imrishabh18/rp2040-motor-controller

engineering/15v-pd.md

# Historical 15 V PD revision — 2026-09-26

Superseded by [the 20 V DRV8825 redesign](20v-drv8825.md). These checks apply only to the archived 15 V circuit.

The user requested that DRV8847 remain. The TMC2209 exploration is abandoned.
The active design requests 15 V, the highest CH224K setting below DRV8847's
18 V recommended operating ceiling. CH224K does not provide an 18 V setting.
The driver has a 20 V absolute maximum; that is not a usable nominal supply.

## Changes

| Reference | New part / setting | Reason |
|---|---|---|
| R_PD_CFG1 | 56 kΩ, 1%, 0603WAF5602T5E, C23206 | CH224K resistor-mode 15 V request; CFG2/CFG3 stay open |
| R_PD_VDD | 1 kΩ, 1%, ERJPA3F1001V, C441922, 330 mW, 0603 | Replaces overloaded 100 mW feed resistor |
| C_MOTOR_BULK | TCJD107M025R0055E, C2159920, 100 µF / 25 V polymer | Voltage headroom and derating; same D case and exact supplier land pattern as former 220 µF / 16 V part |
| D_MOTOR_TVS | SMBJ16CA-13-F, C71870, SMB | 16 V stand-off tolerates the nominal 15 V rail's +5% case |

The bulk capacitor's imported JLCPCB STEP/OBJ model and footprint are retained.
The new footprint is byte-for-byte identical to the former D-case footprint;
its capacitance and part ratings are different. The new import explicitly marks
it polarized and rated 25 V. The two 10 µF motor-rail ceramics are already 25 V;
that rating is now explicit in TSX. The fuse remains 2 A / 30 V.

DRV8847, approximately 1 A peak current regulation, INA241 telemetry, RGB LED,
TMP102/AND-gate shutdown, buzzer, JST and optional bottom DNP encoder are unchanged.
The programming USB remains
fixed 5 V; the motor PD rail is not connected to the 3.3 V regulator input.

## Nominal supply calculations and limits

- 15 V + 5% = 15.75 V, below the driver's 18 V operating ceiling.
- The new bulk capacitor runs at 63% of its 25 V rating at 15.75 V. The old
  16 V part would run at 98.4%, without useful derating or transient margin.
- Worst-case conservative R_PD_VDD dissipation with 3.0 V VDD and a 1% low
  resistor is (15.75 − 3.0)^2 / 990 = 0.164 W. The 330 mW replacement provides
  nominal margin; verify board temperature and the manufacturer's power derating.
- Current regulation is independent of the chosen PD voltage. Do not raise the
  approximately 1 A limit just because the supply voltage increased.

## Not yet production-qualified

The TVS is **not** a precision overvoltage clamp. Its maximum breakdown is
20.5 V and its rated surge clamp is 26 V, exceeding the driver's 20 V absolute
maximum (and the bulk capacitor's 25 V rating at the rated TVS surge). It cannot
be used as evidence that motor regeneration or hot-plug spikes are safe.
The smaller bulk capacitance also increases voltage rise for the same returned
energy. Passing DRC, shorts and GND checks establishes layout connectivity only.

Before manufacturing a production batch, scope VM at the driver's power/ground
pins during hot-plug, start/stop, worst-case deceleration and externally driven
motor motion at maximum intended load and temperature. Verify ripple current,
component heating and USB power sequencing. Keep operating VM within 18 V;
20 V must never be exceeded. If the envelope cannot meet these limits, implement
and qualify an active clamp/braking circuit, or stay with a lower supply. The
existing hardware temperature shutdown does not protect against overvoltage.

Sources:
- [TI DRV8847 datasheet, sections 6.1 and 6.3](https://www.ti.com/lit/ds/symlink/drv8847.pdf)
- [WCH CH224 datasheet, section 5.2.1](https://aitendo3.sakura.ne.jp/aitendo_data/product_img/ic/power/CH224K/CH224K.PDF)
- [KYOCERA AVX TCJ ratings and derating](https://datasheets.kyocera-avx.com/TCJ.pdf)
- [Panasonic ERJPA3F1001V specifications](https://industrial.panasonic.com/ww/products/pt/small-and-high-power-chip-resistors/models/ERJPA3F1001V)
- [Diodes SMBJ16CA specifications](https://www.diodes.com/part/view/SMBJ16CA)

## Local verification

The regenerated circuit has zero embedded DRC errors and zero independent
routing errors. `tsci check` reports zero errors / 166 retained warnings. The
Gerber shorts check passes on all four copper layers. The independent copper
geometry audit finds one GND network and 12 connected pour regions.
All 19 circuit tests and TypeScript checking pass.
Assembly exports use the new four supplier part numbers and still omit U_ENCODER.

The exact copper geometry is unchanged. Explicit capacitor polarity reverses
its generated internal port-ID order; those two IDs were remapped in the saved
route plan after checking that all other input data matched exactly. Positive
pad position and connectivity are regression-tested. The saved ground-pour
regions also match the prior board exactly.

Separate CLI placement analysis still emits existing heuristic recommendations
(its placement command exits 1 despite zero placement DRC errors). Schematic
analysis retains the same 56 advisories as the 12 V baseline after ignoring
internal trace-ID numbering. These are not claimed as clean analysis results.
See [revision-checks.json](revision-checks.json) and local `reports/15v-*.log`.