Section Overview
A short reads near zero everywhere along its rail, so ordinary probing cannot say where it is — the meter reads zero at the fault and zero a hand's-breadth away — and the two workhorse methods here break that deadlock by measuring what does change across a shorted plane: tiny resistance and voltage (confirming-and-characterizing-a-short). That is the theme. The resistance from a point to the short is not quite zero, and it falls as you approach. A resistance gradient is that gentle downhill in resistance — the resistance from a probe point to the short decreases, milliohm by milliohm, as the probe moves closer — so following the falling resistance leads to the fault, a method that works unpowered with a low-ohms, four-wire measurement. The voltage method needs a current to make a gradient. Current injection is deliberately driving a known current into the shorted rail from a current-limited supply, so that the plane, carrying that current through its own copper, develops a minute voltage difference along it (current-measurement-and-in-circuit-limits). That injected current creates a voltage that slopes to the short. A voltage gradient is the tiny voltage the injected current develops along the shorted copper — highest where the current enters, sloping down toward the short, which is where the current leaves — so probing along it and following the voltage downhill converges on the fault. Both methods hunt vanishingly small quantities. Milliohms and microvolts are below what an ordinary meter resolves, so both methods depend on four-wire connections, sensitive meters, and steady technique to see the gradient at all (confirming-and-characterizing-a-short). And both converge to a point. Following either gradient downhill, narrowing each step, converges on the shorted point, which is then confirmed directly. Master these, and a short that reads zero everywhere still tells you exactly where it is.
Why This Matters
Localizing a hard short is one of the genuinely hard problems of repair, because the short defeats the obvious measurement — it reads zero everywhere — so the gradient methods that measure the tiny remaining differences are what make it findable at all (confirming-and-characterizing-a-short). This matters because the short hides in plain sight: a rail shorted to ground reads the same near-zero at every one of its dozens of nodes, so without a method that resolves the tiny differences, the fault is a needle in a haystack that all reads zero. This matters because resistance still slopes to the fault: the copper between a probe and the short has a real, tiny resistance that falls as you approach, so following the resistance gradient is a direct route to the short with only a meter (confirming-and-characterizing-a-short). It matters because an injected current makes the invisible visible: driving a known current through the plane develops a voltage gradient that a sensitive meter can follow downhill, turning a dead-flat rail into a sloped map pointing at the fault (current-measurement-and-in-circuit-limits). It matters because resolution is everything: these gradients are microvolts and milliohms, so the methods succeed only with four-wire connections and sensitive meters, and fail with sloppy technique — knowing that is the difference between finding the short and giving up. And it matters because the methods converge, not guess: each measurement narrows the search toward the point, so the hunt is a directed convergence, not trial and error. Measure what changes across the shorted plane, and the short that reads zero everywhere reveals its exact location.
Required Prerequisites
- Confirming and Characterizing a Short — Section 4.2 confirmed and measured the short with low-ohms and four-wire methods; this section uses those same measurements, followed as a gradient, to localize it.
- Current Measurement and In-Circuit Limits — Section 3.5 taught current and the current-limited supply; injecting a known current to create a voltage gradient builds directly on that.
Recommended Consumables
- A milliohm meter with four-wire leads — to resolve and follow the falling resistance gradient (confirming-and-characterizing-a-short)
- A sensitive millivolt or microvolt meter — to see the tiny voltage gradient an injected current makes
- Fine, sharp probes — to step precisely along a plane or trace without slipping
- A current-limited bench supply — to inject a known, safe current into the shorted rail (current-measurement-and-in-circuit-limits)
- A board layout or schematic — to map the shorted plane and plan the probe path
Recommended Practice Hardware
- A board with a hard short on a plane or wide trace — to practise following a gradient across copper (confirming-and-characterizing-a-short)
- A milliohm meter and four-wire leads — to read the resistance gradient
- A current-limited supply and a microvolt meter — to inject a current and follow the voltage gradient
- A known-good identical board — to compare readings and confirm the located point
- A short whose location you know — to check that the gradient really leads to it
- A magnifier — to inspect the converged-on point for the bridge or failed part
Real-World Applications
The gradient methods are how experienced repairers pin a hard short that reads zero everywhere. A technician with a rail shorted to ground follows the falling milliohms with a four-wire milliohm meter and walks the resistance gradient to the corner of the board where it bottoms out (confirming-and-characterizing-a-short). A repairer on a large ground plane short injects a couple of amps from a current-limited supply and follows the voltage gradient with a microvolt meter downhill to the bridge (current-measurement-and-in-circuit-limits). Someone whose gradient is too small to read raises the injected current within safe limits and switches to a more sensitive meter to bring the gradient into view. A technician converging on a point narrows the probe steps until the reading bottoms out at one spot, then inspects it and finds the solder bridge. And a repairer confirming the find lifts or removes the suspect and watches the short clear, proving the located point was the fault. The failures this prevents: giving up on a short that reads zero everywhere, chasing noise instead of the gradient, and inspecting the wrong end of the board.
Common Challenges
- A rail that reads zero everywhere. Ordinary probing cannot localize it — measure the tiny resistance or voltage gradient that does change (confirming-and-characterizing-a-short).
- A gradient too small to read. Milliohms and microvolts hide below a plain meter — use four-wire connections, a sensitive meter, and more injected current.
- Losing the gradient in noise. Sloppy contacts and thermal EMFs swamp a microvolt gradient — make clean contacts and steady, careful measurements (current-measurement-and-in-circuit-limits).
Safety Notes
Risk Level: Medium. The resistance-gradient method is unpowered and low-risk, but the voltage-gradient method injects a real current into the board, which is powered work, so this section is Medium risk.
Professional Tips Before Starting
- Follow the gradient downhill. The short is where resistance or voltage bottoms out — step toward the falling reading, not away from it (confirming-and-characterizing-a-short).
- Make the gradient readable. Milliohms and microvolts need help to see — four-wire connections, a sensitive meter, and enough injected current.
- Keep injection modest. Too much current cooks the board — inject only enough to make a readable gradient, current-limited (current-measurement-and-in-circuit-limits).
Localizing a Short by Low-Ohms and Voltage-Drop
Recap and Frame
Section 4.2 confirmed and characterized the short; this section localizes it, and the frame to hold is that because a short reads zero everywhere, localization measures the tiny things that still change across the shorted plane — resistance and voltage — and follows them to the fault (confirming-and-characterizing-a-short). The problem is the flat reading. A rail shorted to ground reads the same near-zero at every node, so the obvious measurement gives no direction, and a method that resolves the tiny differences is required (confirming-and-characterizing-a-short). Resistance still slopes. The copper from a probe to the short has a real, tiny resistance that decreases as the probe nears the fault, so the falling resistance is a gradient that points downhill to the short. Voltage can be made to slope. Driving a known current through the plane develops a tiny voltage along its own copper, sloping toward where the current leaves at the short, so an injected current turns the flat rail into a readable gradient. Both need resolution. Milliohms and microvolts are below an ordinary meter, so four-wire connections and sensitive instruments are what make the gradient visible. Both converge. Following either gradient, narrowing each step, converges on the shorted point, which is then confirmed directly rather than assumed. Hold the frame — a short reads flat, but resistance and an injected-current voltage still slope to it, and following those gradients with enough resolution converges on the fault — and the unlocatable short becomes locatable.
Following the Falling Resistance — the Resistance Gradient
The first method needs only a meter: because the resistance from any point to the short is tiny but real and falls as you approach, following that resistance gradient walks you to the fault (confirming-and-characterizing-a-short). Understand the resistance gradient. A resistance gradient is the gentle downhill in measured resistance across a shorted plane — the resistance from a probe point to the short decreases, milliohm by milliohm, as the probe moves nearer the fault — because there is less copper, and so less resistance, between probe and short the closer you get. Reference the shorted-to node. Keep one lead fixed on the node the rail is shorted to — ground, for a short to ground — and step the other probe across the shorted plane, so what you read is the copper from the probe to the short plus the short itself, and only the probe-to-short copper changes, falling toward a minimum as you reach the fault; a reference clamped on the same rail instead measures copper between two points of one net and gives no useful direction (confirming-and-characterizing-a-short). Use a four-wire low-ohms measurement. The gradient is milliohms, far below what a two-wire ohmmeter resolves through its own leads, so a four-wire, low-ohms measurement is essential to see the small changes that make the gradient (confirming-and-characterizing-a-short). Walk downhill. Step the probe in the direction of falling resistance, and where the resistance bottoms out to its lowest is the short, since that is the point of least copper to the fault. Mind the plane's spreading. On a wide plane the resistance falls in two dimensions, so the gradient is followed as a surface sloping to a low point, probing a grid rather than a line to find where it bottoms. Know its limits. The resistance differences are so tiny that on a very low-resistance plane the gradient can be near the meter's resolution, so where the resistance method runs out, the voltage method — with a larger injected current — takes over. The resistance gradient understood, measured from a reference, read four-wire, walked downhill, followed across a plane, and its limits known — and the falling resistance leads to the short. Follow the milliohms downhill, and the short is where they bottom out.
Injecting a Current to Make a Gradient
The voltage method needs a current to work with, and current injection is deliberately driving a known current through the shorted rail so that its own copper develops a voltage gradient to follow (current-measurement-and-in-circuit-limits). Understand current injection. Current injection is connecting a current-limited supply across the shorted rail — between the rail and ground for a short to ground — to drive a known, controlled current through the short and the copper leading to it, so that the plane carries a current whose voltage drop can be measured. Choose the injection points. The current is injected between a convenient point on the rail and the node it is shorted to, so it flows through the shorted plane toward the short, which is where the current crosses to the other node (confirming-and-characterizing-a-short). Set a safe, useful current. The injected current is set large enough to make a readable voltage gradient — often a fraction of an amp to a few amps — but limited so it does not exceed the trace or part ratings or overheat the board, balancing readability against safety. Use a current-limited supply. A current-limited bench supply is the ideal injector because it sets the current precisely and cannot exceed the limit, so it makes a controlled gradient without risking a runaway current through the fault (current-measurement-and-in-circuit-limits). Keep it modest and brief. Because the injected current heats the short and the copper, it is kept only as large and as long as needed to read the gradient, protecting the board while the measurement is made. Mind where the current flows. The injected current takes the shorted path, so it is injected so that its route runs through the region being probed, and not through a sensitive part that the injection could stress. Current injection understood, its points chosen, a safe current set from a limited supply, kept modest, and its path minded — and the plane carries a current to make a gradient. Inject a known current, and the flat rail gains a voltage that slopes to the short.
Following the Voltage Downhill — the Voltage Gradient
With a current flowing, the plane develops a voltage gradient that slopes to the short, and following that voltage downhill is the most powerful short-localization method there is (current-measurement-and-in-circuit-limits). Understand the voltage gradient. A voltage gradient is the tiny voltage the injected current develops along the shorted copper as it flows through the copper's own resistance — highest near where the current enters, falling along the path, and lowest at the short where the current crosses to the other node — so the voltage slopes downhill to the fault. Probe along the plane against a reference. Measure the voltage from a fixed reference point to successive points along the plane, or between adjacent points, so each reading places you on the slope and the direction of falling voltage points toward the short (confirming-and-characterizing-a-short). Follow the voltage downhill. Step toward lower voltage, since the short is the low point where the injected current leaves the plane, and where the voltage reaches its minimum is the fault. Resolve the microvolts. The gradient along good copper is microvolts to millivolts, so a sensitive millivolt or microvolt meter, and clean, low-thermal-EMF contacts, are needed to read it against the noise — the voltmeter draws almost no current, so here it is meter sensitivity and contact quality that matter, not the four-wire lead-cancellation the resistance method needs (confirming-and-characterizing-a-short). Read the two-dimensional slope on a plane. On a plane the voltage slopes toward the short in two dimensions, so the probe follows the surface downhill, sampling a grid to find the minimum, exactly as the resistance method does. Boost the gradient if it is too small. If the voltage gradient is too small to read, increasing the injected current within safe limits raises the whole gradient proportionally, bringing it above the noise (current-measurement-and-in-circuit-limits). The voltage gradient understood, probed against a reference, followed downhill, its microvolts resolved, read across a plane, and boosted when needed — and the voltage leads to the short. Follow the voltage downhill to its minimum, and there is the short.
Getting Enough Resolution
Both gradient methods hunt quantities near the limit of measurement, so getting enough resolution to see the gradient against the noise is not a detail but the difference between success and failure (confirming-and-characterizing-a-short). Use four-wire connections for the resistance gradient. A four-wire, Kelvin connection removes the lead and contact resistance that would otherwise swamp a milliohm resistance measurement, so it is the foundation of resolving the resistance gradient; for the microvolt voltage gradient, where the voltmeter draws almost no current and lead resistance is not the error, what matters instead is a sensitive meter and clean, low-thermal-EMF contacts (confirming-and-characterizing-a-short). Use a sensitive enough meter. The gradients are milliohms and microvolts, so a milliohm meter for resistance and a millivolt or microvolt meter for voltage are used, since an ordinary meter simply cannot resolve the change. Make clean, repeatable contacts. A poor or varying probe contact adds noise larger than the gradient, so sharp probes, firm consistent pressure, and clean pads are what keep successive readings comparable. Beware thermal EMFs. At the microvolt level, the tiny thermoelectric voltages of dissimilar-metal junctions and warm hands rival the signal, so contacts are kept at an even temperature and readings taken steadily to keep thermal EMFs from masking the gradient. Increase the signal where you can. For the voltage method, more injected current raises the gradient above the noise; for the resistance method, measuring over a longer span increases the resistance change per step — so the signal is boosted to beat the noise. Average and repeat. A gradient near the noise is confirmed by repeating and averaging readings, since a real gradient is consistent while noise is not, so a doubtful step is re-measured before it is trusted. Four-wire contacts, a sensitive meter, clean repeatable contacts, thermal EMFs beaten, the signal boosted, and readings averaged — and the tiny gradient is resolved. Give the method the resolution it needs, and a gradient that seemed invisible becomes a clear slope.
Converging on the Short and Confirming
Following a gradient is a convergence, and the final skill is narrowing to the shorted point and confirming it is the fault before any repair, so the localization ends in certainty (confirming-and-characterizing-a-short). Narrow the steps as you close in. Take large probe steps far from the short and smaller ones as the gradient steepens near it, so the search converges efficiently — coarse first, then fine — on the point where the reading bottoms out (fault-isolation-by-divide-and-conquer). Find the minimum. The short is where the resistance or voltage reaches its lowest, so the converged-on point is where stepping in any direction only raises the reading, marking the bottom of the gradient. Inspect the converged point. At the located point, inspect closely with magnification for the physical short — a solder bridge, a piece of swarf, a failed part, a via fault — since the gradient points at the spot but the eye confirms the cause (reading-failure-signatures). Confirm by clearing it. The definitive confirmation is to remove or lift the suspect and watch the short clear — the resistance to ground returns to normal, the gradient vanishes — proving the located point was the fault, not merely near it. Beware multiple shorts. If clearing one suspect only partly restores the rail, another short remains, so the gradient methods are re-run until the rail is fully clear, since a plane can carry more than one fault. Follow the fault to its cause. A localized short is confirmed and then traced to what caused it — an overvoltage, a failed upstream part — so the repair fixes the cause, not only the shorted point (confirming-and-characterizing-a-short). The steps narrowed, the minimum found, the point inspected, confirmed by clearing, multiple shorts checked, and the cause traced — and the short is localized and confirmed. Converge on the minimum and prove it clears, and the short that read zero everywhere is found and fixed.
Common Mistakes
- Trying to localize by the flat reading. A short reads zero everywhere — follow the tiny resistance or voltage gradient that does change (confirming-and-characterizing-a-short).
- Using a two-wire ohmmeter for the gradient. Lead resistance swamps the milliohm change — use a four-wire, low-ohms measurement.
- Injecting too much current. A large injection cooks the board — inject only enough to read the gradient, current-limited (current-measurement-and-in-circuit-limits).
- Chasing noise instead of the gradient. Thermal EMFs and poor contacts mimic a gradient — make clean contacts, average, and beat the noise.
- Stopping at the located point without confirming. The gradient points near the fault — confirm by clearing the suspect and watching the short vanish.
Troubleshooting Guidance
Gradient-localization problems come down to no resolution, a too-small gradient, or an unconfirmed point. If a short reads zero at every node: localize by the resistance or voltage gradient, not the flat reading (confirming-and-characterizing-a-short). If a two-wire ohmmeter shows no gradient: its leads swamp the milliohms — use a four-wire, low-ohms measurement. If the voltage gradient is too small to read: increase the injected current within safe limits and use a more sensitive meter (current-measurement-and-in-circuit-limits). If readings jump around near the noise: thermal EMFs and poor contacts are the cause — make clean, even-temperature contacts and average. If the gradient leads nowhere clear: sample a two-dimensional grid across the plane to find where the surface bottoms out. If you found a point but are unsure: confirm by lifting or removing the suspect and watching the short clear. If the short only partly clears: there is more than one short — re-run the gradient method on the remaining fault. The throughline: measure the gradient with enough resolution, follow it downhill, and confirm the point by clearing it.
Verification & Testing Methods
Confirm you localized the short by gradient and confirmed the point:
- [ ] I followed the falling resistance gradient with a four-wire, low-ohms measurement toward where it bottoms out (confirming-and-characterizing-a-short).
- [ ] I used current injection from a current-limited supply, at a safe, modest current, to create a measurable gradient (current-measurement-and-in-circuit-limits).
- [ ] I followed the voltage gradient the injected current made downhill to its minimum with a sensitive meter.
- [ ] I got enough resolution — four-wire contacts, a sensitive meter, clean contacts, thermal EMFs beaten — to resolve the tiny gradient.
- [ ] I converged on the point, inspected it, and confirmed it by clearing the suspect and watching the short vanish (reading-failure-signatures).
Then try the practice exercises below — gradient-localization practice on shorted boards; scenarios differ from the quiz.
Practice Exercises
- Follow the resistance gradient (5 minutes, hands-on). On an unpowered board with a hard short, use a four-wire milliohm meter to step across the plane and follow the falling resistance to where it bottoms out (confirming-and-characterizing-a-short).
- Inject a current (5 minutes, hands-on). Set a current-limited supply to a safe, modest level and inject it across a shorted rail, confirming the current flows through the shorted plane (current-measurement-and-in-circuit-limits).
- Follow the voltage gradient (5 minutes, hands-on). With the current injected, use a sensitive millivolt or microvolt meter to follow the voltage gradient downhill to its minimum, and note where it points.
- Converge and confirm (5 minutes, hands-on). Narrow to the minimum, inspect it under magnification, and confirm the fault by lifting or removing the suspect and watching the short clear (reading-failure-signatures).
These core steps — following the resistance gradient, injecting a current, following the voltage gradient, resolving the tiny signals, and converging on and confirming the short — are tested in the Chapter Quiz at the end of this chapter, where a score of 80% is required to continue.
Key Takeaways
- A short reads zero everywhere, so it is localized not by the flat reading but by the tiny things that change across the plane — a resistance gradient of falling milliohms toward the fault, read with a four-wire low-ohms measurement (confirming-and-characterizing-a-short).
- Current injection — driving a known, current-limited current into the shorted rail — makes the plane's own copper develop a measurable voltage difference to follow (current-measurement-and-in-circuit-limits).
- A voltage gradient — the microvolt slope the injected current develops along the copper, lowest at the short — is followed downhill to the fault, the most powerful localization method.
- Both methods hunt milliohms and microvolts, so four-wire connections, sensitive meters, clean contacts, and beating thermal EMFs are what make the gradient visible at all.
- Following a gradient is a convergence — narrow to where the reading bottoms out, inspect it, and confirm by clearing the suspect and watching the short vanish (reading-failure-signatures).
Skills Learned
- You can now localize a hard short by following the falling resistance gradient toward it.
- You can now inject a known current into a shorted rail to create a measurable gradient.
- You can now follow the voltage gradient the injected current makes downhill to the short.
- You can now get enough measurement resolution to resolve the tiny gradients.
- You can now converge on and confirm the shorted point.
Glossary Additions
- resistance gradient — the gentle downhill in measured resistance across a shorted plane or trace: the resistance from a probe point to the short is tiny but real, and it decreases, milliohm by milliohm, as the probe moves nearer the fault, because there is less copper between the probe and the short the closer you get. Following the falling resistance therefore leads to the short, which sits at the bottom of the gradient where the reading is lowest. The method works unpowered with only a meter, but because the differences are milliohms — far below what a two-wire ohmmeter resolves through its own lead resistance — it requires a four-wire, low-ohms measurement. Crucially, the reference lead is kept on the node the rail is shorted to (ground, for a short to ground) while the other probe is stepped across the plane, so the reading is the probe-to-short copper plus the short itself and only the probe-to-short term changes; a reference clamped on the same rail instead reads copper between two points of one net and gives no useful direction. On a wide plane the resistance falls in two dimensions, so the gradient is sampled as a surface sloping to a low point.
- current injection — deliberately driving a known, controlled current through a shorted rail — typically from a current-limited bench supply connected between the rail and the node it is shorted to — so that the copper carrying that current develops a measurable voltage drop that slopes toward the short. Current injection is what makes the voltage-gradient localization method possible: with no current there is no voltage difference along a shorted plane to follow, but an injected current turns the dead-flat rail into a sloped voltage map pointing at the fault. The injected current is set large enough to make a readable gradient — often a fraction of an amp to a few amps — yet limited so it cannot exceed a trace or part rating or overheat the board, and it is kept modest and brief because it heats the short and its path; a current-limited supply is used precisely because it sets the current and cannot run away through the fault.
- voltage gradient — the tiny voltage that an injected current develops along a shorted plane as it flows through the copper's own resistance: highest near where the current enters the plane, falling along the path, and lowest at the short itself, where the current crosses to the other node. Because the voltage slopes downhill to the fault, probing along the plane and stepping toward ever-lower voltage converges on the short — the point where the voltage reaches its minimum. The gradient along good copper is only microvolts to millivolts, so it demands a sensitive millivolt or microvolt meter and clean, low-thermal-EMF contacts to read against the noise — because the voltmeter draws almost no current, meter sensitivity and contact quality matter here rather than the four-wire lead-cancellation the resistance method needs — and if it is too small to see the injected current is raised within safe limits to lift the whole gradient proportionally. On a plane the voltage slopes toward the short in two dimensions, followed as a surface sampled on a grid to find the minimum.
Suggested Next Sections
Must read next:
- Thermal and Injection Methods for Shorts — Section 4.4 adds the methods that make the short reveal itself physically: warming it with an injected current until a thermal camera or a finger, or a freeze spray, shows exactly which part or bridge is heating.
Recommended:
- Confirming and Characterizing a Short — the low-ohms and four-wire measurements this section follows as a gradient.
- Current Measurement and In-Circuit Limits — the current and current-limited supply that current injection builds on.