← AXS-044

Circuit review & bench-test guide

AXS-044 — PMS5003 particulate-matter carrier PMOD module

Document purpose#

This document explains the axs-044-particulate carrier at component level and turns the design evidence into a manual-review and bench-test plan. It is based on generate_design.py, the generated schematic, the README, and the Plantower PMS5003 data manual.

The board is an unfabricated prototype: schematic generated, ERC clean (0/0), netlist reviewed, no PCB layout, no fabricated board, no BOM export, no bench results. Two connector MPNs are TBD:. The J2 pad-to-pin mapping is explicitly flagged by the design itself as unverified against hardware.

1. What the board does#

The carrier adapts a Plantower PMS5003 laser-scattering particulate sensor (PM1.0/PM2.5/PM10) to a PMOD Type 3 (UART) socket with two control extras: SET (sleep) on PMOD pin 1 and RESET on PMOD pin 7. The PMS5003 is 5 V powered — its fan needs 5 V — with 3.3 V logic, so external 5 V enters on screw terminal J3 and the PMOD stays strictly 3.3 V. The sensor streams fixed 32-byte frames (start 0x42 0x4D) at 9600 baud in active mode.

Functional block diagram#

EXT 5 V bench supply ──> J3 ──┬──> J2 pins 1/2 (VCC) ──> PMS5003
                              ├── C1 100 nF
                              └── C2 10 µF
PMOD host (ECP5) J1                                PMS5003 (via its supplied cable, J2 2x5 1.27 mm)
  pin 1 PM_SET ────┬───────────────────────────────> SET   (R2 10 k pull-up to 3V3)
  pin 7 PM_RESET_N ┬───────────────────────────────> RESET (R3 10 k pull-up to 3V3)
  pin 2 TXD ───────────────────────────────────────> RX
  pin 3 RXD <── R1 100 Ω ──────────────────────────  TX (PM_TX)
  pins 5/11 GND ── common ground ── J2 pins 3/4 ── J3 GND
  pins 6/12 3V3 ── pull-up rail + decoupling reference only (never powers the sensor)

2. Safety and scope boundaries#

• The PMS5003 contains a laser diode. It is enclosed, but do not open the metal shell ("do not break up the sensor" — Plantower), and stop using a unit with a damaged housing.
• The metal shell is internally connected to GND (Plantower installation note). Do not let it touch any other node, and account for it when probing.
• Never feed 5 V into any PMOD pin. The 5 V domain is J3, C1, C2, and J2 pins 1/2 only.
• Plantower cautions against sustained use in high-concentration environments (kitchens, water mist, outdoors without protection); bench smoke tests should be brief.
• Use a current-limited supply (≈300 mA) for first power-up.

3. Supply sequencing, common ground, and no-host behavior#

1. Common ground first. J3 GND, J2 pins 3/4, and PMOD GND are one net. Buzz continuity before power; never feed J3 from a supply with a separate ground path.
2. Recommended order: PMOD host on first, then external 5 V. The sensor's SET/RESET idle states are guaranteed by R2/R3 pulling to the host 3V3, so the 3V3 rail should exist before the sensor boots. Power down in reverse (5 V off first).
3. 5 V with no PMOD host attached: the sensor runs (its SET/RESET are also pulled up internally, per the datasheet's circuit attentions) and drives UART frames into R1 toward an open pin — harmless. R2/R3 then tie SET/RESET to a dead 3V3 net; the sensor's internal pull-ups leak µA-level current through 10 k into that net — benign.
4. 5 V with the host attached but unpowered is prohibited: PM_TX drives ≈3.3 V through R1 (100 Ω) into the unpowered FPGA input and back-powers the 3V3 rail through its protection diode at up to tens of mA; SET/RESET internal pull-ups add a smaller path through R2/R3.

4. Plug-in module verification (read before first power)#

Mandatory section. The PMS5003's connector is the single most error-prone part of this design, and the design itself flags it.

4.1 Connector and pin-order check#

Two facts must be reconciled on real hardware:

• The carrier assumes the widely documented 10-pin layout (1.27 mm 2×5, pad N = sensor pin N): 1 VCC, 2 VCC, 3 GND, 4 GND, 5 RESET, 6 NC, 7 RX, 8 NC, 9 TX, 10 SET.
• The Plantower data manual V2.3 prints an 8-entry pin table: 1 VCC, 2 GND, 3 SET, 4 RX, 5 TX, 6 RESET, 7/8 NC. Later Plantower documents and the shipped 1.27 mm cable correspond to the 10-pin layout, but note that RESET and SET sit at different positions in the two tables. If the purchased unit follows the 8-pin table, plugging it into a 10-pin-wired carrier would put GND on a VCC pad and SET on a GND pad.

Resolution procedure, sensor in hand and carrier unpowered:

1. Identify the connector on the purchased PMS5003 (count positions: 2×5, 1.27 mm) and locate its pin-1 marking (Plantower's drawing shows PIN1 at one end of the recessed header; the supplied cable is keyed).
2. With a DMM in continuity mode from the sensor end of the supplied cable, map each cable position to a sensor pin. The shell is GND — use it to identify the GND positions confidently.
3. Buzz each cable position through to the carrier J2 pads and confirm: VCC positions land on the EXT_5V net (to J3 pin 1), GND positions on GND, TX on the R1 side, RX on the TXD net, SET on PM_SET, RESET on PM_RESET_N.
4. Verify the mated orientation cannot be reversed (shrouded/keyed header chosen for J2 — the MPN is still TBD:, so this property must be part of the part selection).
5. Only then apply power. Record the mapping with photos.

4.2 Plug-in key specifications (Plantower PMS5003 manual V2.3, 2016-06-01)#

5. Component-by-component review#

#FLG01–#FLG03 are ERC power flags, not physical parts.

Note: Plantower says SET and RESET "should not be connected if without usage" because they are internally pulled up. The carrier connects them through its own 10 k pull-ups plus host pins — functionally equivalent to "unused" when the host tri-states, and controllable when the host drives. This is a deliberate, defensible deviation; verify on the bench that the idle levels read high.

6. Datasheet summary and design interpretation#

Official references: Plantower PMS5003 product page, PMS5003 data manual V2.3 (Adafruit mirror). Later manual revisions exist; obtain the revision matching the purchased unit (this matters for the connector table — see section 4.1).

7. Expected values before bench testing#

8. Manual schematic and assembly review checklist#

• Complete the section 4.1 connector mapping against a purchased PMS5003 + cable before layout, and pick a keyed/shrouded J2 (close the TBD:).
• Pin J1/J3 MPNs; export and availability-check the BOM.
• After layout: 5 V net confined to J3/C1/C2/J2 pins 1/2; 5 V keep-out from the PMOD edge; silkscreen polarity at J3; pad-1 marking at J2 matching the cable key.
• Verify R1 is in the sensor-TX line; R2/R3 to 3V3 (not EXT_5V — a pull-up to 5 V would violate the sensor's 3.3 V logic and the FPGA pins).
• Mounting: per Plantower, keep the shell's inlet/outlet unobstructed and do not screw into the sensor deeper than 5 mm.

9. Ordered bench-test procedure#

Stop at the first abnormal result.

A. Unpowered checks#

1. Complete section 4.1 mapping. Do not skip it.
2. Resistance EXT_5V-to-GND and 3V3-to-GND; investigate hard shorts.
3. Continuity J3 GND ↔ PMOD GND ↔ sensor shell (shell is GND-bonded).

B. First power, active mode#

1. Host on first, then J3 at 5.00 V with a 300 mA limit.
2. Confirm fan audibly spins; record current (spin-up surge, then ≤100 mA).
3. Capture PM_TX at 9600 8N1: frames start 0x42 0x4D, length 28, checksum valid. Log the frame interval (2.3 s stable mode is normal in clean air).
4. Read PM2.5 in room air: a plausible low value with both calibrations present in the frame. Values are not being graded against a reference — Plantower only promises consistency against their own sensors.

C. SET/RESET line tests#

1. Sleep: drive PM_SET low from the host. Fan stops; frames stop; current falls to the µA range (target ≤200 µA — measure in the J3 feed with the host still powered).
2. Wake: release/drive PM_SET high. Fan restarts; frames resume; mark the first frame time and discard 30 s of data before judging values; compare a reading at t+35 s against pre-sleep values.
3. Reset: pulse PM_RESET_N low ≥1 ms. Sensor reboots; frames stop and resume; confirm idle level returns to 3.3 V.
4. With the host tri-stating both lines, confirm the sensor free-runs (pull-up sanity — R2/R3 doing their job).

D. Passive-mode protocol (optional but recommended)#

1. Send 0xE1 (change mode → passive), then 0xE2 read requests; verify one 32-byte answer per request.
2. Send 0xE4 sleep/wake commands and compare behavior with the SET-pin test.
3. Return to active mode and power-cycle to confirm the default state.

E. Aerosol response and soak#

1. Briefly present a controlled particle source near the inlet (e.g., an extinguished match at a distance); PM2.5 should rise within ≤10 s (total response time) and decay afterward. Keep it brief — high concentration exposure is a documented stressor.
2. 2-hour soak in room air: frame-loss/checksum-error count (target 0), current stability, and value drift.

10. Troubleshooting map#

11. Bench record template#

12. Review conclusion#

The design mirrors Plantower's own typical application circuit: external 5 V for the fan, 3.3 V logic straight to the host, and 10 k pull-ups to 3.3 V on SET and RESET. The decoupling and series-resistor conventions match the rest of the board family. The principal risks are (1) the J2 pad-to-pin mapping against the real connector and cable — made sharper by the fact that Plantower's own V2.3 manual prints a different (8-entry) pin table than the 10-pin layout the carrier assumes, with RESET/SET in different places; (2) the unkeyed-header hazard until the TBD: J2 part is pinned as a shrouded, keyed component; (3) back-powering an unpowered host through R1; and (4) the absence of layout, fabrication, and bench evidence.