Signal Integrity Basics for Mining Hardware Repair

Introduction

Every ASIC chip on a hash board communicates with its neighbors through electrical signals traveling along copper traces. When those signals arrive clean and on time, the board hashes at full speed. When they don't, you get missing chips, low hashrate, or intermittent failures that seem impossible to pin down.

Signal integrity (SI) is the discipline of ensuring electrical signals arrive at their destination without corruption. For a repair technician, you don't need to become an RF engineer — but understanding the basics will transform the way you diagnose and fix hash boards. Instead of blindly replacing chips, you'll be able to look at a signal on an oscilloscope and know whether the problem is a cracked trace, a bad solder joint, or a failing chip.

What Is Signal Integrity?

Digital Signals Are Not Ideal Square Waves

In textbooks, digital signals are drawn as perfect rectangles — instant transitions between 0V and 1.8V (or whatever the logic level is). In reality, every signal on a hash board is an analog waveform shaped by the physical properties of the PCB, the components, and the environment.

A real digital signal has:

• Rise time — the time it takes to transition from low to high (typically 1–5 nanoseconds on modern ASIC chips)
• Fall time — the time to transition from high to low
• Overshoot — the signal momentarily exceeds the target voltage before settling
• Undershoot — the signal dips below ground during transitions
• Ringing — the signal oscillates around the target voltage before stabilizing

These imperfections are normal in small amounts. Problems arise when they become severe enough that the receiving chip can't reliably distinguish a 1 from a 0.

Why It Matters for Mining Hardware

A hash board like the Antminer S19 XP has 132 ASIC chips connected in a daisy chain. A signal might pass through dozens of chip-to-chip links, each one a potential point of degradation. At the communication speeds these boards use (typically 1.5–6 MHz clock), signal integrity issues compound across the chain.

PCB Design Basics for Hash Boards

Understanding the physical structure of a hash board PCB helps you make sense of what you see under the microscope and on the oscilloscope.

Multi-Layer PCB Construction

Hash board PCBs are multi-layer boards, typically 4 to 6 layers. Each layer serves a specific purpose:

The ground plane is one of the most important layers. It provides a low-impedance return path for every signal on the board. When the ground plane is damaged (cracked, corroded, or has missing copper), signal integrity degrades across a wide area — not just at the damage point.

Ground Planes and Power Planes

Every signal on the board needs a return path — a way for current to flow back to its source. On a well-designed hash board, the return current flows through the ground plane directly beneath the signal trace.

Think of it like a highway: the signal trace is the road, and the ground plane is the shoulder. If the shoulder is missing, traffic (current) has to take a detour, and things get messy.

Power planes distribute voltage to the ASIC chips. Hash boards divide chips into voltage domains, each fed by its own regulator. The power plane for each domain must deliver stable, clean voltage to every chip in that group. Noise on the power plane directly couples into the signals — more on this below.

Trace Routing for High-Speed Signals

The CI/CO (Clock In/Out) and RI/RO (Receive In/Out) signals are the communication backbone of the hash board. These traces are routed with specific design rules:

• Controlled impedance — trace width and spacing are calculated to maintain a consistent impedance (typically 50 ohms) along the entire length
• Length matching — clock and data traces between the same two chips are kept the same length so signals arrive simultaneously
• Minimized stubs — traces go directly from chip to chip without unnecessary branches
• Adequate spacing — signal traces are kept far enough apart to prevent crosstalk

Chip 0              Chip 1              Chip 2
┌─────┐            ┌─────┐            ┌─────┐
│  CO ├────────────┤ CI  │            │     │
│  RO ├────────────┤ RI  │            │     │
│     │            │  CO ├────────────┤ CI  │
│     │            │  RO ├────────────┤ RI  │
└─────┘            └─────┘            └─────┘
  ← matched length →  ← matched length →

Via Transitions Between Layers

A via is a small plated hole that connects traces between PCB layers. On hash boards, vias are used extensively — to connect surface-mounted chips to inner-layer power and ground planes, and to route signals between the top and bottom of the board.

Vias introduce a small impedance discontinuity. In a well-designed board, this is negligible. But when a via fails (due to thermal stress, corrosion, or manufacturing defects), it can create an open circuit or a high-resistance connection that degrades signal quality.

Common Signal Integrity Issues in Hash Boards

Crosstalk

Crosstalk occurs when a signal on one trace induces a voltage on an adjacent trace through electromagnetic coupling. On hash boards, where dozens of signal traces run in parallel through dense chip arrays, crosstalk is a constant design challenge.

There are two types:

• Near-end crosstalk (NEXT) — noise appears at the same end as the aggressor signal's source
• Far-end crosstalk (FEXT) — noise appears at the far end of the victim trace

In practice, crosstalk on hash boards is most problematic when:

• A repair has added jumper wires running parallel to existing signal traces
• Corrosion has reduced the spacing between adjacent traces
• A solder bridge creates a partial short between two signal traces

Impedance Mismatch

Every trace on a PCB has a characteristic impedance determined by its width, the distance to the ground plane, and the dielectric material between them. When a signal encounters a change in impedance — at a connector, a damaged trace, a via, or a poorly soldered joint — part of the signal energy reflects back toward the source.

These reflections cause:

• Overshoot and undershoot at the receiver
• Double edges that the receiving chip may interpret as two clock pulses instead of one
• Reduced noise margin making the signal vulnerable to other interference

On hash boards, impedance mismatches are most commonly caused by:

• Cracked or partially broken traces (the narrow section has higher impedance)
• Solder bridges that change the effective trace geometry
• Repair jumper wires with different impedance than the original trace
• Corroded or oxidized connectors

Ground Bounce

Ground bounce (also called simultaneous switching noise or SSN) happens when many chips switch their outputs at the same time. Each switching event draws a burst of current through the ground connection, and the inductance of the ground path causes a momentary voltage spike on the ground rail.

On a hash board with 100+ ASIC chips, each containing hundreds of hashing cores switching simultaneously, ground bounce is significant. It effectively raises the "floor" of the ground voltage momentarily, reducing the noise margin for all signals on the board.

Ground bounce is worse when:

• The ground plane has been compromised (cracks, corrosion, missing copper)
• Decoupling capacitors near the chips are missing, cracked, or have poor solder joints
• The board is operating at maximum frequency/voltage

Power Supply Noise

ASIC chips draw varying amounts of current depending on the computations they're performing. This creates ripple on the power rail — small, rapid voltage fluctuations that can couple into nearby signal traces.

On hash boards, power supply noise is managed through:

• Decoupling capacitors (the small capacitors clustered around each chip)
• Bulk capacitors (larger caps near the voltage regulators)
• Proper power plane design

When decoupling capacitors fail (cracked from thermal stress, knocked off during handling), the local power supply noise increases dramatically. This noise couples into the CI/CO and RI/RO signals, degrading communication quality.

How Signal Issues Manifest in Mining

Understanding the symptoms helps you work backward to the root cause.

Test Points and Measurement Techniques

Identifying Test Points on Hash Board PCBs

Most hash boards include dedicated test points — exposed copper pads designed for probing during manufacturing and repair. They are typically:

• Small circular or square pads labeled TP1, TP2, etc.
• Located near voltage regulators (for power rail measurement)
• Placed at the beginning, middle, and end of the daisy chain (for signal measurement)
• Found on the bottom side of the board near chip signal pins

Oscilloscope Probing Techniques

An oscilloscope is the primary tool for measuring signal integrity. For hash board work, you need at minimum a 100 MHz bandwidth scope with at least 1 GSa/s sample rate.

Healthy vs Degraded Signals

Voltage
1.8V  ┌──────┐      ┌──────┐      ┌──────┐
      │      │      │      │      │      │
      │      │      │      │      │      │
0.0V ─┘      └──────┘      └──────┘      └──

Voltage
2.1V  ╱╲
1.8V  ┌──╲╱──┐      ┌───╲╱─┐
      │       │      │       │
      │       │      │       │
0.0V ─┘       └─╲╱──┘       └─╲╱──
-0.3V           ╲╱              ╲╱

Voltage
1.8V

1.0V  ┌──╲╱╲╱╲╱──
      │
0.0V ─┘

Using a Logic Analyzer

While an oscilloscope shows the analog quality of signals, a logic analyzer captures the digital data being transmitted. This is useful for:

• Verifying that the control board is sending valid commands
• Checking that chips are responding to enumeration
• Identifying which specific chip in the chain stops responding
• Capturing job data to verify correctness

Connect the logic analyzer to the CI and RI lines at the start of the chain, and optionally at the suspected failure point. Compare the data patterns to known-good captures.

Practical Repair Implications

This is where theory meets the soldering iron. Understanding signal integrity directly improves your repair outcomes.

Cracked Traces

A cracked trace is the most common signal integrity failure on hash boards. Causes include:

• Flexing the board during handling or heatsink removal
• Thermal cycling stress over months of operation
• Manufacturing defects (hairline cracks that worsen over time)
• Physical impact damage

How to find them: Follow the signal trace from the last responding chip to the first missing chip. Use a magnifying glass or microscope. Hairline cracks may be invisible to the naked eye — use a multimeter in continuity mode to test the trace end-to-end.

How to repair: Scrape away the solder mask on both sides of the crack to expose clean copper. Apply flux and bridge the crack with solder. For longer breaks, use a thin wire (30 AWG magnet wire) soldered to exposed copper on both sides.

Solder Bridges

A solder bridge is an unintended connection between two adjacent pads or traces. On hash boards with fine-pitch chip packages, solder bridges can:

• Short CI to CO (clock loop-back, confuses downstream chips)
• Short a signal to ground (kills the signal entirely)
• Short two data lines together (corrupts data, causes hash errors)

How to find them: Visual inspection under magnification. Look for shiny connections between pads that should be separate. On QFN packages (common for ASIC chips), inspect the edges carefully.

How to repair: Apply flux to the bridge area. Use a clean soldering iron tip to wick away excess solder. For stubborn bridges, use solder wick (desoldering braid). Verify separation with a multimeter.

Component Tombstoning

Tombstoning occurs when a small surface-mount component (typically a capacitor or resistor) lifts off one pad during soldering, standing upright like a tombstone. This creates an open circuit on one side.

When a decoupling capacitor tombstones, it no longer filters power supply noise for its nearby chip. This increases noise on the local power rail, which couples into signal traces and degrades signal quality.

How to find: Visual inspection. Tombstoned components are usually obvious — they stand at an angle or vertically instead of lying flat.

How to repair: Apply flux to both pads. Reflow the component with a hot air station. If the component is damaged, replace it with the same value (check the schematic or compare to the same position near an adjacent chip).

Via Failures

Via failures occur when the plated copper inside a via hole degrades, creating a high-resistance connection or open circuit between layers. This is especially problematic under BGA chips where vias carry critical signals from the chip pads to inner layers.

How to find: Via failures are difficult to detect visually. Suspect a via failure when:

• A chip area shows signal issues but all visible traces and solder joints look good
• The problem appeared after thermal rework (BGA reflow can stress vias)
• Continuity testing shows an intermittent connection when you apply pressure to the board

How to repair: For accessible vias, drill out the failed via and re-plate, or run a jumper wire from the top pad to the bottom pad. For vias under BGA packages, the chip must be removed first — at which point you can inspect and repair the vias before reballing and reflowing the chip.

Proper Repair Techniques to Maintain Signal Integrity

Key Takeaways

1. Digital signals are analog in reality — rise time, overshoot, and ringing affect whether the receiving chip can read the data correctly.

2. Hash boards are multi-layer PCBs with dedicated ground and power planes. Damage to these internal layers can cause widespread signal problems that are hard to localize.

3. The four main signal integrity problems on hash boards are crosstalk, impedance mismatch, ground bounce, and power supply noise. All of them reduce the noise margin available for reliable communication.

4. Symptoms map to signal issues: missing chips mean a broken signal chain; low hashrate means degraded but functional signals; intermittent failures suggest temperature-dependent marginal signals; hash errors indicate data corruption.

5. An oscilloscope is your most important diagnostic tool for signal issues. Learn to recognize healthy, degraded, and failed signals.

6. Repair technique matters for signal integrity — short jumper wires, proper routing, and replacing damaged decoupling capacitors can mean the difference between a repair that works and one that introduces new problems.

Apply This Knowledge

Now that you understand signal integrity basics, put this knowledge to work:

• Practice oscilloscope probing on a known-good hash board. Measure CI/CO at several points along the chain to see what healthy signals look like on your specific scope.
• Inspect a failed board under magnification. Follow the signal traces from the last good chip to the first missing chip, looking for cracks, bridges, or damaged components.
• Build a reference library of oscilloscope captures showing good and bad signals for each board model you commonly repair.