Sprites mods - Magic Printer Cartridge Paintbrush

ESP32 Inkjet Cartridge Controller Project - Hardware Debugging Help Needed

I'm reproducing Jeroen Domburg's HP63 cartridge controller project (Magic Printer Cartridge Paintbrush) and have encountered several hardware failures. Looking for advice on debugging strategy and potential design issues.

Project Status: Successfully achieved some ink output (cyan, occasional yellow) before hardware failures occurred. Using Jeroen's original KiCad files and exact component specifications.

Hardware Architecture:

• 3-board system: PSU board (3.3V/9V/16V rails), ESP32 board, cartridge control board
• MC14504B level converters for 3.3V to 9V/16V translation
• Custom power protection circuit for nozzle drive (10µs pulse limiting)
• ESP32-S3 as programmer, GPIO22 substituted for GPIO12 (to avoid using bootstrapping pin)

Current Issues:

1. Level Converter Behavior (MC14504B):
   • Inconsistent signal propagation delays under load
   • Some cartridges require timing adjustments to function
   • DCLK signal integrity issues between ESP32 output and level converter output
   • Suspected latch-up when VCC pins left floating during initial assembly
2. Power Supply Problems:
   • 9V rail jumping to 15.7V when cartridge connected (should remain 9V)
   • Current spikes causing brownout detection on ESP32 (triggers at 2.44V threshold)
   • Final failure: VCC/GND short on ESP32 after power supply voltage drop
3. Assembly Sequence Issues:
   • Initial assembly with floating VCC pins on level converters caused component damage
   • Replacement of U13 (MC14504B) resolved initial voltage issues
   • Subsequent failure during operation with cartridge connected

Measurements (V_in = 4.2V):

• Idle (no cartridge): 45mA
• Cartridge connected, no dispensing: 45mA
• Dispensing without cartridge: ~80mA
• Dispensing with cartridge: ~150mA

Logic Analyzer Results:

• ESP32 outputs appear correct per waveform templates
• Power protection circuit functions correctly (10µs pulse limiting verified)
• DCLK signal shows inconsistencies between ESP32 and level converter outputs

Specific Questions:

1. Assembly Strategy: What's the recommended power-up sequence for MC14504B-based designs? Should VCC always be applied before input signals?
2. Level Converter Issues: Given MC14504B's limited current output and propagation delays, are there better alternatives for 3.3V to 9V/16V level shifting in this application?
3. Protection Recommendations: What additional protection (diodes, current limiting resistors) would prevent ESP32 damage from power supply issues?
4. DCLK Signal Integrity: How can I debug and correct the timing inconsistencies in the DCLK path through the level converters?