learning-garden/sources/robotics-lab-pi-replacement.md

---
pageType: source
id: source.robotics-lab-pi-replacement
title: ROBOTICS-LAB-PI-REPLACEMENT
sourceType: local-file
sourcePath: /home/topher/.openclaw/workspace-crash-bot/projects/ROBOTICS-LAB-PI-REPLACEMENT.md
ingestedAt: 2026-05-02T21:20:13.835Z
updatedAt: 2026-05-02T21:20:13.835Z
status: active
growth: seed
---

# ROBOTICS-LAB-PI-REPLACEMENT

## Source
- Type: `local-file`
- Path: `/home/topher/.openclaw/workspace-crash-bot/projects/ROBOTICS-LAB-PI-REPLACEMENT.md`
- Bytes: 26679
- Updated: 2026-05-02T21:20:13.835Z

## Content
```text
# Robotics Lab — Pi 4 Replacement Project

**Status:** Research Phase (Purchase: Next School Year)
**Last Updated:** 2026-03-30
**Decision:** LattePanda IOTA via Mouser (preferred) — CDW business mini PCs (backup)

---

## Quick Summary (TL;DR)

**Problem:** Pi 4s are underpowered (ESP32 compiles: 2-3 min) + Arduino IDE on ARM is 2 years old

**Solution:** x86 boards (Arduino IDE official, auto-updates, fast compiles)

**Recommended:** LattePanda IOTA (DFRobot) via Mouser.com
- **Why:** Approved vendor, educational product, minimal IT scrutiny, x86 = Arduino works
- **Cost:** ~$4,300 for 22 units (includes spares)
- **Timeline:** Order next school year

**Backup:** Business mini PCs (Dell/HP/Lenovo) via CDW

**Keep Watching:** New x86 SBCs, better pricing, educational discounts

---

## How to Update This Doc

**When you find a new board:**
1. Add to "Candidates to Evaluate" table
2. If interesting, add detailed section below
3. Update comparison table
4. Note vendor + pricing + availability
5. Mark status: Research / Watching / Recommended / Rejected

**When pricing changes:**
- Update budget estimates
- Note date + source

**When new info arrives:**
- Arduino IDE updates? → Update ARM vs. x86 section
- New vendor approved? → Update Purchasing Constraints
- Better deal found? → Update Recommendation

---

**Context:** E2/Robotics classroom desks need replacement for underpowered 4GB Raspberry Pi 4s.

**Current Setup:**
- **Hardware:** Raspberry Pi 4 (4GB) — 20-22 units
- **Problem:** Underpowered for workload + Arduino IDE 2 years old (ARM port stale)
- **Primary Workloads:**
  - Arduino IDE (ESP8266/ESP32 compiles: 2-3 min first, 1-2 min subsequent)
  - Node-RED (local IoT, MQTT extensively)
  - Browser + Arduino IDE = sluggish on some machines (suspect SD cards)
- **Environment:** Classroom desks (22 units needed: 20 seats + 2-3 spares)
- **Major Project:** Self-scoring robot competition (FRC-style)

**Arduino Boards in Use:**
- Arduino Uno
- Arduino Leonardo
- ESP8266 (NodeMCU, Wemos, etc.)
- ESP32 (various boards)

**Class Breakdown:**
| Class | Students | Arduino Requirement | Notes |
|-------|----------|---------------------|-------|
| **E2 (Electronics 2)** | 4-10 | **ESP board definitions required** | Local Arduino IDE must have current ESP32/ESP8266 support |
| **Robotics** | 10-20 | Can use Arduino Cloud (but clunky + costs district yearly) | Prefers local IDE, but Cloud is fallback |

**Critical Constraint:**
- **Does NOT want hybrid** (two device types = two instruction sets)
- **Budget might force hybrid** even if undesirable
- **One instruction set for all desks** = priority (E2 + Robotics use same room/desks)

**Budget:** Under $200/unit (can stretch if needed)
**Total Budget:** ~$4,400-5,000 for 22 units

---

## Requirements

**Must Have:**
- [ ] Significantly faster ESP32/ESP8266 compiles (target: <1 min first, <30 sec subsequent)
- [ ] Smooth Browser + Arduino IDE multitasking
- [ ] **Arduino IDE 2.0+** — simple interface for intro classes (PlatformIO too complex)
- [ ] USB ports for Arduino boards (CH340, CP2102, ATmega16U2 drivers)
- [ ] Ethernet + WiFi (classroom network, MQTT)
- [ ] HDMI display output
- [ ] Reliable storage (SD card issues suspected)
- [ ] 5V power (simple classroom setup)

**Important Context:**
- **PlatformIO was evaluated** — rejected for intro classes (too complex)
- Students need the "green arrow" simplicity of Arduino IDE
- Build systems, platformio.ini, etc. = cognitive overhead for beginners
- **Arduino IDE must work reliably** — not a hacky port that breaks

**Nice to Have:**
- [ ] PoE support (cleaner cable management)
- [ ] SSD boot (faster, more reliable than SD)
- [ ] Educational pricing / bulk discounts
- [ ] Good documentation for students
- [ ] Durable (student handling)
- [ ] Easy disk imaging for 22 units

**Constraints:**
- 22 units needed
- Must work with existing Arduino boards (USB driver compatibility)
- MQTT network traffic (22 nodes publishing/subscribing)
- Student-proof (recovery when things break)

---

## Root Cause Analysis

**ESP32 Compile Times (2-3 min first):**
- ESP32 toolchain is heavy (xtensa-esp32-elf-gcc)
- First compile = full build, subsequent = incremental
- Pi 4 CPU (1.5GHz quad) is the bottleneck
- **Solution:** Faster CPU + more cores + SSD for file I/O

**Browser + Arduino Sluggish:**
- Chromium + Arduino IDE (Electron-based) = RAM hungry
- SD card I/O = slow + can corrupt
- **Solution:** 8GB+ RAM + SSD boot

**SD Card Issues:**
- Cheap SD cards = slow + unreliable
- Student abuse (power pulls, crashes) = corruption
- **Solution:** SSD boot (USB 3.0) or eMMC

---

## Candidates to Evaluate

### ARM Boards (Arduino IDE = 2 years old, community port)

| Board | CPU | RAM | Storage | Price | Notes |
|-------|-----|-----|---------|-------|-------|
| **Pi 4 (current)** | Quad 1.5GHz | 4GB | SD | ~$55 | Underpowered, SD bottleneck |
| **Pi 5 (4GB)** | Quad 2.4GHz | 4GB | SD/PCIe SSD | ~$60 | 2-3x CPU, PCIe for SSD |
| **Pi 5 (8GB)** | Quad 2.4GHz | 8GB | SD/PCIe SSD | ~$80 | ✅ Sweet spot for RAM |
| **Orange Pi 5 (8GB)** | Octa 2.4GHz (RK3588S) | 8GB | eMMC/SD | ~$65 | ✅ 4x CPU cores, eMMC option |
| **Orange Pi 5 (16GB)** | Octa 2.4GHz | 16GB | eMMC/SD | ~$75 | Overkill RAM, good price |
| **Orange Pi 5 Plus** | Octa 2.4GHz | 8/16GB | eMMC/SD/PCIe | ~$80-90 | Dual Ethernet, M.2 SSD |
| **Rock Pi 4C+** | Hexa 1.8GHz (RK3399) | 4GB | eMMC/SD | ~$55 | Pi 4 form, eMMC option |
| **RDK X5** | Octa-core A55 (ARM, 10 TOPS NPU) | 4/8GB | eMMC | ~$100-130 | ROS/AI focus, ARM (same Arduino issue) |

### x86 Boards (Arduino IDE = Official, auto-updates)

| Board | CPU | RAM | Storage | Price | Notes |
|-------|-----|-----|---------|-------|-------|
| **LattePanda IOTA** (DFRobot) | Quad 3.6GHz (Intel N150) | 8/16GB | 64/128GB eMMC | ~$129-179 | ✅ SBC form, RP2040 co-processor |
| **Orange Pi X** | Quad 3.4GHz (Intel N100/N150) | 8/16GB | eMMC/SSD | ~$100-130 | ✅ Budget x86, new 2025 |
| **Intel N100/N150 Mini PCs** (GMKtec, Beelink, Acemagic) | Quad 3.4-3.6GHz | 16GB | 512GB SSD | ~$160-220 | ✅ Complete, includes case/PSU |
| **SeeedStudio Odyssey X86** | Quad 2.5GHz (J4105) | 8GB | 64GB eMMC | ~$150-180 | Older gen, but proven |

---

## x86 Options — Detailed Comparison

### Option 1: LattePanda IOTA (DFRobot) — SBC Form Factor
**Vendor:** DFRobot (one of your preferred vendors)
**Price:** ~$129-179/board (8GB/64GB to 16GB/128GB)

**Per-Unit Cost (22 units):**
| Component | Unit Price | Qty | Total |
|-----------|------------|-----|-------|
| IOTA 8GB/64GB | $129 | 22 | $2,838 |
| Active Cooler | $5 | 22 | $110 |
| WiFi 6E Module (AX210) | $20 | 22 | $440 |
| Case + PSU | $25 | 22 | $550 |
| **Total** | **~$179/unit** | | **~$3,938** |

**Pros:**
- DFRobot = preferred vendor, educational support
- SBC form factor (familiar to you)
- RP2040 co-processor (not needed, but available)
- Ubuntu/Windows support
- GPIO via RP2040 (not needed for your use)

**Cons:**
- Requires accessories (cooler, WiFi, case, PSU)
- Final cost ~$179/unit

---

### Option 2: Orange Pi X — Budget x86
**Vendor:** Orange Pi (new 2025 release)
**Price:** ~$100-130/board (estimated)

**Per-Unit Cost (22 units):**
| Component | Unit Price | Qty | Total |
|-----------|------------|-----|-------|
| Orange Pi X 8GB | ~$100 | 22 | ~$2,200 |
| Cooler + Case | ~$20 | 22 | ~$440 |
| WiFi Module | ~$15 | 22 | ~$330 |
| PSU | ~$10 | 22 | ~$220 |
| **Total** | **~$145/unit** | | **~$3,190** |

**Pros:**
- Cheapest x86 option
- Intel N100/N150 (same CPU as IOTA)
- New 2025 release

**Cons:**
- New product = unproven, limited docs
- Requires accessories (like IOTA)
- Less educational support than DFRobot
- Availability uncertain

---

### Option 3: Intel N100/N150 Mini PCs — Complete Solution
**Vendors:** GMKtec, Beelink, Acemagic (Amazon, Newegg)
**Price:** ~$160-220/unit (complete, includes everything)

**Examples (2025-2026 pricing):**
- GMKtec N150, 16GB RAM, 512GB SSD: ~$180-200
- Beelink S13 (N150), 16GB, 512GB SSD: ~$200-220 (was $169 Aug 2025)
- Acemagic V1 (N150), 16GB, 512GB SSD: ~$180-200

**Per-Unit Cost (22 units):**
| Component | Unit Price | Qty | Total |
|-----------|------------|-----|-------|
| Mini PC (N150, 16GB, 512GB SSD) | ~$180 | 22 | ~$3,960 |
| **Total** | **~$180/unit** | | **~$3,960** |

**What's Included:**
- ✅ Case
- ✅ Power supply
- ✅ WiFi 6 + Bluetooth
- ✅ Gigabit Ethernet
- ✅ HDMI/DP outputs
- ✅ Multiple USB ports
- ✅ 512GB NVMe SSD (5-8x faster than eMMC)
- ✅ 16GB RAM (vs. 8GB on IOTA base model)

**Pros:**
- **Complete solution** — no accessory shopping
- **16GB RAM + 512GB SSD** — better specs than IOTA base
- **Proven products** — GMKtec, Beelink have reviews
- **Same N150 CPU** as IOTA
- **Arduino IDE works officially** (x86)
- **Easier replacement** — if one dies, swap with any mini PC

**Cons:**
- Larger than SBC (but still compact)
- Not from DFRobot/Seeed/M5Stack (but Amazon/Newegg return policies are good)
- Less "educational" branding (but same hardware)

---

### Option 4: SeeedStudio Odyssey X86 — Older but Proven
**Vendor:** SeeedStudio (one of your preferred vendors)
**Price:** ~$150-180/board

**Specs:**
- CPU: Intel Celeron J4105 (quad 2.0-2.5GHz) — older, slower than N100/N150
- RAM: 8GB
- Storage: 64GB eMMC
- Dual Gigabit Ethernet
- WiFi + Bluetooth

**Per-Unit Cost (22 units):**
| Component | Unit Price | Qty | Total |
|-----------|------------|-----|-------|
| Odyssey X86 (J4105, 8GB, 64GB) | ~$160 | 22 | ~$3,520 |
| Case + PSU | ~$20 | 22 | ~$440 |
| **Total** | **~$180/unit** | | **~$3,960** |

**Pros:**
- SeeedStudio = preferred vendor
- Proven product (been around since ~2020)
- Dual Ethernet (great for MQTT/network lessons)

**Cons:**
- **Older CPU** (J4105 vs. N150) — ~30-40% slower
- Only 64GB storage (vs. 512GB on mini PCs)
- Similar price to mini PCs but worse specs

---

## Recommendation: LattePanda IOTA via Mouser

**Selected Path:** LattePanda IOTA (DFRobot) through Mouser.com

**Why This Wins:**
1. **Mouser = Approved vendor** — You already order DFRobot/SeeedStudio from here
2. **Educational product** — SBC looks like "educational equipment" not "consumer desktop"
3. **IT scrutiny = Minimal** — DFRobot is known in education, Mouser is industrial/educational supplier
4. **x86 = Arduino IDE works** — Official support, auto-updates, ESP32/ESP8266 current
5. **N150 = Same CPU** as consumer mini PCs (3.6GHz quad)
6. **No justification needed** — Unlike "why are we buying 22 mini PCs from CDW?"

**Mini PCs via CDW = Backup Option** (if IOTA unavailable or budget issues)

---

## Deployment Plan (IOTA via Mouser)

### Phase 1: Test One Unit (~1-2 weeks)
- [ ] Order 1x LattePanda IOTA (8GB/64GB) + cooler + WiFi module
- [ ] Install Ubuntu 22.04 or 24.04
- [ ] Install Arduino IDE 2.x (verify it's current, not 2-year-old)
- [ ] Test ESP32 compile times (target: <1 min first, <30 sec subsequent)
- [ ] Test ESP8266 compile times
- [ ] Test Node-RED + browser multitasking
- [ ] Verify CH340/CP2102 USB drivers work (for student Arduino boards)
- [ ] Document setup process (for imaging 22 units)

### Phase 2: Order 22 Units (~2-4 weeks)
- [ ] Contact Mouser/DFRobot — Educational bulk pricing for 22 units?
- [ ] Order: 22x IOTA (8GB/64GB), 22x coolers, 22x WiFi modules
- [ ] Order: Cases + PSUs (or use existing Pi 4 supplies if compatible)
- [ ] Order: 2-3 spare units (classroom attrition)

### Phase 3: Image & Deploy (~1-2 weeks)
- [ ] Create master Ubuntu image (Arduino IDE, Node-RED, browser, MQTT tools)
- [ ] Clone to all 22 units
- [ ] Set up auto-login (classroom workflow)
- [ ] Configure network (Ethernet preferred for MQTT stability)
- [ ] Test all 22 units with student Arduino boards
- [ ] Document recovery procedure (SD card corruption = reflash)

### Phase 4: Training (~1 week)
- [ ] Update lesson plans (any IOTA-specific steps?)
- [ ] Test self-scoring robot competition setup
- [ ] Verify MQTT network handles 22 nodes
- [ ] Student orientation (new hardware)

---

## Budget Estimate (22 Units)

| Component | Unit Price | Qty | Total |
|-----------|------------|-----|-------|
| LattePanda IOTA (8GB/64GB) | ~$129 | 22 | $2,838 |
| Active Cooler | ~$5 | 22 | $110 |
| WiFi 6E Module (AX210) | ~$20 | 22 | $440 |
| Case + PSU | ~$25 | 22 | $550 |
| Spare Units (2x IOTA + accessories) | ~$180 | 2 | $360 |
| **Total** | **~$179/unit** | **24** | **~$4,298** |

**Vs. Arduino Cloud Licensing:**
- Cloud: ~$800/year × 5 years = ~$4,000
- IOTA: ~$4,298 one-time
- **Break-even: ~5 years** (but students get current Arduino IDE forever)

**Potential Savings:**
- Educational bulk discount from DFRobot/Mouser?
- Reuse existing Pi 4 cases/PSUs (if compatible)?
- Skip WiFi modules (use Ethernet for classroom)?

---

## IT Conversation Prep

**If IT asks:**

| Question | Answer |
|----------|--------|
| "What are these?" | Single-board computers for robotics/AIoT education |
| "Why not use existing Pis?" | Arduino IDE on ARM is 2 years old, ESP32 support broken |
| "Why x86?" | Arduino.cc officially supports x86, students get current toolchains |
| "Why DFRobot?" | Educational vendor, Mouser is approved supplier |
| "What OS?" | Ubuntu Linux (open-source, no licensing costs) |
| "Network security?" | Ethernet to classroom VLAN, MQTT local only (no cloud) |
| "Student data?" | No cloud accounts, all local storage |

**Key Talking Points:**
- Educational purpose (robotics, IoT, programming)
- Approved vendor (Mouser)
- Open-source software (Ubuntu, Arduino IDE, Node-RED)
- No cloud dependencies (local MQTT, local compilation)
- Replaces aging Pi 4s (same use case, better performance)

---

## Risk Mitigation

| Risk | Mitigation |
|------|------------|
| IOTA unavailable on Mouser | Backup: CDW business mini PCs |
| Ubuntu driver issues | Test one unit first, document fixes |
| Student breaks unit | 2-3 spares on hand |
| Compile times still slow | SSD upgrade option (M.2 NVMe slot) |
| WiFi unreliable | Use Ethernet (all IOTAs have Gigabit) |

---

## For your classroom, LattePanda IOTA via Mouser is the right choice:

| Factor | Mini PC (N150) | LattePanda IOTA | Orange Pi X | Pi 5 + SSD |
|--------|----------------|-----------------|-------------|------------|
| **CPU** | N150 (3.6GHz) | N150 (3.6GHz) | N150 (3.6GHz) | ARM (2.4GHz) |
| **Arduino IDE** | ✅ Official | ✅ Official | ✅ Official | ⚠️ 2 years old |
| **RAM** | 16GB | 8GB | 8GB | 8GB |
| **Storage** | 512GB SSD | 64GB eMMC | 64GB eMMC | 128GB SSD |
| **Includes** | Everything | Board only | Board only | Board only |
| **Accessories** | None needed | ~$50/unit | ~$45/unit | ~$30/unit |
| **Total (22 units)** | ~$3,960 | ~$3,938 | ~$3,190 | ~$2,860 |
| **Vendor** | Amazon/Newegg | DFRobot | Orange Pi | Pi Foundation |

**Why Mini PCs Win:**
1. **Complete solution** — no accessory hunting
2. **16GB RAM + 512GB SSD** — better than IOTA/Orange Pi
3. **Same N150 CPU** as IOTA (Arduino works officially)
4. **~$3,960 total** — same as IOTA, but includes everything
5. **Easier to replace** — if one dies, any N100/N150 mini PC works

**The Only Reason to Choose IOTA:**
- You want DFRobot specifically (educational support, relationship)
- You want SBC form factor (GPIO availability, even if not used)
- You want RP2040 co-processor (not needed for your use)

---

## Watch List — Interesting But Not (Yet) Recommended

### Orange Pi AI Station
**Added:** 2026-03-30
**URL:** http://www.orangepi.org/html/hardWare/computerAndMicrocontrollers/details/Orange-Pi-AI-Station.html

**Specs:**
- **CPU:** Ascend 310 series, 16 CPU cores @ 1.9GHz (ARM)
- **AI:** 10 AI cores + 8 Vector cores @ 1.08-1GHz
- **Performance:** **176 TOPS** (massive AI inference power)
- **RAM:** 48GB or 96GB LPDDR4X @ 4266MHz
- **Storage:** NVMe SSD M.2 2280 slot
- **I/O:** USB3.0, HDMI, Gigabit Ethernet, TF slot, 40-pin GPIO
- **Use Cases:** Smart security, transportation, healthcare, industrial automation, embodied intelligence

**Why It's On The Watch List (Not Recommended):**
- ❌ **ARM architecture** — same Arduino IDE 2-year-old problem
- ❌ **Ascend processor** — Huawei AI chip, niche ecosystem
- ❌ **Overkill for classroom** — 176 TOPS is for industrial AI inference
- ❌ **Price unknown** — likely $300-500+ (industrial pricing)
- ❌ **Linux support** — Ascend requires special drivers/toolchains

**Why It's Interesting:**
- ✅ 176 TOPS = serious AI/ML robotics projects (YOLO, SLAM, etc.)
- ✅ 48-96GB RAM = massive multitasking
- ✅ NVMe SSD = fast storage
- ✅ 40-pin GPIO = robotics/sensor compatibility
- ✅ Could run local LLMs, advanced vision, etc.

**Best Use Case:**
- Advanced robotics capstone projects
- AI/ML research station (not general classroom)
- Teacher demo station for computer vision lessons
- **Not** for intro E2/Robotics classes (Arduino IDE won't work properly)

**Status:** Watch for pricing + Arduino compatibility. Might be worth 1-2 units for advanced projects.

---

## Purchasing Constraints

**School System Approved Vendors:**
- ✅ **Mouser.com** — Approved industrial supplier
- ❌ **Amazon** — NOT allowed for school purchases
- ❌ **Newegg** — Likely not approved (need to verify)
- ? **DFRobot Direct** — Need to verify (educational vendor)
- ? **SeeedStudio** — Need to verify (educational vendor)
- ? **CDW / Insight / PCM** — Common school suppliers (need to verify)

**Implication:** Consumer mini PCs (GMKtec, Beelink, Acemagic) are NOT purchasable through Mouser.

**Mouser-Compatible Options:**
- ✅ **LattePanda IOTA** (DFRobot) — Mouser carries DFRobot products
- ✅ **SeeedStudio Products** — Odyssey X86, sensors, accessories
- ✅ **Industrial SBCs** (IEI, Advantech, etc.) — Expensive, overkill
- ⚠️ **Orange Pi X** — Unlikely on Mouser (consumer brand)

**CDW-Compatible Options:**
- ✅ **Business Mini PCs** — Dell OptiPlex Micro, HP ProMini/EliteDesk, Lenovo ThinkCentre Tiny
- ✅ **Educational SBCs** — May carry Raspberry Pi, possibly LattePanda
- ✅ **Volume pricing** — CDW specializes in bulk education sales

**Strategy:**
1. **Mouser** → LattePanda IOTA (DFRobot) + SeeedStudio accessories
2. **CDW** → Check for N100/N150 mini PCs (business brands: Dell, HP, Lenovo)
3. **Compare pricing** — Mouser (IOTA) vs. CDW (business mini PCs)

---

## Next Steps

1. **Verify school vendor list** — What suppliers ARE approved besides Mouser?
2. **Contact DFRobot** — Educational pricing + school purchase order for 22x IOTA?
3. **Contact SeeedStudio** — Educational pricing for Odyssey or alternatives?
4. **Check CDW / Insight / PCM** — Do they carry N100/N150 mini PCs?
5. **Test one unit** — Whatever vendor works, test Arduino IDE + ESP32 compiles first
6. **Verify Ubuntu imaging** — How to deploy 22 units efficiently?

---

## Mouser Research Notes

**What Mouser Carries (Relevant to This Project):**

| Category | Examples | Price Range | Notes |
|----------|----------|-------------|-------|
| **LattePanda (DFRobot)** | Older LattePanda models | ~$150-200 | IOTA may be available |
| **Industrial SBCs** | IEI, Advantech, WinSystems | $300-800+ | Overkill, expensive |
| **Embedded Boards** | Intel Atom, Celeron based | $200-500 | Older CPUs, industrial pricing |

**What Mouser DOESN'T Carry:**
- GMKtec, Beelink, Acemagic (consumer mini PC brands)
- Orange Pi X (consumer brand)
- Most N100/N150 mini PCs (too new, consumer market)

**Strategy:**
- **LattePanda IOTA via DFRobot** — Best Mouser-compatible option
- **Alternative vendors** — Need to find school-approved suppliers for mini PCs
- **Industrial SBCs** — Only if budget allows ($6,600-17,600 for 22 units)


**Want me to:**
- Find specific mini PC models with best reviews?
- Research Amazon Business/educational pricing?
- Look into disk imaging for 22 units?
- Compare warranty/support options?

---

### LattePanda IOTA (x86 — Native Arduino Support)

**Added:** 2026-03-30
**Price:** ~$129-179/unit (8GB/64GB to 16GB/128GB with Win11)
**URL:** https://www.dfrobot.com/kit-005.html

**Specs:**
- **CPU:** Intel N150 (Twin Lake), quad-core up to 3.6GHz (**x86**)
- **GPU:** Intel UHD Graphics (32 EUs)
- **RAM:** 8GB or 16GB LPDDR5
- **Storage:** 64GB or 128GB eMMC (expandable via M.2 NVMe)
- **Co-processor:** RP2040 — *not relevant for this use case*
- **I/O:** 3x USB 3.2 Gen2, Gigabit Ethernet, HDMI 2.1, eDP
- **Wireless:** Intel AX210 (WiFi 6E, BT 5.2) or BE200 (WiFi 7)
- **Size:** 88×70mm (palm-sized)
- **TDP:** 6W (fanless) to 15W (active cooler)
- **OS:** Linux (Ubuntu/Debian) or Windows 11 IoT Enterprise

**The Arduino on ARM Problem (Your Current Pi 4s, Pi 5, Orange Pi, RDK X5):**
- Arduino 2.0+ on ARM Linux is a **community port** (github.com/koendv/arduino-ide-raspberrypi)
- **Not officially supported by Arduino.cc**
- **Current version is 2+ years old** (as of 2026-03-30)
- Students get **stale ESP32/ESP8266 board packages** (missing features, bug fixes, security patches)
- Toolchain quirks (especially ESP32/ESP8266 board support)
- **You're stuck waiting for ARM builds** that may never come
- **PlatformIO alternative = rejected** (too complex for intro classes)

**The Arduino on x86 Reality (LattePanda IOTA):**
- **Official support** from Arduino.cc
- **Automatic updates** via Arduino IDE (just like Windows/Mac)
- **Native toolchains** (avr-gcc, xtensa-esp32-elf, etc.) — optimized by Espressif
- **No waiting** for ARM compatibility layers
- **ESP32/ESP8266 board managers** update automatically

**Why x86 Matters for Your Lab:**
1. **Arduino IDE works as intended** — not a hacky port
2. **ESP32 compiles are fast** — 3.6GHz x86 vs. 1.5GHz ARM
3. **No toolchain drama** — CH340/CP2102 drivers work on Linux out of box
4. **Future-proof** — Arduino.cc develops for x86 first (Windows/Mac/Linux)

**Potential Downsides:**
- **Price:** ~$129-179/unit = ~$2,838-3,938 for 22 units
- **Cooling required:** Heatsink + fan (~$5-10/unit)
- **WiFi module required:** AX210 or BE200 (~$15-25/unit)
- **Total per unit:** ~$150-210 (with cooler + WiFi)
- **Linux docs:** Less polished than Raspberry Pi (but works fine)
- **GPIO:** Not needed (you use external Arduino boards via USB)

**Expected ESP32 Compile Performance:**
- Intel N150 (3.6GHz) vs. Pi 4 (1.5GHz) = ~2.4x single-core speed
- x86 toolchain optimizations = additional boost
- **Estimate:** ~30-45 sec first compile, ~15-25 sec subsequent (vs. 2-3 min on Pi 4)

---

## Recommended Configurations

### Option A: Pi 5 + SSD (Safe Choice)
| Component | Qty | Unit Price | Total |
|-----------|-----|------------|-------|
| Pi 5 (8GB) | 22 | $80 | $1,760 |
| Pi 5 Active Cooler | 22 | $5 | $110 |
| USB 3.0 SSD (128GB) | 22 | $25 | $550 |
| Case + PSU | 22 | $20 | $440 |
| **Total** | | | **~$2,860** |

**Pros:** Best support, docs, Arduino compatibility, ecosystem
**Cons:** More expensive than Orange Pi, SSD is external

**Expected Performance:**
- ESP32 compile: ~45-60 sec first, ~20-30 sec subsequent
- Browser + Arduino: Smooth
- Storage: 5-10x faster than SD

---

### Option B: Orange Pi 5 + eMMC (Performance/Value)
| Component | Qty | Unit Price | Total |
|-----------|-----|------------|-------|
| Orange Pi 5 (8GB) | 22 | $65 | $1,430 |
| eMMC Module (64GB) | 22 | $15 | $330 |
| Case + PSU | 22 | $20 | $440 |
| **Total** | | | **~$2,200** |

**Pros:** 4x CPU cores, eMMC built-in, $660 savings vs. Pi 5
**Cons:** Less polished, Arduino USB driver quirks possible, smaller community

**Expected Performance:**
- ESP32 compile: ~30-45 sec first, ~15-25 sec subsequent
- Browser + Arduino: Very smooth
- Storage: 3-5x faster than SD (eMMC)

---

### Option C: Hybrid (Pi 5 + Orange Pi 5)
- **5x Pi 5 (8GB) + SSD** for teacher demo + advanced stations
- **17x Orange Pi 5 (8GB) + eMMC** for student desks
- **Total:** ~$2,400

**Pros:** Best of both, teacher station has maximum reliability
**Cons:** Two images to maintain, different hardware

---

## Key Decision Factors

### 1. Arduino USB Driver Compatibility
- **Pi 5:** All Arduino boards work out of box (CH340, CP2102, ATmega16U2)
- **Orange Pi 5:** May need manual driver installs for some ESP boards
- **Risk:** Orange Pi = potential classroom time lost to driver issues

### 2. Storage Reliability
- **SD cards:** Cheap, but corrupt easily (student abuse)
- **eMMC:** Built-in, more reliable, moderate speed
- **USB SSD:** Fastest, most reliable, external (cable management)

### 3. Classroom Management
- **Disk imaging:** Need to clone 22 units
  - Pi: `raspi-clone`, `Etcher`, network boot options
  - Orange Pi: `dd`, `Etcher`, less tooling
- **Student recovery:** Boot recovery partition? Network boot fallback?
- **Configuration management:** Ansible? Simple scripts?

### 4. MQTT Network Load
- 22 nodes publishing/subscribing simultaneously
- Need good network stack + Ethernet preferred over WiFi
- Pi 5 + Orange Pi 5 both handle this fine

---

## Research Questions

1. **Orange Pi 5 Arduino USB drivers:** Do CH340/CP2102 work on Armbian out of box?
2. **eMMC vs. SSD:** Is eMMC reliable enough for classroom, or go full SSD?
3. **Bulk pricing:** Educational discounts from Pi Foundation, DFRobot, etc.?
4. **Cases:** Bulk classroom cases with cooling?
5. **Imaging:** What's the workflow for 22 units?

---

## Next Steps

- [ ] **Test Pi 5 + SSD** — borrow one unit, measure ESP32 compile times
- [ ] **Test Orange Pi 5** — borrow one, verify Arduino USB drivers work
- [ ] **Get bulk quotes** — Pi Foundation education, DFRobot, Amazon Business
- [ ] ** Decide: SSD vs. eMMC** — reliability vs. cost
- [ ] **Plan imaging workflow** — how to deploy 22 units efficiently
- [ ] **Check existing SD cards** — are they the bottleneck? (Test before replacing all)

---

## Related

- Self-scoring robot competition specs (what sensors/scoring?)
- MQTT broker setup (Mosquitto on Pi? Cloud?)
- Student recovery procedures (what happens when they break it?)

---

## Changelog

| Date | Update |
|------|--------|
| 2026-03-30 | Created doc, identified x86 requirement (Arduino IDE 2 years old on ARM) |
| 2026-03-30 | Added LattePanda IOTA, Orange Pi X, mini PCs as candidates |
| 2026-03-30 | Confirmed Mouser + CDW are approved vendors (Amazon not allowed) |
| 2026-03-30 | Selected LattePanda IOTA via Mouser (minimal IT scrutiny) |
| 2026-03-30 | Added deployment plan + budget (~$4,300 for 22 units) |
| 2026-03-30 | Added Orange Pi AI Station to Watch List (176 TOPS, overkill for classroom) |

---

*Created: 2026-03-30*
*Next Review: Before next school year budget cycle (Summer 2026)*
*Status: Research Phase — Purchase planned for next school year*

```

## Notes
<!-- openclaw:human:start -->
<!-- openclaw:human:end -->

## Related
<!-- openclaw:wiki:related:start -->
### Referenced By

- [README](sources/readme.md)
<!-- openclaw:wiki:related:end -->