Nintendo 3DS Family Repair Reference

This page covers all six retail models of the Nintendo 3DS family: the CTR-001 3DS, the SPR-001 3DS XL, the FTR-001 2DS, the KTR-001 New 3DS, the RED-001 New 3DS XL and the JAN-001 New 2DS XL. Across US, EU and Japanese units.

Two things before you read further, because both change how you should weight what follows.

First: I have not had a 3DS-family board on my own bench. This page is built from reverse-engineered references, manufacturer datasheets, community repair documentation and board photographs, not from my own repair count. Where I would normally say “I see this constantly,” I have written “the sources report” instead. The data tables at the bottom carry per-item confidence tags for the same reason, and nothing on this page is tagged bench.

Second: there is no OEM service manual, schematic or boardview for any 3DS-family board, and that was re-checked rather than assumed. Nintendo never released one and none has leaked. What exists instead is Martin Korth’s GBATEK pages, reverse-engineered from real boards; the 3dbrew wiki, written by different authors using a different method, including a test-point survey actually measured with a scope; Nintendo’s own operations manuals, which are product documentation and never name a chip; manufacturer datasheets for the handful of catalogue parts; and the boards’ own silkscreen, read at magnification off high-resolution photographs. That silkscreen turns out to be the strongest pin-1 datum available on this platform, because it is independent of both wikis.

The consequence is that almost every numbered pinout below came off exactly one board, C/CTR-CPU-01, the original 3DS. It is the only 3DS-family mainboard with a complete published connector map and a silkscreened test-point field. Do not carry a designator or a test-point number onto another board without buzzing it.

Two organising ideas for this machine, and they pull in the same direction.

This is not a capacitor console. Skip the retro reflex. The failure record here is mechanical, connector and fuse, and a blanket recap is not a service, it is an invented job.

The model name on the box is a bad guide to the part you need to order. The battery codes do not follow the model names, the SD reader is modular on some boards and soldered on others in a pattern that is not an Old-versus-New split, and one of the three New models has no button board despite the other two having one. Nearly every expensive mistake on this platform is somebody generalising from the model name.

Safety first: the cell, not the console

Nothing on a 3DS board will shock you. There is no high-voltage backlight inverter here and no bias rail worth respecting. If you want the console in my reference set that can actually bite a person, that is the Game Gear with its +34 V rail and kilovolt inverter.

The hazard on this platform is the lithium cell, and it is a real one.

A swollen pack is a fire and mechanical hazard. Do not puncture it, do not pry against its face, and do not put a swollen cell back in a shell to “test it one more time.” Swelling is also destructive in slow motion before it is dangerous: an expanding pack bows the lower shell, lifts the board off its mounts, and pushes the top half open against the hinge until the hinge cracks. A console that arrives with a cracked hinge and a bulged back cover usually has one root cause, not two.

The second hazard is to the board rather than to you, and it comes from the charge port.

Do not ground pin 5 of the DC input connector. A widely read repair answer speculates that pin 5 “is probably a ground for the connector shell.” It is not. Pin 5 is the second 4.6 V input, the sprung contact the charging cradle uses, and tying it to ground shorts the adapter straight through fuse F1. The shields are pins 3 and 4. The full pad order, the silkscreened pin-1 triangle, and what each source does and does not actually cover are on the P2 card in the component data below.

And the standing rule that this platform in particular punishes: never leave a fuse jumpered. Both board fuses blow because of something downstream. Find the short first.

Know your board first

Grade the shell from across the room, then read the mainboard code before you order anything.

Model	Code	Generation	Tells
Nintendo 3DS	CTR-001	Old	smallest clamshell, glossy top bezel, one circle pad, no ZL/ZR, uncoloured ABXY
Nintendo 3DS XL / LL	SPR-001	Old	large rounded clamshell, one circle pad, rectangular Start/Select/Home
Nintendo 2DS	FTR-001	Old	non-folding slate, no 3D slider, physical sleep switch, mono speaker
New Nintendo 3DS	KTR-001	New	small clamshell, interchangeable faceplates, C-stick, ZL/ZR, coloured ABXY
New Nintendo 3DS XL / LL	RED-001	New	large clamshell, C-stick, ZL/ZR, camera above the top screen
New Nintendo 2DS XL / LL	JAN-001	New	folding clamshell, camera between the screens, no 3D slider, Home button bottom-left

The quick discriminator: New models have a C-stick, ZL and ZR, and microSD. Old models have full-size SD and none of those. “XL” is the Americas and European label; “LL” is the Japanese label for the same hardware.

Inside, the mainboards are silkscreened to a family convention of PWB/<assembly>/<factory> <MODEL>-CPU-NN. A listing quoting only “SPR-01” or “RED-01” is quoting a fragment of that string, not a different board.

Model	Mainboard code	Test-point text layer	Published connector map
CTR-001	`C/CTR-CPU-01`	yes, TP1-TP259 plus cut-links	full
SPR-001	`SPR-CPU-01`	yes	eMMC pads only
FTR-001	`FTR-CPU-01`	yes, including TP260 and up	eMMC pads only
KTR-001	`KTR-CPU-01`	none	none
RED-001	`RED-CPU-01`	none	chip list only
JAN-001	`JAN-CPU-01`	none	none

The three New boards have test points but no text layer on the silkscreen. That is a physical property of those boards, stated by GBATEK and confirmed by the board photographs, not a research shortfall. Do not go looking for a TP number on a New board.

Three board-identity facts decide most parts orders, and none of them follow the model names:

SD reader construction is not an Old-versus-New split. CTR-001, KTR-001 and RED-001 all use a modular reader that unplugs and swaps as a unit. SPR-001 and JAN-001 are soldered. Get this backwards and you desolder a reader that only needed reseating.
Only two of the three New models have a button board. RED-001 has a full one carrying the ABXY, C-stick, ZL/ZR and power functions plus a fuel gauge and a magnet sensor. KTR-001 has a power and ABXY board silkscreened “ABXY-01”, but which chips it carries is undocumented. JAN-001 has no button board at all and its power switch is on the mainboard.
“100 % interchangeable” is a fitment claim, not a layout claim. A later CTR-001 board revision has been found carrying a sixth pad in the eMMC test-point cluster that the documented board does not have. Boards that drop into the same shell can still differ where you are about to solder.

Common problems and fixes

Ordered by how often the community reports each one. That is a triage prior, not measured data. Nobody has published 3DS failure statistics, and I have no count of my own to offer instead.

What stands out about this list compared with every other console on this site is how little of it is electrical. The top five faults are a fatigued flex, a torn connector, another fatigued flex, a spent battery and cracked plastic.

Shoulder button dead or intermittent

The most common fault on the platform. L or R stops registering, or works only on a hard press, or drops out during rapid inputs.

The shoulder button is a metal dome on a small flex ribbon that plugs into a socket on the mainboard. The dome cracks, or the ribbon fatigues and tears at the fold where it is actuated.

Reseat before you diagnose anything. A half-seated flip-lock connector mimics a dead switch exactly, and this is the first move on three of the faults on this page. If reseating does not fix it, ohm the dome closed at the ribbon end.

One detail that saves a wasted continuity test: seven of the eight contacts on each shoulder connector are ground. A ribbon with a cracked trace will almost always still pass continuity on something. Ring out the one signal contact specifically, which is pin 4 on the right board and pin 3 on the left. Both cards are in the component data below.

When you order, order per board. The New-series flex is a different part from the Old-series flex.

No charge, dead charge LED, or a port that wobbles

Either no orange LED at all when you plug in, or the port wiggles and the console only charges at one cable angle.

On CTR-001 and SPR-001 this is usually the port itself. GBATEK’s own note on the CTR part says it “breaks off easily, but it can be soldered back in without problems,” and it is fitted with glue on a surface-mount footprint. Pads lift, or a pin goes cold. The 2DS uses a sturdier through-hole connector without the goldplate contacts, so an FTR-001 with no charge is more likely a fuse or the charger IC than a torn port.

The measurement that separates them: TP4 on an Old board is the charger input node and reads 0 V unplugged, around 4.6 V with a good charger. No voltage there with a known-good supply means the port, its solder, or the fuse upstream of it. Wiggle-test under the meter for the intermittent case.

If the port is sound and there is still no charge, go to the dead-console ladder below.

Top screen dead, flickering, or dropping lines

Top screen black, flickering, showing line dropouts, or cutting in and out as the lid angle changes. Bottom screen and everything else normal.

The top screen’s ribbon runs through the hinge and is flexed on every open and close until the fold cracks. It is the highest-prior-probability cause by a wide margin, and the lid-angle test is close to diagnostic on its own: if opening and closing the lid changes the behaviour, it is the flex.

Everything after that is where people go wrong, so here is the part that matters. Healthy connector rails and a released reset do not exonerate the mainboard. Four separate reasons:

The reset line at TP244 is shared with the bottom screen. If the bottom screen images at all, reset tells you nothing top-specific.
The backlight is not on the video connector at all. It is on P7, with on-board PWM and boost.
VCOM is written by the MCU over I2C, so it can be present and still be wrong.
Both the connector’s ZIF socket and the SoC’s lane drivers are board-side.

So fork on image-dead versus dark-dead first, which means: is the top backlight lit? Then inspect the socket and its pads under magnification. Then swap in a known-good top assembly. If the fault follows the board, it is the board.

One more thing about this connector that shapes the whole diagnosis: the top panel is fed a low-swing differential serial link, roughly 200 mV on data. Top screen video cannot be eyeballed with a meter the way the parallel-RGB bottom screen can. The label “MIPI DSI” that circulates for it traces back to a single author, so I have tagged it accordingly rather than repeating it as fact.

The 2DS has no hinge, so this specific failure mode does not apply to it.

Won’t hold charge, dies on unplug, or a bulging back cover

Spent or swollen cell. Measure the pack unloaded; a cell resting well under about 3.5 V, or any visible swelling, means replace. See the safety section above for what swelling does to the rest of the console.

The battery codes are where the model names lie to you. Nintendo assigned exactly three across the whole family. There is no FTR-003, no RED-003 and no JAN-003, and the reason that error is so plausible is that the -004 stylus codes do follow the model prefix.

Pack	OEM rating	Fits
CTR-003	3.7 V, 5 Wh (1300 mAh)	CTR-001 3DS, FTR-001 2DS, JAN-001 New 2DS XL
SPR-003	3.7 V, 6.5 Wh (1750 mAh)	SPR-001 3DS XL, RED-001 New 3DS XL
KTR-003	3.7 V, 5.2 Wh (1400 mAh)	KTR-001 New 3DS, only

Three traps in that table. The New 3DS XL reuses the Old XL pack, which is counterintuitive and is stated both by Nintendo’s shared manual page and independently by iFixit’s compatibility list. The New 2DS XL takes the small pack despite the XL body, and Nintendo prints no part code for it at all while declaring it non-user-replaceable. And KTR-003 is the scarce one: no OEM supply, third-party only, physically unique to the New 3DS small shell. Price that scarcity into a New-3DS-non-XL refurb before you buy the console.

A rounding detail that stops a false “degraded” call: CTR-003’s “5 Wh” is rounded up. 1300 mAh at 3.7 V is 4.81 Wh, so a pack metering about 4.8 Wh is at spec.

Ignore listings claiming 5800 mAh. That is roughly four times the OEM energy in the same volume.

Cracked hinge or a top shell split at the pivot

The plastic hinge is moulded into the top case. That makes it a structural weak point, it is notorious on the 3DS XL, and it has a procurement consequence people get wrong constantly: this is a partial shell swap, not a hinge part. You need the midframe, the hinge and the LED diffuser, and not necessarily a full shell.

Transfer the top-screen ribbon carefully. A hinge job is the classic way to introduce the top-screen fault above.

The 2DS slate is exempt.

Totally dead, no LED, with a good battery

Now you are in the electrical part of the platform, and the topology hands you the triage split for free.

There are two board fuses on an Old board. F1 is the external-power fuse, F2 is the internal-battery fuse. So:

No charge LED at all with a known-good supply points at F1 and the input path.
A charge LED that lights and then dies under battery points at F2 and the battery path.

Buzz each for continuity. A blown chip fuse reads open and usually looks untouched.

On RED-001 the battery fuse moves to the button board and the mainboard keeps only the charger fuse. That placement is disputed: one widely read repair writeup describes checking two fuses with the second one on the mainboard. Buzz every fuse you can find on both boards rather than assuming a location. JAN-001 keeps everything on the mainboard.

If the fuses are good, the charger IC is next, and it is one of the very few board ICs on this platform you can genuinely order. It is a TI BQ24072, marked CKP, and the rail order it gives you is a clean three-probe test:

IN, then OUT, then BAT: TP4, then TP8, then TP259 on an Old board. Those three probes separate a dead jack or fuse from a dead charger from a dead battery path, in that order. Full pin table and the strap-resistor values are on the BQ24072 card below.

If the fuse is good and TP8 is dead, the candidates are the charger IC and the PMIC, a Nintendo-custom TI part with no datasheet and no published pinout. Ordering or reflowing the wrong one of those two wastes the repair, so localise by rail rather than by guess.

One more cause worth clearing early on a New 3DS XL, because it is cheap to check and expensive to miss: a shorted lower-LCD flex. Unplug it before you condemn anything.

A New unit is dead but the port and battery are fine

This is where the button board becomes a candidate, and only a candidate.

On RED-001 a cracked joint or a dead IC on the button board will kill power-on with everything else intact. But the same source everyone cites for that fix checks the mainboard fuses first, and the charger, the PMIC, the external fuse, the MCU and the CPU are all still on the RED mainboard. Clear the ladder above before condemning a board.

And per model, not per generation: RED-001 has the full button board, KTR-001 has a power and ABXY board whose chip complement nobody has documented, and JAN-001 has no button board at all. Treat that one as a mainboard job.

SD or microSD not detected

Rule out a failing card or a bad format before you open the console.

Then confirm SD VDD is 3.3 V and that the card-detect line changes state on insertion, before you suspect the data lines or the reader.

Then reseat, because on CTR-001, KTR-001 and RED-001 the reader is a modular board on a ribbon and reseating is free. Only SPR-001 and JAN-001 need desoldering.

One CTR-001 wrinkle: the SD daughterboard is electrically modular but mechanically bonded to the eMMC shielding plate. GBATEK tells you to desolder the plate first; iFixit’s guide simply pries the adhesive pad.

Game card not read

Inspect and clean the slot for bent pins with a wood or plastic pick before you desolder anything.

Card-not-detected-at-all, as opposed to a read error, often traces to the insert-detect switch rather than the bus. Verify the slot supply is 3.3 V at pin 8 first.

Note that the card protocol is encrypted, so probing buys you a supply, a clock and lines toggling. It does not buy you the payload. And on the 3DS the card supply is software-gated, unlike the original DS and DS Lite where it is live whenever a card is detected, so 0 V there with a card inserted is not automatically a fault.

On RED-001 the cart slot is on its own board, which isolates that repair from the mainboard.

Dead speakers, headphones fine

The headphone jack has an insert-detect leaf that mutes the speakers. A leaf stuck in the “inserted” position tells the console a plug is always present. Check the detect contact with nothing plugged in; clean or replace the jack.

Two things that are not faults. The 2DS is mono, one speaker, so one channel is correct behaviour. And if speakers and headphones are both dead, this is the wrong diagnosis: that is the amplifier path or the audio rail.

A measurement trap on this connector, if you go probing: pin 1 is a driven mid-rail reference, not chassis ground. It swings around 2.2 V peak-to-peak. Measure left and right against that pin, not against board ground, or you will read a phantom offset on a healthy amplifier.

Blue light comes on, then goes off

The fork here is how the LED dies, not what you did to the console last.

Fades out slowly: almost always software. Specifically, it is a missing or invalid boot.firm, or an unreadable SD card, on a console that has custom firmware installed. The modified bootloader shuts the system down gracefully when it cannot find the file it needs. Fix it on the SD card: reseat, clean, reformat, and put a fresh copy of the file back.
Cuts off instantly, especially with an audible pop: usually hardware, and the list is much wider than a misseated ribbon. Open or shorted backlights, shorted or open LCD high-voltage rails, a missing or badly seated LCD, power rails shorted together, data lines shorted to a rail through a damaged camera flex, or an open speaker line.

Then ask whether the lid angle changes the behaviour. If it does, you are back at the top-screen flex.

A misseated or damaged top-screen ribbon is a hardware fault, so it belongs in the second branch, not the first. That distinction matters because it is exactly the case people misfile as “software” after a screen job.

Exhaust the SD-level fixes before touching the console’s internal storage.

Top screen washes out at an angle

Not a fault. That is a TN panel behaving normally.

This one is a grading call rather than a repair, and it is worth real money. New 3DS (KTR-001) and New 3DS XL (RED-001) shipped with either IPS or TN panels at random, top, bottom, both or neither. Every Old model is TN, and the New 2DS XL is effectively guaranteed TN.

The tilt test works on every model with no software: view the lit screen at a steep angle. Image stays visible means IPS; it washing out into the backlight means TN.

There is a software check, but scope it correctly: it reads panel type only on New-family hardware running custom firmware, because the underlying call is stubbed on the Old 3DS. On CTR, SPR and FTR the tilt test is all you have.

Two consequences for parts. Do not mix a TN and an IPS panel in one unit expecting matched colour: iFixit’s own copy for the IPS part warns it has a wider viewing angle but slightly lower luminance, so the swapped unit can look dimmer without anything being wrong. And on any swap, match the model’s screen FFC pin count: 41 and 53 pin on the 3DS XL and New 3DS XL against 37 and 41 pin on the 3DS.

Yellowing of a panel from UV and heat is a separate, permanent grading defect and is not repairable.

3D gone or ghosting with a fine 2D image

The parallax barrier is its own driven layer, separate from the main panel, and the depth slider is its own part. A clean 2D image with broken 3D isolates the fault to one of those two rather than to the LCD.

Only the four 3D models can have this: CTR, SPR, KTR and RED. The 2DS and New 2DS XL have no parallax layer and no slider.

Screen dark but faintly imaged

Backlight or its boost circuit, not the video path.

The discriminator is worth stating plainly because it sends you to two completely different repairs. Lit but black is a video or logic fault. Dark but faintly imaged is a backlight fault. On the lower screen, the backlight connector’s plus line is pulsed at roughly 14 V; absence of that pulse points at the boost circuit or a broken backlight ribbon.

Cross-cutting bench notes

The case screws are JIS, not Phillips. A true Phillips driver cams out and strips them. Use a JIS #00 or a JIS-compatible bit. A stripped case screw is the most avoidable self-inflicted failure on this platform.
Reseat before you desolder. Half-seated flip-lock ribbons mimic dead components on the shoulder buttons, the top screen and the SD reader.
The New 3DS non-XL has essentially no repair-guide coverage. For that board, the closest documented analogue is the New 3DS XL layout.

Inside the family: what differs

The 2DS is one LCD, not two. Both “screens” are windows onto a single panel. That is two-sourced, and it has a hard repair consequence: a crack in either window is one part, and no 3DS or 3DS XL panel is a donor for it.

“400x240” means two different things on the two 2D models. On the 2DS it is genuine 400x240 square-pixel geometry, because there is no parallax barrier over the upper window. On the New 2DS XL the panel is physically 800x240 (the same glass family as the New 3DS XL upper panel) driven with adjacent column pairs ganged so the addressable dot count is 400x240. The bench tell is that a New 2DS XL shows the same 2D-mode pixel-pairing artefact as the 3D models, which a true 400x240 panel would not. Do not read “400x240” off a manual and conclude the two panels are the same class of part.

Region does not change the mainboard silicon. It is firmware, account territory and a cart region check. On the Old models the region is printed in the model number itself, as in CTR-001(EUR); the New-series manuals print the bare code, so read the case label on those. Taiwan and China were never released for any New-series model.

The charger IC is not confirmed across all six boards. One wiki’s table spans the whole family, but that spanning cell was a formatting change by one editor who left question marks on neighbouring rows for the same boards. Direct evidence exists for three boards. Nothing refutes it on the other three. Read the marking rather than the table.

The battery thermistor never reaches the charger. The charger’s thermistor pin is tied to a fixed resistor rather than to the pack. That is a single-source claim, but it is the explanation for why two-wire third-party packs charge at all.

Mods worth knowing

I am not reproducing anyone’s install guide. This is an orientation to what is worth doing, what does not exist, and where the money is.

Custom firmware

The single biggest resale lever on this platform, and it is software-only with no soldering. It replaces the boot chain so the console runs unsigned code: region-free loading, homebrew, emulators, save and theme management.

The reason it has gone from optional to load-bearing is that a stock 3DS in 2026 has lost most of its online surface. The eShop stopped taking purchases and free downloads in March 2023, and Nintendo Network online play ended in April 2024. What still works on a stock unit is redownloading titles already purchased, StreetPass, and the Pokémon transfer services, which run outside Nintendo Network.

Every retail console in the family, any region, any firmware, can be modified. The only real decision is the entry route, and that is driven by the exact firmware version and whether the unit is Old or New, not by region alone. The current-firmware routes are the ones that matter on a refurb bench, because most units arriving today are fully updated.

Three things to be straight about in a listing. Take the storage backup immediately after the install, not before: the tool that makes it cannot run until the modification is in place, and that ordering catches people out. Sanitise the unit of the previous owner’s accounts and data before it ships. And do not advertise “eShop restored”: the community network replacement covers online play, and its own project has stated it will not add an eShop revival.

The IPS screen field

This is the most misunderstood hardware topic on the platform, so get the vendor landscape right before quoting anyone.

There is no Game-Boy-style IPS kit for any 3DS model. Not a panel plus driver board plus ribbon adapter plus cut lens. No vendor sells one. The kit vendors people name in this context sell laminated IPS kits for the Game Gear, the DS Lite and Game Boy-class hardware, and their 3DS catalogue is batteries, buttons, hinges, membranes, touchscreens and lenses.

There is exactly one catalogued aftermarket IPS upper panel, and it fits the New 3DS XL only. New 3DS non-XL has no catalogued IPS panel at all and is generally the scarcer parts situation of the two.

The third route is the one the grading market actually pays for, and it is free: the factory lottery on New 3DS and New 3DS XL. Tilt-test every one of those you buy, because you are grading the unit rather than modifying it.

Electrically an IPS swap is a non-event, since there is no panel-type check on the mainboard. The problems are calibration and mechanics. A dim top screen after a correct IPS install is usually the panel’s own lower luminance, which its vendor documents, plus a probable calibration mismatch: the console stores per-unit factory display calibration, including flicker and contrast voltages, and it is tuned for the panel that left the factory. Treat “looks wrong after the swap” as expected work, not as a bad panel.

USB-C charge port

A popular quality-of-life and resale upgrade, and the correct expectation to set is important: every one of these feeds the console’s existing charge circuit. It is a port or breakout swap, not a power-delivery sink. The battery still charges through the console’s own charger IC and F1.

Open-source boards exist per model, and the compatibility statements are mechanical and need reading literally. One well-known board covers the 3DS XL only and its own README excludes the 2DS, New 2DS XL, New 3DS, New 3DS XL and the original 3DS by name, pointing at a different project for the original 3DS. Its one-line repository description is looser than its README, and reading the description instead of the README is how people order the wrong board. On the New 3DS XL there is also an additive route that adds a port without removing the OEM one, which is the lowest-risk version of this job.

On charge rate, say this before a customer asks. The OEM supply gives 900 mA at 4.6 V, and a bare USB-C sink is entitled to 500 mA at 5 V by specification, which sounds like a downgrade. In practice it is not, and the reason is on the board: the charger’s input current limit is set by a strap resistor to about 0.82 A, so the console is the binding constraint, not the supply. A measured report of 800 mA at 5.1 V lands dead centre of that computed range.

One buying detail with a real failure mode behind it: the cheap boards that omit the two 5.1 kΩ configuration resistors will not charge from a C-to-C cable at all, only from a USB-A-to-C one. And the characteristic failure of every charge-port mod is F1, not the new port.

Cover plates, shells and hinges

The New 3DS non-XL is the only model in the family with swappable faceplates, front and back. Fast, no soldering, and a genuine listing differentiator.

Aftermarket shells exist across models. A full shell is a complete teardown, so pair it with other internal work rather than doing it alone. And the hinge, as above, is a midframe job rather than a hinge part.

What actually changes value

Lead with the model code and the panel grade, because those are what a buyer is paying for: the mainboard code, whether the panels are IPS or TN on a New unit, whether the console has custom firmware, and the battery’s real state. Be specific about which modifications are reversible. And do not describe a sixty-second software job as a “mod” at a premium if the buyer’s real question is whether the shoulder buttons work.

Recap and parts

The headline: do not blanket-recap a 3DS.

The two boards with published component lists are BGA SoC, FCRAM and eMMC, SMD ceramic decoupling, and a handful of small-outline ICs. There is no bulk aluminium electrolytic in the Game Gear or Genesis sense on any documented 3DS board, there is no published cap map for any of the six, and the community failure record is mechanical, connector and fuse rather than dried electrolyte. The community wikis have no 3DS recap page because there is nothing to put on one.

If a rail is dead, chase the fuse, the charger IC, the PMIC or a shorted flex. The only capacitor work that pays here is replacing one you cooked yourself during a hot-air job.

That makes the consumables list short and unusual for this site: fuses, cells, ports, flexes and plastic.

The fuses are the one place I have to leave you with an open question rather than a number. GBATEK read a “K” off the fuse body. Cross-referenced to a manufacturer part-marking table, K is 1.50 A, H would be 1.00 A and J would be 1.25 A, and that marking table is identical across every revision of the datasheet I checked. What is inferred is the identification of the fuse as that manufacturer’s part in the first place, and other manufacturers use different letters. The forum record does not settle it either: the same thread set carries both a 1 A figure and a 1.5 A figure, and the 1 A one is a claim about an acceptable substitute rather than about the original rating.

So: read the marking letter under magnification and match it against the manufacturer’s own table rather than trusting a number off a forum. Do not fit a 1 A part into a 1.5 A position: it will nuisance-blow on inrush and you will spend an afternoon chasing a short that is not there.

Everything else is covered in the tables below: the three battery codes and what they actually fit, the charge port per model, the shoulder flex, the top screen assembly and its FFC pin counts, the SD reader modules, and the hinge midframe.

Replacement parts

The consumables a 3DS job actually eats, and the one where the answer is still “read the marking.”

Non-cap consumables

Function	OEM part	Why replaced	Substitute	Note
Board fuses F1 (charge input) and F2 (battery)	0603 chip fuse, body marked 'K'	The number one board-level consumable. F1 opens on a charge-path fault, F2 on a battery-path fault	Littelfuse 046701.5NR (1.5 A, marking K) if the 467-series identification holds	read the marking letter under magnification and match it against the manufacturer's table rather than trusting a forum number ^{single source}^{gbatek,littelfuse-467}
notes The value is genuinely unresolved and it matters. GBATEK read a “K” off the part body. Cross-referenced to Littelfuse’s 467-series part-marking table, K is 1.50 A; 1.00 A would be marked H and 1.25 A would be J. That marking table is identical across every revision of the datasheet checked, so the table itself is solid. The inference is the identification of the fuse as a Littelfuse 467-class part in the first place, and GBATEK never names a manufacturer. Bel, Schurter and AEM use different letters. The forum record is not a clean disagreement either: the same thread set carries both “0603 sized 32 volt 1 amp fuse will work fine as a replacement” and “the fuses are 1.5 A in a 0603 package”. The first is a claim about an acceptable substitute, not about the original rating. Practical guidance until it settles. Do not fit a 1 A part in a 1.5 A position: it will nuisance-blow on inrush and you will chase a phantom short. A healthy 1.5 A 467-series part ohms out around 0.039 Ω on the current datasheet revision, about 6.5 % below the older printed figure that still circulates. Never jumper a fuse except as a momentary diagnostic. Both of these blow because of something downstream: the PMIC, the charger IC, a shorted lower-LCD flex, or a reverse-inserted battery. Find the short first. On `RED-CPU-01` the battery fuse moves to the button board and the mainboard keeps only the charger fuse. That placement is disputed (one widely read repair writeup describes two fuses with the second on the mainboard), so buzz every fuse you can find on both boards. `JAN-CPU-01` keeps everything on the mainboard.
Battery pack: 3DS, 2DS, New 2DS XL	Nintendo CTR-003	Swell and capacity loss. Swelling bows the shell, lifts the board and cracks hinges from the inside	1300 mAh 3.7 V, protection PCM required	Nintendo rates it 5 Wh, which is rounded up: 1300 mAh at 3.7 V is 4.81 Wh, so a pack metering ~4.8 Wh is at spec, not degraded ^✓^{nintendo-manuals,ifixit}
notes Match the code, not the console name. Nintendo assigned exactly three battery codes across the whole family, and the mapping is not what the model names suggest. There is no FTR-003, no RED-003 and no JAN-003. The `<MODEL>-004` pattern that does hold is for the stylus, which is probably where the mistake comes from. CTR-003 fits the CTR-001 3DS, the FTR-001 2DS and the JAN-001 New 2DS XL, the last of those despite being an XL body. Nintendo’s New 2DS XL manual gives type and capacity but no part code and declares the pack non-user-replaceable; vendors ship a CTR-003-class 1300 mAh pack for it and the capacities agree at 5 Wh. Ignore “5800 mAh” marketplace listings. That is roughly four times the OEM energy in the same volume.
Battery pack: 3DS XL, New 3DS XL	Nintendo SPR-003	Same swell and capacity loss	1750 mAh 3.7 V, 6.5 Wh	the New 3DS XL reuses the Old XL pack. This is stated by Nintendo's shared manual page and independently by iFixit's compatibility list ^✓^{nintendo-manuals,ifixit}
Battery pack: New 3DS (non-XL) only	Nintendo KTR-003	Same swell and capacity loss	1400 mAh 3.7 V, 5.2 Wh	the scarce one. No OEM supply, third-party only, and physically unique to the New 3DS small shell: price the scarcity into that model's margins ^✓^{nintendo-manuals,ifixit}
notes Vendor capacity claims disagree across listings, from 1200 to 1400 mAh. Nintendo’s 5.2 Wh is the figure to trust, and 5.2 Wh divided by 3.7 V is about 1400 mAh.
DC charge port	P2 assembly, per model	Tears off or goes cold. The CTR-001 part is surface-mount, glued, and GBATEK's own note says it breaks off easily	New-series port (KTR/RED/JAN) is a catalogued repair part; Old-series ports come from handheld parts vendors	the 2DS uses a sturdier through-hole connector, so an FTR no-charge is more likely a fuse or the charger IC than a torn port ^✓^{gbatek,ifixit}
Shoulder button flex (L / R)	per model, Old and New parts differ	The metal dome cracks or the ribbon fatigues at the fold. This is the most common fault on the platform	aftermarket flex assemblies, sold per side	reseat the ribbon in its mainboard socket before ordering anything: a half-seated flip connector mimics a dead switch exactly ^✓^ifixit
notes The New-series flex is a different part from the Old-series flex. Order per board, not per family. The button itself is retained by a metal pin and a small spring, and on reassembly the contact and ribbon can lift out with the black cover. Note the orientation before it comes apart.
Top screen assembly / hinge ribbon	per model; the FFC pin count differs	The ribbon runs through the hinge and is flexed on every open and close until the fold cracks	full upper assemblies are sold for CTR, SPR, RED and JAN	match the model's FFC pin count: 41 and 53 pin on the 3DS XL and New 3DS XL against 37 and 41 pin on the 3DS ^{single source}^ifixit
notes A hinge repair is the classic way to introduce this fault, because the ribbon has to be transferred. There is one catalogued factory-fit IPS upper panel in the whole family and it is for the New 3DS XL. Its own vendor copy warns that it has a wider viewing angle than the TN part but slightly lower luminance, so a swapped unit can look dimmer without anything being wrong.
SD / microSD reader module	'SD 02 DM 2-2' on CTR-001; board-plus-flex modules on KTR and RED	Contact wear, bent pins, or a reader that has un-seated	pre-mounted replacement modules are sold for CTR, KTR and RED	modular on CTR, KTR and RED; soldered on SPR and JAN. Get this backwards and you desolder a reader that only needed reseating ^✓^{gbatek,ifixit}
notes On CTR-001 the board is electrically modular but mechanically bonded: GBATEK says to desolder the eMMC shielding plate before removing it, while iFixit’s guide simply pries the adhesive pad.
Hinge / midframe assembly	midframe + hinge + LED diffuser, per model	The plastic hinge is moulded into the top case, so it cracks with drops and age	aftermarket midframe kits	this is a partial shell swap, not a hinge part. Notoriously common on the 3DS XL; the 2DS slate has no hinge and is exempt ^✓^{ifixit,consolemods}

Chip and connector pinouts

The component library: each connector and IC defined once, with a pin table and, where the package geometry supports it, an interactive diagram. The badge on each card is the mainboard code the table was drawn from.

Read those badges carefully on this device in particular. There is no OEM schematic for any 3DS board, most of what follows came off one board, and several cards are deliberately empty with a note saying what would fill them.

Components & pinouts

Each part is defined once. The badge on every card is the mainboard code the table was actually drawn from, not the model name on the box. Almost everything numbered below came off one board, `C/CTR-CPU-01` (the original 3DS): it is the only 3DS-family mainboard with a complete published connector map and a silkscreened test-point field. Do not carry a designator or a TP number onto another board without buzzing it.

Every diagram rotates and flips to match the board in front of you, zooms from a whole-package overview up to 3.2×, and saves at the orientation you are looking at — ↓ SVG matches the screen, and ↓ SVG (print) gives you the same drawing as ink on white. Both are vector, so print them at any size you like, which is the better route for the long connector pinouts since printing this page has to shrink them to fit the sheet. The # beside a part name links to that card, and the one in the pin panel links to a single pin — either link opens the card for whoever you send it to.

Battery charger / power path TI BQ24072, marked CKP (U7 on the Old mainboard) C/CTR-CPU-01 · RED-CPU-01 qfp-16^✓^{ti-bq24072,gbatek,3dbrew}#

VQFN-16 (RGT), 3.00 x 3.00 mm. The pin assignment is datasheet-confirmed; the '3DS wiring' notes on individual pins are GBATEK-only

The pin numbers, the package, the CKP top marking and the 4.35-6.6 V input range all match between TI’s datasheet and GBATEK’s board-wired table. What no second author has traced is the wiring column: the TP numbers, both strap resistor values, the EN1/EN2 strapping and the TS pin tied to a fixed resistor instead of a pack thermistor.

“BQ24072 on all six boards” does not hold up per board. Direct evidence exists for three: C/CTR-CPU-01 and RED-CPU-01 are two-sourced, and SPR-CPU-01 has an independent bench read behind it. For FTR-CPU-01, KTR-CPU-01 and JAN-CPU-01 the only support is one wiki’s spanning table cell, and GBATEK affirmatively disclaims those boards because no PCB photos exist. Nothing refutes the part there. Read the marking.

Three bench consequences fall straight out of the table.

The rail order for a no-charge board is IN (13) then OUT (10/11) then BAT (2/3), and each has its own test point: TP4, then TP8, then TP259. Three probes separate a dead jack or fuse from a dead charger from a dead battery path.

EN2 high and EN1 grounded means the input current is set by the ILIM resistor. The 1.96 kΩ strap gives roughly 0.82 A, which is why a plain USB-C source charges at about OEM speed: the console is the binding constraint, not the supply. Lifting EN1 or shorting EN2 silently drops the console into 100 mA mode.

The exposed thermal pad is bonded to VSS internally, but TI is explicit that it must not be used as the primary ground input. A reflow that wets the pad and misses pin 8 behaves erratically rather than failing cleanly.

0° 1×

Click a pin to see its signal, net, and sources.

Interactive diagram needs JavaScript. The full pinout is in the table below.

Pin	Signal	Cat	Net	Note
1	TS	signal	ts	thermistor input. On the 3DS it is a dummy 10 kΩ to ground: the pack thermistor is not used ^✓
2	BAT	rail	vbat	charger output and battery sense. TP259 to F2 to TP22 bridge-ok ^✓
3	BAT	rail	vbat	same net as pin 2 bridge-ok ^✓
4	/CE	signal	ce	charge enable, active low; internal ~285 kΩ pull-down. TP11 ^✓
5	EN2	signal	en2	input-current-limit select. 10 kΩ to TP8, i.e. high ^✓
6	EN1	signal	gnd	input-current-limit select, tied to ground. With EN2 high this selects resistor-programmed mode ^✓
7	/PGOOD	signal	pgood	power good, open drain, low when the input is valid. TP7 ^✓
8	VSS	gnd	gnd	ground, and the thermal-pad potential. Wet this pin, not just the pad ^✓
9	/CHG	signal	chg	charging status, open drain, low while charging. TP10 ^✓
10	OUT	rail	vsys	system supply output. TP8 bridge-ok ^✓
11	OUT	rail	vsys	same net as pin 10 bridge-ok ^✓
12	ILIM	signal	ilim	input current limit, 1100 Ω to 8 kΩ to VSS. 1.96 kΩ on the 3DS, giving roughly 0.82 A ^✓
13	IN	rail	vin	external supply, operating range 4.35-6.6 V; survives 26 V by suspending. Fed from EM1, F1 and TP4 ^✓
14	TMR	signal	tmr	safety-timer program, 18-72 kΩ to VSS. 56 kΩ on the 3DS ^✓
15	TD	signal	gnd	termination disable, checked at startup only. Tied to ground, so termination is enabled ^✓
16	ISET	signal	iset	fast-charge current program, 590 Ω to 8.9 kΩ to VSS. 1.5 kΩ on the 3DS. The voltage here mirrors actual charge current, so it is a measurement point ^✓

eMMC (NAND) access pads TP249-TP253 plus CLK (Old boards) C/CTR-CPU-01 · SPR-CPU-01 · FTR-CPU-01^✓^{gbatek,3dbrew}#

the highest-traffic pin work on the platform, and the only pin-level data that exists for the New boards at all. DAT0 + CMD + CLK + GND is a working 1-bit hardmod

The data lines are two-sourced. The clock is where boards diverge, and it is where a wiring mistake destroys the console.

On C/CTR-CPU-01 there is no CLK test point. That negative is two-sourced. The commonly circulated tap is a small unlabelled via beside C114/C113/C106, and that specific location rests on one 3dbrew photograph (the ElOtroLado diagram is a re-crop of the same image, not an independent observation). A second documented tap that is easier to solder: the 1x1 120 Ω series pass-through resistor in the eMMC CLK line. Either end is the same net and both are larger targets than a bare via.

Do not solder CLK to TP254. 3dbrew’s own table carries a speculative row “TP254? || NAND-CLK?” with question marks on both the number and the net and a blank voltage column. GBATEK instead identifies TP254 as the gyroscope’s pin 8 supply, on cut-link CL56, and TP254 is silkscreened on side A, next to the circle-pad test points, while the eMMC cluster is on side B. The numeric run does not continue; it jumps to the other face of the board. A hardmod wired to a gyro supply pin is a destroyed board.

Board revisions differ. A later C/CTR-CPU-01 revision has been found carrying a sixth pad in the TP249-TP253 cluster, which is the CLK pad. Look at the cluster in front of you before assuming the photograph’s tap point. If there are six pads, use the sixth.

SPR-CPU-01 and FTR-CPU-01 both have a real CLK test point, TP260, which does not exist on the 3DS. That is a genuine board change, not a transcription artefact.

On KTR-CPU-01 and RED-CPU-01 there are no test-point numbers at all: bare vias and package pads clustered on the eMMC’s board edge, CLK isolated above the package. That is image-only pin location and cannot be reduced to a table. Work from a photograph and buzz DAT0 to the corresponding eMMC ball before applying power. Identify the eMMC by its marking rather than by the model: both Samsung and Toshiba parts shipped, in different packages with different pad layouts.

No eMMC access-point documentation exists for JAN-CPU-01 as of 2026-08-05.

Pin	Signal	Cat	Net	Note
TP249	eMMC DAT0	bus	emmc_dat0	required for a 1-bit hardmod ^✓
TP250	eMMC DAT1	bus	emmc_dat1	optional; buys transfer speed only ^✓
TP251	eMMC DAT2	bus	emmc_dat2	optional ^✓
TP252	eMMC DAT3	bus	emmc_dat3	optional ^✓
TP253	eMMC CMD	bus	emmc_cmd	required ^✓
TP260	eMMC CLK	bus	emmc_clk	exists on SPR-CPU-01 and FTR-CPU-01 ONLY. There is no CLK test point on C/CTR-CPU-01 ^✓
via	eMMC CLK (CTR)	bus	emmc_clk	unlabelled via beside C114/C113/C106 on C/CTR-CPU-01, or either end of the 120 Ω series resistor in the CLK line ^{single source}
GND	Ground	gnd	gnd	any board ground; 3dbrew references the square pad at the bottom-left corner ^✓

New-series board connectors KTR-CPU-01 / RED-CPU-01 / JAN-CPU-01 KTR-CPU-01 · RED-CPU-01 · JAN-CPU-01^{single source}^{gbatek,3dbrew}#

no numbered connector map exists for any New board. This card is here so the gap is explicit rather than silently absent

GBATEK’s connector page covers C/CTR-CPU-01 and nothing else, and its component list marks the 3DS XL, 2DS, New 3DS and New 2DS XL as unknown. For RED-CPU-01 there is a chip list but no connector table.

The New boards also have no test-point text layer. GBATEK states they have testpoints “but only half as many, and without text layer, so they could be used/described only via photos or drawings”, and the KTR, RED and JAN board photographs show no TP silkscreen. That is a physical property of those boards, not a research shortfall. Do not go looking for a TP number on a New board.

Two structural facts that are established, and that change the job:

RED-CPU-01’s external power connector is 2-pin, on side A near the charger. A CTR-shaped mental model of four grounds around a centre pin does not apply.

RED-CPU-01 carries a separate button board (“F-7, PWB, /ABXY, /AU-C, /RED-01”) holding the ABXY, C-stick, ZL/ZR and power functions, a fuel gauge, a magnet sensor and the battery connector. KTR-CPU-01 has a power and ABXY board too, silkscreened “ABXY-01”, but which ICs it carries is undocumented. JAN-CPU-01 has no button board at all: its button pads and its power switch are on the mainboard.

No pins captured yet — bench stub.

SD-card daughterboard connector P14, 20-pin, board 'SD 02 DM 2-2' C/CTR-CPU-01 header-20^{single source}^gbatek#

GBATEK is the only author for this numbering. Odd pins are all ground except 17 and 19, which are VDD

On CTR-001 the SD slot is a separate daughterboard on this 20-pin cable, so it swaps as a unit. GBATEK adds the part that catches people out: the board “is firmly glued to eMMC shielding plate”, and it tells you to desolder the eMMC plate first. iFixit’s guide instead just pries the adhesive pad. Electrically modular, mechanically bonded.

Construction differs per board and it is not an Old-versus-New split. C/CTR-CPU-01, KTR-CPU-01 and RED-CPU-01 all use a modular reader that unplugs. SPR-CPU-01 and JAN-CPU-01 are soldered. Getting this backwards sends you desoldering a reader that only needed reseating.

The card end is the ordinary SD bus, on 13 solder pads plus shield: pin 1 DAT3, 2 CMD, 3 VSS, 4 VDD, 5 CLK, 6 VSS, 7 DAT0, 8 DAT1, 9 DAT2. microSD on the New boards is the same bus on 8 contacts.

Before suspecting the CMD and CLK lines, confirm VDD is 3.3 V at TP13 and that /Inserted at TP94 changes state when a card goes in.

0° 1×

Click a pin to see its signal, net, and sources.

Interactive diagram needs JavaScript. The full pinout is in the table below.

Pin	Signal	Cat	Net	Note
1	GND	gnd	gnd	^{single source}
2	Data1	bus	sd_dat1	TP100 ^{single source}
3	GND	gnd	gnd	^{single source}
4	Data0	bus	sd_dat0	TP101 ^{single source}
5	GND	gnd	gnd	^{single source}
6	GND	gnd	gnd	bridge-ok ^{single source}
7	GND	gnd	gnd	^{single source}
8	/Unlock	signal	sd_unlock	write-protect slider. TP93, cut-link CL50 ^{single source}
9	GND	gnd	gnd	^{single source}
10	/Inserted	signal	sd_insert	card-detect. TP94, cut-link CL49 ^{single source}
11	GND	gnd	gnd	^{single source}
12	GND	gnd	gnd	bridge-ok ^{single source}
13	GND	gnd	gnd	^{single source}
14	Cmd	bus	sd_cmd	via a 57 Ω series resistor. TP96 ^{single source}
15	GND	gnd	gnd	^{single source}
16	Clk	bus	sd_clk	TP103 ^{single source}
17	VDD 3.3 V	rail	v33_sd	TP13 bridge-ok ^{single source}
18	Data3	bus	sd_dat3	TP97 ^{single source}
19	VDD 3.3 V	rail	v33_sd	same net as pin 17. TP13 bridge-ok ^{single source}
20	Data2	bus	sd_dat2	TP98 ^{single source}

Right shoulder button board P15, 8-pin C/CTR-CPU-01 header-8^✓^{gbatek,3dbrew}#

seven of the eight contacts are ground. Ring out pin 4 specifically: a cracked ribbon almost always still passes continuity on something

The shoulder button is a metal dome on a small flex that plugs into a socket on the mainboard, and the fold fatigues where it is actuated. That makes it the single most common fault on the platform, and it is why reseating the ribbon is the first move rather than the last.

3dbrew’s measured survey corroborates the function: TP82 reads 1.8 V open-collector in the main-buttons domain.

The New-series flex (KTR, RED, JAN) is a different part from the Old-series one. Order per board.

0° 1×

Click a pin to see its signal, net, and sources.

Interactive diagram needs JavaScript. The full pinout is in the table below.

Pin	Signal	Cat	Net	Note
1	GND	gnd	gnd	bridge-ok ^✓
2	GND	gnd	gnd	bridge-ok ^✓
3	GND	gnd	gnd	bridge-ok ^✓
4	Button R	signal	btn_r	the only signal contact. 1.8 V open-collector. TP82 ^✓
5	GND	gnd	gnd	bridge-ok ^✓
6	GND	gnd	gnd	bridge-ok ^✓
7	GND	gnd	gnd	bridge-ok ^✓
8	GND	gnd	gnd	bridge-ok ^✓

Left shoulder button board P16, 8-pin C/CTR-CPU-01 header-8^✓^{gbatek,3dbrew}#

same construction as the right board, but the signal sits on pin 3 rather than pin 4

Note the asymmetry: the left board’s button contact is pin 3, the right board’s is pin 4. Everything else on both is ground. 3dbrew measures TP83 at 1.8 V open-collector, matching TP82 on the right side.

0° 1×

Click a pin to see its signal, net, and sources.

Interactive diagram needs JavaScript. The full pinout is in the table below.

Pin	Signal	Cat	Net	Note
1	GND	gnd	gnd	bridge-ok ^✓
2	GND	gnd	gnd	bridge-ok ^✓
3	Button L	signal	btn_l	the only signal contact. 1.8 V open-collector. TP83 ^✓
4	GND	gnd	gnd	bridge-ok ^✓
5	GND	gnd	gnd	bridge-ok ^✓
6	GND	gnd	gnd	bridge-ok ^✓
7	GND	gnd	gnd	bridge-ok ^✓
8	GND	gnd	gnd	bridge-ok ^✓

Game-card slot P1, 19-pin (17 card + 2 insert switch) C/CTR-CPU-01 header-19^✓^{gbatek,3dbrew,consolemods}#

one physical slot serves DS, DSi and 3DS cards. Card pins 1-17 are 3-sourced; the board-level insert-switch numbering (18, 19) and every TP number are GBATEK-only

Card pins 1-17 are the DS/DSi/3DS Slot-1 interface and agree across three independent tables: GBATEK, 3dbrew’s separately written Gamecards page, and the ConsoleMods NDS:Connector Pinouts Slot-1 table.

Two mechanical facts worth carrying to the bench. The cartridge’s own contacts at positions 1, 8 and 17 (the supply and ground trio) are physically longer than the other fourteen: it is the card’s fingers that are extended, not the console’s spring fingers. And card VCC at pin 8 is software-gated on the 3DS: 3dbrew records it as enabled only when the power-supply bits of CFG9_CARDSTATUS are set, unlike the original DS and DS Lite where it is live whenever a card is detected. So a slot that reads 0 V at pin 8 with a card inserted is not automatically a fault.

The card protocol is encrypted, so probing buys you 3.3 V, a clock, and lines toggling. It does not buy you the payload.

0° 1×

Click a pin to see its signal, net, and sources.

Interactive diagram needs JavaScript. The full pinout is in the table below.

Pin	Signal	Cat	Net	Note
1	GND	gnd	gnd	card pin 1; longer contact on the cartridge ^✓
2	CLK	signal	card_clk	serial clock; DS 6.7/4.2 MHz, 3DS up to 16.75 MHz (67 MHz / 4). TP95 ^✓ The ‘16.6 MHz’ figure in circulation is a rounding of 16.75 MHz, not a separate rate.
3	NC	nc		not connected in the console; 3dbrew suggests it may be a card-programming line ^✓
4	/RomCS	signal	card_rcs	ROM chip select, active low. TP107 ^✓
5	/Reset	signal	card_rst	reset; also switches the ROM out of encrypted mode, active low. TP108 ^✓
6	/SaveCS	signal	card_ecs	save-chip select, active low. TP99 ^✓
7	IRQ	signal	card_irq	card removal / interrupt. TP102 ^✓
8	VCC 3.3 V	rail	v33_card	card supply, software-gated on 3DS. TP16 ^✓
9	D0	bus	card_d0	ROM data bus. TP110 ^✓
10	D1	bus	card_d1	ROM data bus. TP111 ^✓
11	D2	bus	card_d2	ROM data bus. TP112 ^✓
12	D3	bus	card_d3	ROM data bus. TP104 ^✓
13	D4	bus	card_d4	ROM / save data. TP105 ^✓
14	D5	bus	card_d5	ROM / save data. TP113 ^✓
15	D6	bus	card_d6	ROM / save data. TP114 ^✓
16	D7	bus	card_d7	ROM / save data. TP106 ^✓
17	GND	gnd	gnd	card pin 17; longer contact on the cartridge ^✓
18	/Insert	signal	card_insert	slot insert-detect switch, active low. TP109 ^{single source} The board-level pin numbering and the TP109 mapping are GBATEK-only. What is independently corroborated is that a mechanical insert switch exists and is read through `CFG9_CARDSTATUS` bit 0.
19	Insert GND	gnd	gnd	insert-switch return ^{single source}

DC input jack P2, 6-pad footprint, 4.6 V in C/CTR-CPU-01 header-6^✓^{gbatek,silkscreen,ifixit,nintendo-manuals}#

the single most damage-relevant connector on the console. Pad order on the board reads 6 4 2 1 3 5, left to right, with a triangle under pin 1

The physical layout, read off the CTR-CPU-01 side-B silkscreen with the jack at the top edge and its opening facing away from you:

6        4         2 | 1         3        5
goldplate  shield  GND   +4.6V   shield   goldplate

The two large side tabs are the shields, the two small centre pads are the power pair, and the two outer small pads are the sprung goldplate contacts that the CTR-007 charging cradle uses. A triangle is silkscreened directly under pin 1’s pad.

Do not ground pin 5. A widely read iFixit answer guesses pin 5 is “a ground for the connector shell”. It is not: pin 5 is the second 4.6 V input, and tying it to ground shorts the adapter straight through F1. The shields are pins 3 and 4.

GBATEK’s own note on this part is worth repeating: it “breaks off easily, but it can be soldered back in without problems (the 2DS uses a better connector with thru-holes; and without goldplates)”. It is also fitted with glue.

Per-board construction differs and the CTR mental model does not transfer. FTR-CPU-01 is through-hole with no goldplates. RED-CPU-01 drops to a 2-pin external power connector on side A near the charger IC. JAN-CPU-01 has three mounting pins to desolder.

If you are working without the board photograph in front of you, do not trust a remembered pin 1. Probe TP4 (4.6 V with the charger in, 0 V without) against TP1.

0° 1×

Click a pin to see its signal, net, and sources.

Interactive diagram needs JavaScript. The full pinout is in the table below.

Pin	Signal	Cat	Net	Note
1	+4.6 V IN	rail	vin	centre-right pad, marked with the silkscreen triangle. TP4 ^✓
2	GND	gnd	gnd	centre-left pad, the return of the power pair. TP1 ^✓
3	GND (shield)	gnd	gnd	large mounting tab bridge-ok ^✓
4	GND (shield)	gnd	gnd	large mounting tab bridge-ok ^✓
5	+4.6 V IN (goldplate)	rail	vin	sprung contact for the CTR-007 charging cradle. Same net as pin 1, NOT a ground. TP4 ^✓
6	GND (goldplate)	gnd	gnd	sprung contact for the charging cradle ^✓

Headphone jack P3, 5-pin, stereo plus insert switch C/CTR-CPU-01 header-5^✓^{gbatek,3dbrew}#

pin 1 is a driven mid-rail reference, not chassis ground. Measure left and right against TP48, not against board ground

Two sources agree on function and disagree usefully on character. GBATEK names the nets; 3dbrew’s measured TP survey shows the audio ground at pin 1 swinging 2.2 Vpp at roughly 33 kHz. Referencing a scope or meter to board ground instead of TP48 puts a phantom 2.2 V offset on a healthy amplifier.

The insert-detect leaf at pin 4 is what mutes the speakers. It reaches the SoC through Q3 and cut-link CL20 as GPIO 1:0, where 0 means no plug and 1 means connected. A leaf stuck in the “inserted” position is the classic dead-speakers-but-headphones-fine fault.

Unlike the DS and DSi, the 3DS has no external microphone contact in this jack.

0° 1×

Click a pin to see its signal, net, and sources.

Interactive diagram needs JavaScript. The full pinout is in the table below.

Pin	Signal	Cat	Net	Note
1	Audio ground (sleeve)	signal	audio_ref	driven mid-rail reference, 2.2 Vpp at ~33 kHz, NOT chassis ground. TP48 ^✓
2	Left (tip)	signal	audio_l	2.2 Vpp AC referenced to pin 1. TP50 ^✓
3	Right (ring)	signal	audio_r	2.2 Vpp AC referenced to pin 1. TP51 ^✓
4	Insert detect	signal	hp_detect	pulled low when a plug is inserted; mutes the speakers. Open-collector, ~1 V. TP49 ^✓
5	GND (chassis)	gnd	gnd	jack body ^✓

Battery connector P4, 3-pin C/CTR-CPU-01 header-3^✓^{gbatek,silkscreen}#

polarity is silkscreened on the board, so it does not need a meter. GBATEK labels the contacts +, centre and −; the 1-2-3 numbering below is positional in that order

A “+” is printed beside the outer contact at the P4 footprint on C/CTR-CPU-01 side B, with a triangle under the connector body. That makes the polarity two-sourced: GBATEK’s naming plus Nintendo’s own silkscreen.

Two facts here decide third-party-battery questions.

The pack thermistor does not reach the charger. GBATEK’s BQ24072 wiring column gives the TS pin as a dummy 10 kΩ to ground, so the charger sees a fixed resistance rather than the cell’s temperature. There is a separate board-side thermistor read elsewhere (TP241). That is the explanation for why two-wire third-party packs charge at all. Single-source, GBATEK.

The centre pin is a detect or alert line sitting near 0.01 V, not a second supply. Do not bridge it to either outer contact.

0° 1×

Click a pin to see its signal, net, and sources.

Interactive diagram needs JavaScript. The full pinout is in the table below.

Pin	Signal	Cat	Net	Note
1	Battery +	rail	vbat	outer contact, silkscreened '+'. Runs to F2, then TP259 and the charger's BAT pins. TP22 ^✓
2	Battery type / detect	signal	bat_detect	centre contact, typically 0.01 V. Not a supply: do not bridge it. TP23 ^{single source}
3	Battery −	gnd	gnd	outer contact, ground return ^✓

Upper LCD video connector P6, 29-pin, PCB side A C/CTR-CPU-01 ffc-29^{single source}^gbatek#

deliberately left empty. The per-pin map exists in one source only, and the top-screen link is low-swing differential serial, so it cannot be eyeballed with a meter the way the lower screen can

The top panel is fed over P6 as a low-swing differential serial link, two data pairs plus a clock pair, around 200 mV on data and 100 mV on clock. GBATEK labels that “Serial MIPI DSI format”, and that label is single-source: 3dbrew has no top-screen pinout and never says MIPI, and the one apparent web corroboration self-attributes back to GBATEK.

The practical consequence is the reason this card exists at all. Healthy rails and a released reset do not exonerate the mainboard on a dead top screen. /RESET at TP244 is shared with the bottom screen, so if the bottom screen images at all it tells you nothing top-specific. The backlight is not on this connector: it is on P7, with on-board PWM and boost. VCOM is written by the MCU over I2C, so it can be present and still wrong. And both the P6 ZIF socket and the SoC’s LCD1 lane drivers are board-side.

Fork on image-dead versus dark-dead first, inspect the socket and its pads under magnification, then swap in a known-good top assembly. If the fault follows the board, it is the board.

Healthy supplies GBATEK measured on a working CTR board: 5.0 V (TP32), 1.8 V (TP5), VCOM 1.7 V (TP24), AVDD 3.43 V (TP35), +10.0 V (TP36) and -5.0 V (TP40). Single-probe readings on one unit, not a spec.

No top-screen connector pinout exists for SPR-CPU-01, KTR-CPU-01, RED-CPU-01 or JAN-CPU-01 as of 2026-08-05. Only the hinge-routing half of this fault generalises across models.

0° 1×

Click a pin to see its signal, net, and sources.

Interactive diagram needs JavaScript. The full pinout is in the table below.

No pins captured yet — bench stub.

Lower LCD video connector P8, 43-pin, parallel 24-bit RGB C/CTR-CPU-01 ffc-43^{single source}^gbatek#

stub. Unlike the top screen this is a parallel RGB bus, so it can be probed conventionally, but the per-pin map is single-source and is not reproduced here

The bottom panel runs parallel 24-bit RGB, which is the reason a lower-screen fault is tractable with a meter and an upper-screen fault is not.

Its backlight is separate, on P12: minus at TP42 near 0 V, plus at TP41 pulsed around 14 V. That split is the whole diagnosis. A lit but black lower screen is a video or logic fault. A dark but faintly imaged screen is a backlight or P12 fault.

A shorted lower-LCD flex is a documented no-power cause on the New 3DS XL. Unplug it before condemning a fuse or an IC.

0° 1×

Click a pin to see its signal, net, and sources.

Interactive diagram needs JavaScript. The full pinout is in the table below.

No pins captured yet — bench stub.

Power management IC ("powerman") TI 93045A4, Nintendo-custom, non-catalogue C/CTR-CPU-01 · RED-CPU-01^✓^{gbatek,ifixit}#

identity is two-sourced but there is no datasheet and no published pinout. Deliberately empty

The part identity holds up beyond GBATEK: parts vendors sell it explicitly as the 3DS / New 3DS / New 3DS XL PMIC, and a TI E2E thread confirms 93045A4 is a Nintendo-custom, non-catalogue TI power-management part.

What does not exist is a datasheet, so there is no pin map to publish. It generates the console’s regulated rails downstream of the charger’s system output. On a board where the fuse is good and TP8 is live but nothing comes up, this and the charger IC are the two candidates.

Ordering or reflowing the wrong one of the two wastes the repair, so confirm which by rail rather than by guess.

No pins captured yet — bench stub.

Schematic facts

OEM input: DC 4.6 V, 900 mA (adapter WAP-002) (the charging cradle CTR-007 is rated the same in and out)^✓^{nintendo-manuals,gbatek}
Charger IC input range: 4.35 V to 6.6 V operating; survives 26 V by suspending (which is why a plain 5 V USB source is inside spec with no PD negotiation)^✓^ti-bq24072
Input current ceiling: approximately 0.82 A, set by the 1.96 kΩ ILIM strap (0.77 to 0.88 A across the datasheet's K_ILIM range)^✓^{ti-bq24072,gbatek}
notes
This is the console’s own computed limit, not the supply’s. It is why a USB-C conversion charges at about OEM speed rather than at the 500 mA a bare USB-C sink is entitled to by specification.
Bulk electrolytics on any documented board: none (there is no published cap map for any of the six boards, and no recap page in the community wikis)^✓^{gbatek,consolemods}
notes
The two boards with published component lists are BGA SoC plus FCRAM plus eMMC, SMD ceramic decoupling and a handful of small-outline ICs. The community failure record on this platform is mechanical, connector and fuse, not dried electrolyte. Chase the fuse, the charger IC, the PMIC or a shorted flex, not the caps.
Game-card clock: up to 16.75 MHz on 3DS (67 MHz / 4); 6.7 or 4.2 MHz in DS mode ('16.6 MHz' in circulation is a rounding of the same figure)^✓^{gbatek,3dbrew}
Case screws: JIS, not Phillips (a true Phillips driver cams out and strips them. JIS #00 or a JIS-compatible bit)^✓^ifixit
Test-point text layer: present on C/CTR-CPU-01, SPR-CPU-01 and FTR-CPU-01; absent on all three New boards (the absence on New boards is a property of those boards, not a documentation gap)^✓^{gbatek,3dbrew}
notes
GBATEK also warns that on the 3DS XL and 2DS it is unknown whether TP1 through TP259 mean the same thing as on the 3DS, beyond the eMMC cluster which is confirmed identical on the 2DS. Do not carry any other TP number across boards without buzzing it.
Region codes: U (US/Canada), E (Europe), J (Japan), K (Korea), T (Taiwan/Hong Kong), C (China) (T and C were never released for any New-series model)^✓^consolemods
notes
Region does not change the mainboard silicon. It is firmware, account territory and a cart region check. On the Old models the region is printed in the model number itself, for example CTR-001(EUR); the New-series manuals print the bare code, so read the case label on those.
Latest stock firmware: 11.17 for US, EU and JP; 11.16 for Taiwan, China and Korea (firmware level, not region, is what decides the software entry route at intake)^✓^consolemods

Where my two sources disagree

Four numbers on this device have two published readings and no way to pick between them from a desk. They are not marked unverified because nobody checked; they are unresolved because two careful sources checked and got different answers. Meter these rather than trusting the table.

TP3 voltage. GBATEK says +3.3 V, 3dbrew measured +2.8 V. GBATEK is internally consistent, but 3dbrew’s own TP6 also reads 2.8 V, which looks like a transcription slip in its table. Probe before you rely on either.
TP96. GBATEK calls it SD/MMC CMD through a 57 ohm series resistor, 3dbrew reads +3.2 V on it. Nobody has reconciled those.
TP126, the shell sensor. GBATEK’s test-point page says 1 = low = closed, and the same author’s GPIO page says 1 = Shell Closed. Those are compatible only if the bit inverts between the pin and the register, so confirm the sense on your board before wiring anything to it.
RED-001 battery fuse. GBATEK puts it on the button board; Sthetix’s repair walkthrough finds a second fuse on the mainboard. One source each way, so check both PCBs.

Two gaps are worth knowing before you start a job on the less common models. SPR-001 is the least documented board of the six: no PMIC, charger, gyro, accelerometer, IR or MCU part data appears in any source I found, and the mainboard silkscreens for SPR, FTR and KTR are photographed nowhere I can reach at a resolution that resolves text. And U10’s real part is unidentified by anyone. The H0X/H1X marking has never been published as a part number, so the DRV5032 you will find suggested for it is refuted, not merely unconfirmed: that device shipped in April 2017, six years after the console.

Reference confidence key

How to read the confidence tags and source citations above. Nothing on this page is tagged bench: this device is documentation-derived and none of these figures are my own measurement.

Sources

3dbrew: 3dbrew.org Hardware, Pinouts, TP_Nets and Gamecards wiki pages: an independently derived community source with different authors and a different method (the TP_Nets survey was measured with a PicoScope 2206B), plus high-resolution mainboard and eMMC photographs. Retrieved 2026-08-05
consolemods: ConsoleMods wiki 3DS pages: model differences, region information, power-fuse replacement, hinge replacement, the 2DS USB-C charging-port mod and the New 3DS XL wireless-charging mod
gbatek: GBATEK 3DS pages (Martin Korth, problemkaputt.de): reverse-engineered from real boards; the only source with complete numbered connector tables, a chip-pinout list with board wiring, and a numbered test-point map. Its connector page is headed 'Connectors in Old 3DS with Mainboard C/CTR-CPU-01' and its component list marks the 3DS XL, 2DS, New 3DS and New 2DS XL as unknown. Retrieved 2026-07-25
ifixit: iFixit device pages, repair guides, parts catalogue and Answers threads for the six models. Guide-grade for procedure and part fitment; Answers threads are forum-tier and are labelled as such where they carry a claim
littelfuse-467: Littelfuse 467 series datasheet, Part Marking System table on p.3 (identical across the 2012, 2017, 2018 and 2021 revisions) and the current-revision electricals on p.1
nintendo-manuals: Nintendo operations manuals for the 3DS, 3DS XL, 2DS, New 3DS / New 3DS XL and New 2DS XL (OEM product documentation, downloaded 2026-08-05). Authoritative for adapter ratings, battery pack codes, panel sizes and model numbers. They never name a chip
sdspec: SD card physical interface assignment as tabulated on Wikipedia's SD card article, retrieved 2026-08-05: the standard 9-contact full-size and 8-contact microSD bus, not a Nintendo-specific document
silkscreen: Nintendo's own silkscreen, read at magnification off the 3dbrew CTR-001 side-A and side-B board photographs and the JAN-001 board photographs. Carries the connector designators, the end pin numbers, the P2 pin-1 triangle and the P4 battery polarity mark, independent of both wikis
ti-bq24072: Texas Instruments BQ24072 datasheet SLUS810N, Figure 7-1, Table 7-1 and Table 7-2: pin assignment, VQFN-16 (RGT) 3.00 x 3.00 mm package, the 'CKP' top marking, the 4.35-6.6 V input range and the input-current-limit programming equation

Confidence: verified two+ sources agree · single source one source / scan-derived, treat with care · bench my own measurement or practice.

Sources and further reading

These are the outside references I lean on for the 3DS family. I link them rather than copy them.

GBATEK, Martin Korth (the connector tables, the chip pinouts with board wiring, and the numbered test-point map; everything on this page tagged single-source traces here)
3dbrew wiki: Hardware and Pinouts (the independent second source, the per-model silicon table, and the board and eMMC photographs)
3dbrew: Gamecards (the card interface, and the software gating on card VCC)
ConsoleMods: 3DS model differences and region information
ConsoleMods: power fuse replacement and hinge replacement
ConsoleMods: 2DS USB-C charging port mod
iFixit: Nintendo 3DS (per-model teardowns and repair guides; the device pages for the other five models branch from here)
repair.wiki: blue light turns on then off (the slow-fade versus instant-cut fork, and the hardware branch’s fault list)
3ds.hacks.guide (the canonical entry-route decision logic and the post-install troubleshooting path)
boot9strap, SciresM and Luma3DS
Texas Instruments BQ24072 datasheet (the charger pin assignment, package, marking and the input-current-limit equation)
Nintendo support: 3DS eShop discontinuation and online services discontinuation (the dates behind the CFW value argument)
rorosaurus/3ds-xl-usb-c and makhowastaken/3DS_C (open-source USB-C boards; read the READMEs for compatibility, not the repository descriptions)
giltesa Nintendo 3DS USB-C Kit and 2DS USB-C

If you would rather buy a console that has already had this work done, everything I restore is in the shop. I do not have 3DS listings up yet, and I will link them here when I do rather than pointing you at something that only nearly fits.