The short story: Global pandemic causes chip shortage. LPC4330 goes out of stock worldwide. Team redesigns around STM32H745. Three years late but better hardware.
The long story is more interesting.
Early Development (2015-2019):
The Kickstarter (2019):
The Shortage (2020):
The Pivot (2020-2021):
The Delivery (2023-2025):
Total Delay: ~3 years from original April 2020 planned delivery to May 2023 actual delivery.
The Chip Shortage: The COVID-19 pandemic caused unprecedented disruption to semiconductor manufacturing and supply chains. The automotive industry, expecting reduced demand, cancelled chip orders. When demand rebounded faster than expected, they panic-bought everything available. Consumer electronics demand surged with work-from-home adoption. The result: massive chip shortage lasting years.
Impact on Smoothieboard:
Critical Timing: The shortage was discovered weeks before the final production prototype was scheduled. The design was nearly complete. Functional prototypes had been tested. But with no chips available and no timeline for when they would be, continuing with LPC4330 meant indefinite delays - potentially years.
The “Escape Hatch”: Fortunately, the team had built design flexibility into the PCB layout. They call it an “escape hatch” - the layout allowed swapping MCUs if supply issues occurred. This foresight, which seemed paranoid at the time, saved the entire project.
Options Considered:
Why STM32H745:
Turns out the chip shortage forced an upgrade, not a compromise.
Most specifications were not finalized before cancellation. The following represents the design intent based on Kickstarter descriptions, prototype testing, and early development work.
| Component | Specification |
|---|---|
| MCU | NXP LPC4330 |
| Architecture | Dual-core ARM Cortex-M4 + M0 |
| M4 Clock | Up to 204 MHz |
| M0 Clock | Up to 204 MHz |
| Flash | External SPI/QSPI (LPC4330 has no internal flash) |
| RAM | 264 kB SRAM |
| FPU | Single precision (M4 core only) |
| Package | BGA (exact type not finalized) |
Dual-Core Plan:
The lack of internal flash was a design consideration - external flash would need to be added to the PCB, adding cost and complexity.
The driver situation evolved during development:
Original Plan: TMC2260
Revised Plan: TMC2590
The TMC2260 testing on LPC4330 prototypes was valuable - it proved the driver interface worked, and that knowledge transferred to the STM32H745 redesign.
Gadgeteer Expansion:
SD Card:
Ethernet:
USB:
Power:
MOSFETs:
Thermistors:
Endstops:
The exact details of protection circuits, MOSFET configuration, and power architecture were not finalized before the project was cancelled.
Functional prototypes were built and tested:
What Was Built:
What Was NOT Completed:
Prototype Count: Likely fewer than 10 units. These are in possession of the development team and were never sold. They’re not compatible with v2 Prime firmware (different MCU architecture).
Paid For: LPC4330-based Smoothieboard v2 Prime ($155)
Received: STM32H745-based Smoothieboard v2 Prime ($155)
| Feature | LPC4330 (Paid For) | STM32H745 (Received) | Result |
|---|---|---|---|
| Cores | M4 + M0 | M7 + M4 | ✅ Better (both cores more capable) |
| M4/M7 Speed | 204 MHz | 480 MHz | ✅ 2.35× faster |
| Secondary Core | M0 (low-power) | M4 with FPU | ✅ Much more capable |
| RAM | 264 kB | 1 MB | ✅ 3.8× more |
| Flash | External | 2 MB internal | ✅ Better (simpler, faster) |
| FPU | Single precision (M4 only) | Double precision (M7) + Single (M4) | ✅ Better |
| Availability (2020) | None | Available | ✅ Actually exists |
| Price | $155 | $155 | ✅ Same |
| Delivery Date | April 2020 (planned) | May 2023 (actual) | ❌ 3+ years late |
The Bottom Line: Backers received a free upgrade to significantly more powerful hardware. The STM32H745 is better in every technical measure. The downside was the 3-year delay.
Community Reaction:
The team’s open communication about the challenges helped maintain community trust despite the significant delays.
Campaign Details:
Variants Offered:
v2-mini:
v2-prime:
v2-pro:
Only the v2-prime variant eventually shipped, and only with STM32H745. The v2-mini and v2-pro were permanently cancelled.
When design work began (2015-2019), choosing the LPC4330 was logical:
Team Experience: The Smoothieboard v1 used the NXP LPC1769. The team had years of experience with NXP’s LPC series, development tools, and programming models.
Logical Evolution: LPC4330 was a natural progression from LPC1769:
No Availability Concerns: In 2015-2019, chip availability wasn’t a major concern. The LPC4330 was readily available from multiple distributors. No one predicted a global pandemic would cause years-long chip shortages.
Dual-Core Appeal: The M4+M0 architecture was attractive for separating real-time tasks (M4) from auxiliary tasks (M0). This would have improved responsiveness over v1’s single-core design.
The choice made complete sense given the information available at the time. The chip shortage was unprecedented and unpredictable.
This project demonstrates several valuable principles for hardware development:
The “escape hatch” in the PCB layout saved the project. The team deliberately designed in the ability to swap MCUs if needed. At the time, this seemed like paranoid over-engineering. In retrospect, it was brilliant foresight.
Takeaway: Build flexibility into designs. Don’t assume your primary component will remain available.
The 2020-2023 chip shortage affected everyone from hobbyists to Fortune 500 companies. Automotive manufacturers, with their massive purchasing power, couldn’t get chips. Small open-source projects had no chance with single-source dependencies.
Takeaway: Supply chain risk is real and can last years. Have backup plans.
The Smoothieboard team communicated openly with backers throughout the crisis. They explained the shortage, the redesign decision, the challenges, and the timeline. This honesty maintained community trust despite 3-year delays.
Takeaway: Transparent communication during crises maintains community support.
The forced redesign resulted in a better product. The STM32H745 is objectively superior to the LPC4330. Without the shortage forcing the change, backers would have received less capable hardware.
Takeaway: Sometimes constraints force improvements that wouldn’t have happened otherwise.
Redesigning around a different MCU is not a simple swap. Here’s what was involved:
Complete PCB Redesign:
Firmware Rewrite:
Component Sourcing:
Prototyping and Testing:
Timeline Breakdown:
Each phase took longer than expected due to ongoing supply chain challenges and the complexity of the redesign.
The LPC4330 design is permanently cancelled. There will be no LPC4330 firmware, no LPC4330 support, and no revival of this design.
All current Smoothieboard v2 boards use STM32H745.
If you want a Smoothieboard v2, you want the v2 Prime with STM32H745.
Q: Can I still buy the LPC4330 version?
A: No. It was never manufactured for customers. Only a handful of development prototypes exist.
Q: Is the LPC4330 version better than the STM32H745?
A: No. The STM32H745 is better in every measurable way:
Q: Why did the team choose LPC4330 originally?
A: Experience with NXP from v1, logical evolution from LPC1769, good availability in 2015-2019, dual-core architecture was attractive. It was a sensible choice given available information.
Q: Could the team have predicted the chip shortage?
A: No. The COVID-19 pandemic and resulting multi-year chip shortage was unprecedented. Even major corporations with massive resources were caught off-guard. Design started years before the shortage (2015-2019).
Q: Why did it take 3 years after the redesign decision?
A: Complete PCB redesign, firmware rewrite for different architecture, ongoing component shortages affecting other parts, multiple prototyping rounds, and testing. The redesign was not a simple MCU swap - it was effectively designing a new board.
Q: Will there ever be LPC4330 firmware support?
A: No. All development is on the STM32H745 version. The LPC4330 design is permanently abandoned.
Q: What happened to v2-mini and v2-pro?
A: They were cancelled along with the LPC4330 design. Only v2-prime was eventually manufactured, and only with the STM32H745.
This design represents an interesting case study in open-source hardware development during unprecedented supply chain disruption.
What It Demonstrates:
For Developers:
For Historians:
This page documents the cancelled LPC4330 design for historical purposes. Last updated: 2025-10-21