Fully Wired · Volume 6
What Goes Wrong, What It Costs, and What to Decide
The failure modes of a tethered gun and the guard against each, the bench checks that turn this dive's assumptions into measurements, the whole wired setup priced, and the decision rows it proposes
6.1 What Goes Wrong
A tethered gun fails mostly at its ends. The table lists every failure this dive’s design has an answer to, the answer, and where it is argued.
Table 1 — 6.1 What Goes Wrong
| Failure | What the player sees | Guard | Where |
|---|---|---|---|
| Pulse current in the USB ground | Disconnect sound when the coil fires; gun re-enumerates | Separate 24 V return conductor and pin; one join in the gun; floating brick | Vols 3, 5; Power Supplies Vol 5 |
| A second ground join in the box | The same, intermittently | Plastic box; J1’s sleeve joined to core 2 only | Vol 4, 4.2–4.3 |
| GX mated one position out | Dead gun until the key drops in | 24 V on the centre pin; VBUS between the pedal pins; brick unplugged during mating | Vol 3, 3.5 |
| Brick plugged in live at the box | Spark at J1; overshoot at J12 | Mating order on the lid; LED shows “live”; electrolytic reservoir and SMBJ28A | Vol 4, 4.6 |
| Thin USB cable | Gun browns out or will not enumerate on some ports | 28/24 (or 28/22) cable: 63 mV ground drop at the gun’s draw over 4 m | Vol 2, 2.3 |
| Flex fatigue at the gun end | Intermittent USB, a cut-out when the gun is twisted | Constant-flex cable for the power cores; heat-shrink boot sets an 8 × D bend; jackets stopped inside the boot | Vol 2, 2.6–2.8 |
| Untwisted pair at the cups | Nothing, until it is long | Under about 25 mm each side | Vol 5, 5.2 |
| Umbilical crushed by a chair | Kick stops; gun keeps aiming | The brick hiccups on a 24 V short; USB is separate | Vol 4, 4.4 |
| Earthed brick substituted | Disconnects return | Meter check: output 0 V to mains earth must be open | Vol 5, 5.6 |
| Wrong brick | Blown fuse or TVS; or a weak kick | 6.2 | |
| GX contacts worn or oxidised | Rising drop; then intermittent USB | Contact-resistance check, new and after use | 6.3 |
| GX coupling ring backs off | Intermittent everything | Ring checked at every session; see 6.3 | 6.3 |
| Cable snag | Gun yanked from the hand, PC pulled off the desk | 6.4 | |
| USB-A plug wear at the PC | Loose fit | Leave it in the PC; the box takes the handling. The USB-A connector is specified for 1,500 insertion cycles (USB 2.0 Table 6-7) | Vol 4 |
6.2 The Wrong Brick
The junction box’s barrel jack is a 5.5 × 2.1 mm socket, and every brick of every voltage with that plug fits it. Power Supplies Vol 2 makes the point for the gun; with the jack moved to the box, it applies to the box.
Table 2 — 6.2 The Wrong Brick
| Brick plugged in | What happens |
|---|---|
| 12 V | The regulated stages run with half the headroom. A coil chosen for 24 V may no longer reach its set current: a weak kick, nothing damaged. |
| 19 to 20 V laptop brick | Inside the 18 to 28 V window (D14); headroom reduced (Vol 2, 2.4 gives the largest coil at 18 V as about 6.7 Ω at 2.4 A). Some laptop bricks have earthed outputs (Vol 5, 5.6), so check. |
| 24 V, the intended class | As designed. |
| 36 V, such as a supply bought for an AimTrak (Ultimarc specifies 24 to 36 V) | Above the SMBJ28A’s stand-off and the window’s 28 V top. The TVS conducts heavily, and either F1 opens or the TVS fails. This is the realistic hazard in a house with more than one light gun. |
The guards are the label beside J1 (“24 V ONLY, centre +”), keeping the gun’s brick with the gun, and, if the hazard is judged real, a different jack for the box. A 5.5 × 2.5 mm jack would refuse the 2.1 mm plugs most bricks carry, but the GST60A24-P1J has a 2.1 mm plug, so that would need a re-plugged brick. That trade is not worth making for the first build.
6.3 Connector Wear and a Ring That Backs Off
Contact wear. No GX16 seller publishes a mating life; the one maker figure for an M16 connector of the same size is binder’s “more than 500 mating cycles” (Vol 3, 3.2). A gun unplugged once per session reaches 500 matings in a year of daily play. Wear shows first as rising contact resistance. Because the 24 V contact runs under half its lowest rating, and the regulated stage tolerates a volt of extra drop (Vol 2, 2.4), the first symptom would most likely be on USB: a gun that occasionally fails to enumerate. Check W3 measures it.
The practical policy: leave the gun plugged in at its station. The GX is a connector for packing the gun away, not for every session; the USB-A plug stays in the PC; the brick is unplugged at the wall, not at the gun.
The coupling ring. A threaded ring on a part that takes a solenoid’s kick dozens of times a minute may work loose. A loosened ring lets the plug rock in the socket and the contacts fret. No published figure exists for GX rings under vibration. The check is simple: mark the ring and the panel part with a paint line once tight, and look at it after a session of autofire (W8). If it moves, a thin rubber O-ring behind the ring, or a printed collar that clips over it, holds it.
6.4 Snags
A cable on the floor will be stepped on, caught by a chair, and walked through. The question is what gives way.
- In this design, the GX and the base plate take the pull, then the gland at the box, then the USB-A plug in the PC. If the umbilical is caught near the box, the box (screwed down by its flanges) holds and the gun is pulled from the player’s hand. If it is caught near the gun, the base plate takes it.
- The PC end is the expensive end. A floor box screwed to the desk leg or floor stops a yank reaching the PC’s USB port. This is the strongest argument for screwing the box down.
- A breakaway connector would let go before anything else. No stocked magnetic or breakaway connector was found with eight contacts and a rating for the 24 V pulse (Vol 3, 3.9). The GX should therefore be treated as a non-breakaway joint, and the cable routed accordingly: from the box straight to the floor in front of the player, never across a walkway.
- Tripping. A 3 m cable lying from the screen to the player is a trip hazard for anyone else in the room. The fully wireless architecture’s main advantage over this one is exactly this (Fully Wireless).
6.5 The Bench Checks
Every one uses the lab’s own instruments (the Analog Discovery Pro ADP3450, the programmable bench supplies, the NanoVNA, calipers, a luggage scale). They are numbered so that other volumes can cite them.
Table 3 — 6.5 The Bench Checks
| # | Check | How | Settles |
|---|---|---|---|
| W1 | Fit | Calipers: the GX16 plug’s clamp bore; the panel part’s hole and nut against the printed base plate | Whether the boot of Vol 2, 2.8 fits; GX16 or GX12 |
| W2 | Can a slip happen? | Calipers: key projection ahead of the pin tips; pin recess in the shell; then try to touch a pin to a wrong socket with the key out of its slot | Whether the slip of Vol 3, 3.4 is physically possible at all |
| W3 | Contact resistance | Four-wire: 1 A from a bench supply through each mated contact, ADP3450 across it; new, after 100 and after 500 matings | The 20 mΩ seller figure; wear |
| W4 | Reservoir discharge | Scope the GX centre pin to position 7 after unplugging with the brick live a moment before | How long the gun’s pins stay live (Vol 3, 3.6) |
| W5 | Hot-plug overshoot | Single-shot capture of V_SOL at J12 as the barrel goes into a live brick | Whether the SMBJ28A works hard; whether C4 needs damping (Vol 4, 4.6) |
| W6 | Ground offset | Gun ground minus the PC’s USB shell, two channels, idle and in autofire | The 63 mV estimate; that firing does not move it (Vol 5, 5.4) |
| W7 | USB under fire | USB decode or host disconnect log through several minutes of autofire; then with a 0.5 m lead for comparison | Packet integrity; no disconnects (Vol 5, 5.5) |
| W8 | Mechanical | 40 N pull at the gun and at the box gland; ring witness mark after autofire | Strain relief; ring backing off |
| W9 | 24 V drop | Brick plug and J12 voltages during a regulated pulse | The real loop resistance against Vol 2’s 0.36 Ω |
| W10 | Cable delay and impedance | NanoVNA, time-domain, on the open-ended 4 m USB pair | The cable meets 90 Ω and 26 ns (Vol 5) |
6.6 What the Whole Wired Setup Costs
Prices as read on 2026-10-19 (2026-10-18 for the brick, from Power Supplies), from the pages cited in the volumes. Several sellers refused automated reading; those rows give no price rather than an unverified one.
Table 4 — 6.6 What the Whole Wired Setup Costs
| Item | Pick | Price | Source |
|---|---|---|---|
| 24 V brick | Mean Well GST60A24-P1J | $19.40 (DigiKey) to $24.86 (TRC), 2026-10-18 | Power Supplies Vol 2 |
| USB cable, 28/24, 15 ft | Monoprice 5440 or equivalent | not read | Vol 2 |
| 4 × 20 AWG flex cable, 3 m | Alpha Wire 85004 (or an unrated 4-core) | not read | Vol 2 |
| Expandable PET sleeve, 3 m, 1/2 in | any | not read | Vol 2 |
| GX16-8 connector pair | Tayda A-8801 and its mate, or equivalent | $1.09 for the A-8801 part | Vol 3 |
| Heat shrink, adhesive-lined, for the boot | stock | stock | |
| Micro-USB pigtail | a short micro-USB cable, cut | stock | Vol 3 |
| JST VH plug for J12 | stock | Reference Electronics | |
| Junction box: FB27, gland, bushing, PJ-392, LED, resistor | Tayda | about $4.00 | Vol 4 |
| Barrel jack J1 | Same Sky PJ-005A (or Tayda DC-099, $0.06) | not read | Vol 4 |
| Known total | about $25 to $30, plus the cables and the PJ-005A |
The cables and sleeve are the unknowns, and they are the bulk of whatever is missing: a few dollars for the USB cable and the sleeve, more for flex-rated control cable bought by the foot. A builder using a cheap unrated four-core cable for the first bench umbilical (Vol 2, 2.10’s fallback) will land under $40 for the whole wired setup including the brick.
For scale: Ultimarc charges $25 more for the AimTrak with recoil ($120.00) than without ($95.00), on 2026-10-19, and ships the recoil gun without its supply (Vol 1, 1.6). The wired architecture is the cheap part of a recoil gun; the solenoid, its driver and the shell are where the money goes.
6.7 Decision Rows Proposed
For DECISIONS.md, all Proposed:
Table 5 — For DECISIONS.md, all Proposed
| # | Part / choice | Candidates | Pick | Why, and why not the others | Vol |
|---|---|---|---|---|---|
| WD-1 | Umbilical | USB 28/24 + 4 × 20 AWG flex in one sleeve; + 2 × 18 AWG and a pedal pair; unrated 4-core; Cat 6A S/FTP; multicore with untwisted data; chainflex; 28/28 USB; coiled cord | USB 2.0 28/24 (or 28/22) shielded + 4 × 20 AWG constant-flex (Alpha Wire 85004 class), PET sleeve, 3 m | Real 90 Ω pair; separate 24 V return; pedal on spare cores; flex-rated. Under regulated drive the gauge moves the regulable coil only 8 % from 26 to 18 AWG, so 20 AWG is chosen for robustness and diameter. Cat 6A: 24–26 AWG power and no spare; untwisted multicore: no defined impedance; chainflex: too stiff; 28/28: fails 125 mV | 2 |
| WD-2 | GX16 pin map | 24 V on the circle; VBUS beside ground; six or seven pins | GX16-8: centre +24 V; 1 Pedal, 2 VBUS, 3 Alt Pedal, 4 D+, 5 D−, 6 USB GND, 7 24 V RTN (clockwise from the key, gun’s mating face) | Meets D31 by geometry: no rotation moves the centre. VBUS slips into a pedal input behind 1 kΩ (about 4 mA), never onto a data pin or a reversed supply. Settles CI-7’s pin question | 3 |
| WD-3 | Far-end junction box | plastic passive box; diecast 1590A/B; moulded Y-split; USB-B socket on the box; box with fuse/switch/hub | Passive plastic box (FB27 class): PJ-005A, PJ-392, LED + 10 kΩ, M20 gland, bushing; USB passes through; pedal sleeve to USB GND by a window tap | A metal box joins the jack sleeves (a second ground join); every electronic addition duplicates the brick’s or the carrier’s protection or adds a pulse-loop contact | 4 |
| WD-4 | USB shield termination | both ends; PC end only; gun end only | PC end only (USB-A shell); folded back at the GX | No spare GX pin; no chassis in a printed gun; a gun-end bond is a parallel ground path | 5 |
| WD-5 | Mating order | as D31; plus LED; plus switch | GX first with the brick unplugged; barrel in with the brick unpowered; brick into the wall last; “live” LED in the box; order on the lid | Keeps the only live source during GX mating the current-limited VBUS; avoids hot-plug overshoot; a switch is a pulse-loop contact and easily left on | 1, 4 |
| WD-6 | Pedal ground | USB GND conductor; 24 V return core | USB GND, tapped in the box | The pedal input is referenced to logic ground; the 24 V return carries up to 0.5 V of pulse drop | 4 |
Correction to a sibling, minor. Power Supplies Vol 2 and D31 give the GX16’s current rating as “4 to 7 A” from sellers. On 2026-10-19 Olimex listed its four-pin GX16 at “20 A / 220 V”, so the seller range is wider still (5 to 20 A). The conclusion is unchanged: no maker datasheet was found, and the design runs every contact at under half the lowest figure.
6.8 Questions for the Builders
- Adopt the pin map (WD-2)? It also confirms CI-7: the pedal jack goes in the box.
- Base-plate width (W1): does a 16 mm GX16 panel part and its 18.8 mm nut fit the printed magazine’s floor plate, or does it fall back to GX12?
- LED or switch in the junction box (WD-5)?
- The pedal TVS part for the box, chosen with the carrier’s input protection.
- First umbilical: buy Alpha Wire 85004 by the foot, or start with an unrated 4-core and upgrade if it fails at the boot?
- Brick discipline: is there a 36 V AimTrak-class brick, or any 12 V brick with a 5.5 × 2.1 mm plug, in the same room? If so, consider a keyed connector at the box (6.2).
6.9 What Could Not Be Verified, Across the Dive
- GX16 maker data: none exists that was found; contact resistance at eight pins, mating life and clamp bore are seller figures or unknown (W1, W3).
- Prices of the USB cable, the 85004, PET sleeve and the PJ-005A (seller pages refused automated reading on 2026-10-19).
- Alpha Wire’s flex-test method behind “up to 14 million” cycles.
- The gun’s 5 V draw (150 mA is Power Supplies’ estimate).
- The slip itself (W2): whether a GX16-8 can be mated one position out at all.
- Hot-plug overshoot at J12 (W5) and reservoir discharge after unplug (W4).
- Coupling and packet integrity under autofire (W6, W7): expected clean, unmeasured.
- Linear Technology AN88 was identified but not read; its content is cited only for the general mechanism.
Comments (0)