”); the contact table (SDN-21T-P1.5; 0.33 to 0.83mm²; AWG #22 to 18; insulation O.D. 1.6 to 3.0mm; 5,000 per reel; phosphor bronze, tin-plated (reflow treatment)); the crimping tooling table (AP-K2N machine, MKS-L applicator, MK/SDN-21-15 die, APLMK SDN21-15 crimp applicator with dies, with the note “Contact JST for fully automatic crimping applicator”); the housing table giving with-boss versions 2P-SDN to 12P-SDN and without-boss versions 2P-SDN-A to 12P-SDN-A, each with an A and a B dimension (3.96/8.96mm at 2 circuits through 43.56/48.56mm at 12), quantity per bag of 1,000 for 2 to 5 circuits and 500 for 6 to 12, and the material PA 66, UL94V-0, natural (white); and the model-number allocation for both the contact (applicable wire range AWG #22 to #18; surface finish T·Tin-plated (reflow treatment); material P·Phosphor bronze; terminal size) and the housing (no. of circuits 2 to 12; shape: none = with polarizing boss, A = without polarizing boss). JST’s SDN catalogue publishes no contact resistance, no durability cycle count, no insertion or withdrawal force, and no current derating table, and no other JST SDN document was obtainable, so none of those four rows is compared against the original in this guide. KONNRA figures come from PS-KR3963-01 revision A1 (2022/2/26, 6 pages — 3.96mm pitch KR3963 series connector specification); the KR3963 product page; the two KR3963 component pages (/components/3963-h/ and /components/3963-t/); and the KR3963 series and component drawings (KR3963-Series-Drawing, KR3963-H_1, KR3963-T_2). From the specification: the parts list (housing H396301**0101A; terminal T39630BT0101A; wafer “None”); the materials (housing PA66 UL94 V-0; terminal Brass Tin plated over nickel; wafer rows for straight and right-angle marked N/A for base, contact and solder tab); the ratings (250V AC/DC; 7A (18AWG); −40 to +105°C; AWG 18# to 22#; insulation O.D. 2.50mm maximum); §5.1 contact resistance of 20mΩ maximum by dry circuit at 20mV and 100mA per EIA-364-23C; §5.2 insulation resistance of 1,000MΩ minimum at 500VDC per EIA-364-21B; §5.3 dielectric strength of 1,500V AC for one minute with no breakdown or flashover per EIA-364-20A; §6.1 terminal insertion force of 1.5kgf (14.7N) maximum; §6.2 terminal-to-housing retention of 3.0kgf (29.4N) minimum at 25.4 ± 3 mm per minute per EIA-364-13D; §6.3 crimp geometry and strength of 9.07 / 6.80 / 4.54 kgf minimum at 18 / 20 / 22 AWG with a stripping length of 3.5 to 4.2mm; §7.1 temperature rise of 30°C maximum; the vibration clause numbered §7.32 with amplitude 1.5mm peak-to-peak, sweep 10→55→10 Hz in one minute and duration two hours in each axis per EIA-364-28B; §7.3 shock of 490m/s² (50g), three strokes in each axis per EIA-364-27B; §7.4 heat resistance of 96 hours at 105 ± 2°C; §7.5 cold resistance of 96 hours at −40 ± 2°C; §7.6 humidity of 96 hours at 40 ± 2°C and 90–95% R.H. requiring 40mΩ maximum, dielectric strength per §5.3 and insulation resistance of 100MΩ minimum; §7.7 thermal shock over 5 cycles of −40°C / room / +105°C / room; §7.8 salt spray of 8 hours at 35 ± 2°C from a 5 ± 1% solution; §7.9 solderability at 245 ± 5°C for 3 ± 0.5 seconds with 95% wetting; and §7.10 solder resistance at 250°C for 3 to 5 seconds measured 0.8mm from the terminal tip per EIA-364-71B. Notes on what was not obtainable. The KR3963 series and component drawings are vector PDFs with no text layer, and the KR3963 package specification likewise carries no extractable text, so no dimension in this guide was taken from a drawing — every KONNRA figure above comes from the product specification or a page. PS-KR3963-01 publishes no circuit range, no applicable board thickness, no board layout and no durability clause, which is why those rows read as gaps rather than values. Where a KONNRA figure and a JST figure differ, both are printed side by side. Where a KONNRA document disagrees with another KONNRA document — the terminal base metal, the insulation diameter, the circuit range, the “Material” row listing a flammability standard among materials, the contact-resistance basis not being stated on the pages, and the two unsupported advantage claims — the discrepancy is reported rather than smoothed over, and the internal contradiction is treated as the more urgent of the two. Every force comparison is stated with its basis: our figures are per contact and published in both kilogram-force and newtons with the test rate and standard, and because the original publishes no force data for SDN there is no figure above to bracket them against.
