Electrical Verification After Rework

Section Overview

The last two sections proved a joint can be judged to look right — its fillet, its wetting, its surface, under magnification and the right light (10.2; 10.3). But looking right and working right are not the same thing, and this section is the second half of verification: proving the reworked board actually functions electrically. The reason is simple and important: some faults are electrically real yet visually invisible. A joint can look perfect and still be a marginal, high-resistance connection; the wrong-value or wrong part can be fitted and look identical to the right one; a lead can have an internal open the surface hides; a hairline solder bridge can escape even a microscope. Inspection cannot catch these — only a meter and a powered test can — so a rework is not verified until it has been checked electrically. You work a safe sequence. Unpowered first: with no power on the board, check continuity of the new joints, probe for pin-to-pin shorts, and confirm no power rail is shorted to ground (9.6). Then the cautious powered check: bring the board up on a current-limited supply — the smoke test — watching the current draw against what you expect, because excess current means a fault; then a functional test that the repaired circuit does its actual job (9.6). Beyond the handheld meter, production verifies at scale: an in-circuit test probes every node of the board through a bed-of-nails fixture or a flying probe to check each component in place, and boundary scan tests the pins and interconnects of complex ICs through a dedicated test access port — reaching nets you cannot physically touch, such as the balls hidden beneath a ball-grid array. Finally, you match the rigor to the stakes: a hobby board may need only a meter and a power-up, while a Class 3 board may require documented, automated electrical test (10.1), and you record what you tested and how it came out. Inspection says it looks right; electrical verification says it works right — and only the second finishes the repair.

Why This Matters

A repair that looks perfect and does not work is still a failed repair — and the only way to tell the difference is to test it electrically, because the eye cannot. This matters because appearance and function diverge: the faults that most often survive a good inspection — a high-resistance joint, a wrong part, an internal open, a fine bridge — are exactly the ones that look fine and fail electrically, so inspection alone leaves them uncaught (10.2; 10.3). This matters because the unpowered checks are cheap safety: a continuity check and a short circuit check take a minute with the board off and catch the faults that would otherwise turn a first power-up into a fire (9.6). It matters because the powered check is where proof and hazard meet: powering a reworked board is the moment a hidden short or wrong part can overheat or catch fire, so doing it current-limited and watchful is both how you verify function and how you stay safe (Chapter 2). It matters because scale needs automation: a handheld meter cannot check every node of a dense board or reach under a ball-grid array*, so production leans on* in-circuit test and boundary scan to verify what hands cannot (10.5). It matters because the right amount of testing depends on the product: over-testing a toy wastes effort and under-testing a medical board is dangerous, so matching the verification to the acceptance class is part of doing the job correctly (10.1). And it matters because a documented test is a defensible repair: recording what you verified and the result makes the work traceable for warranty, safety, and the next technician. Verify electrically, and the repair is proven; skip it, and a good-looking board can still fail in the field.

Required Prerequisites

• Using Microscopy for Inspection — Sections 10.2 and 10.3 confirmed the joint looks right; this section confirms it works. You should also know the meter-before-power and cautious first power-up discipline from the post-install check (9.6) — continuity, pin-to-pin and rail-to-ground shorts, the smoke test*, and the* functional test — and the electrical-safety basics for powering a board (Chapter 2).

Recommended Consumables

• A repair and test log — to record what was tested — continuity, shorts, current, function — and the outcome
• Fine probe tips and clip leads — for reaching pins and holding a measurement without a slip
• Isopropyl alcohol — to clean flux so probes make clean contact and shorts read true (Chapter 3)
• Eye protection — for the powered check, in case a part vents

Recommended Practice Hardware

• A multimeter with continuity, resistance, and diode-test modes — the core hand tool for the unpowered checks (multimeter; continuity)
• A current-limited bench power supply — the safest way to do the powered check (highly recommended)
• A board with a known-good and a known-faulty rework to compare — to practice telling a passing test from a failing one
• If available, access to an in-circuit tester or a JTAG/boundary-scan setup — to see automated verification in action (not required)
• An ESD-safe bench and, for mains devices, the electrical-safety setup of the earlier sections

Real-World Applications

Electrical verification is where a repair is finally proven or sent back, and it catches what no amount of looking can. A technician who reworked a fine-pitch part inspects it clean under the scope, then meters continuity and finds one lead reads open — a joint that looked wetted but never truly connected — and reflows it (10.3). A repairer metering rails before power catches a low resistance from a supply rail to ground, traces it to a bridge under a part, and clears it before it can do harm (9.6). Someone bringing a board up current-limited watches the current climb far past what the circuit should draw, cuts power instantly, and finds a wrong-value part that would have cooked on a bench supply without a limit. A production line verifying a dense assembly runs an in-circuit test on a bed-of-nails to confirm every component is present and correct, then a boundary scan to check the interconnects under its BGAs that no probe can reach. And a shop returning a Class 3 board runs and documents a full electrical test because the product's class demands a defensible record, not just a technician's word (10.1). The failures electrical verification prevents: an open joint that looked fine, a rail short caught before power, a wrong part caught by current draw, an unreachable net verified by boundary scan, and an undocumented repair on a critical board — all caught by testing, not just looking.

Common Challenges

• Trusting a clean inspection as proof it works. A perfect-looking joint can still be electrically bad — meter it and power it before you call it verified (10.2).
• Powering up before metering for shorts. A hidden short turns a power-up into a hazard — always do the unpowered continuity and short checks first (9.6).
• No way to reach a net. You cannot probe under a ball-grid array by hand — that is what boundary scan and X-ray are for (10.5).

Safety Notes

Risk Level: Medium. The unpowered checks are low-risk, but this section powers the reworked board — and a hidden short or wrong part can, on power, overheat, vent, catch fire, or damage the board.

Professional Tips Before Starting

• Meter fully before you ever power. Continuity of the new joints, pin-to-pin shorts, and every rail to ground, all with the board off — the unpowered checks are cheap and they are what make the power-up safe (9.6).
• Power up current-limited and watch the current. Set a sensible current limit, bring it up slowly, and compare the draw to what the circuit should take — excess current is a fault, and the limit turns a fault into a trip instead of a fire.
• Test to the product, not to habit. A simple board needs a meter and a power-up; a critical one needs documented, maybe automated test — match the rigor to the acceptance class (10.1).

Proving the Rework Works Electrically

Why Electrical Verification

The starting point is a hard truth about inspection: it confirms how a joint looks, not whether it works, and the two can disagree. A visual and microscope inspection can pass a joint on every criterion — good fillet, clean wetting, no visible bridge — and that joint can still be electrically faulty (10.2; 10.3). Consider the faults inspection cannot see. A joint can look fully wetted and yet be a marginal, high-resistance connection that drops voltage, runs warm, or fails intermittently — the resistance is invisible to the eye. The wrong part can be fitted — a resistor of the wrong value, a capacitor of the wrong rating, a part rotated or substituted — and look identical to the correct one. A lead can have an internal open inside the package or under the joint that no surface view reveals. A hairline solder bridge can hide between fine-pitch pins even under magnification. Every one of these is electrically real and visually silent, so the only way to catch them is to measure and to power the board. That is why electrical verification is the necessary second half of finishing a rework: inspection asks "does it look right?" and electrical test asks "does it work right?", and a repair is not done until both are answered yes. This is not a replacement for inspection — you inspect first, because many faults are visible and cheap to catch by eye — but inspection alone is never proof of function. Look first, then test: only when the board both looks right and measures and works right is the rework truly verified.

Unpowered Electrical Checks

Electrical verification always begins with the board unpowered, because the checks you can do cold are the ones that keep the first power-up safe. With no power applied, take a multimeter and work through three checks — the same meter-before-power discipline you learned after an install (9.6). First, continuity: probe each new joint from the component lead to the trace or a known test point and confirm it actually connects where it should, catching an open a visual missed. Second, pin-to-pin shorts: probe between adjacent leads and confirm they are not connected except where the design joins them, catching a solder bridge the eye or even the scope did not. Third, the power rails: measure from each supply rail to ground and confirm it is not a dead short — a near-zero reading that does not climb signals a bridge or a backwards part that would be catastrophic on power. Where the design allows, you can go further: an in-circuit resistance or diode-test measurement of the reworked component itself — checking a resistor reads near its value, a diode conducts one way — though remember that other parts on the same net can influence an in-circuit reading, so interpret it with the schematic in mind. All of this is done cold, because the whole point is to find a fault with an ohmmeter before power can turn it into damage. The unpowered checks are quick, safe, and decisive: they confirm the new connections are made and no short is waiting, so that when you do apply power, you are powering a board you already have good reason to trust.

The Cautious Powered Check

Only once the unpowered checks pass do you apply power — and even then carefully, because the power-up is where a hidden fault becomes a hazard and where function is finally proven. Bring the board up on a current-limited bench supply set to a sensible limit — this is the smoke test from the post-install check, and the current limit is what makes it safe (9.6). As power comes up, watch the current draw and compare it to what the circuit should take: a board drawing far more than expected has a fault — a short, a wrong part, a backwards component — and you cut power immediately and go back to the meter. Watch and feel, too, for a part heating, and for any smoke or smell, cutting power at the first sign of trouble. A current-limited supply turns a fault into a gentle trip at the limit instead of a destructive surge, which is exactly why you use one for a first power-up. If the board powers up cleanly — normal current, no heat, no smoke — you move to the test that actually proves the repair: a functional test (9.6). Powering without a fault is necessary but not sufficient: a board can draw normal current and still not do its job if the wrong part was fitted or a joint is marginal, so you verify that the repaired circuit performs its intended function — the device turns on, the signal is present, the feature works. For a mains-powered board, remember that you cannot current-limit the mains with a bench supply — there the equivalent tools are a series-lamp "dim-bulb" limiter, which does the current-limiting, and an isolation transformer, which provides shock-safety isolation rather than any current limiting, per the earlier electrical-safety guidance (Chapter 2). Power up cautiously, watch the current, then prove the function: that sequence both keeps you safe and confirms the rework works.

Automated and Production Verification

A handheld meter and a bench supply are enough for a single repair, but production and complex boards need verification that reaches further and faster than a hand can — and two automated methods do exactly that. The first is in-circuit test — an automated tester that probes the individual nodes of a populated board to check each component in place. It makes contact either through a bed-of-nails fixture (a bed of spring-loaded pins that touch every test point on the board at once, custom-built for that board) or a flying probe (a small number of fast-moving probes that visit each node in turn, needing no custom fixture). An in-circuit test verifies that each part is present, the right value, the correct orientation, and properly connected — catching a wrong or missing or misconnected component automatically, across the whole board, far faster and more completely than probing by hand. The second is boundary scan — a method standardized as JTAG (IEEE-1149.1) that tests the pins and interconnections of complex ICs through a dedicated test access port built into the chips. Each pin of a boundary-scan device has a small register cell, and the chips are chained together so a tester can drive and read every pin through a few dedicated signals — without a physical probe on each net. That is what lets it verify connections you cannot physically reach, above all the balls hidden underneath a ball-grid array*, where no bed-of-nails pin or flying probe can touch.* Many boards also run a built-in self-test or are checked during in-system programming, exercising their own circuitry. You will not own this equipment as a bench repairer, but knowing it exists matters: it is how a factory verifies at scale and how the un-probeable nets get tested, and it sets the standard a critical rework is ultimately held to.

Matching Verification to the Stakes

Not every board needs the same depth of electrical verification, and matching the rigor to the product is part of doing the job right — the same idea as the acceptance classes. The verification a rework needs scales with the product's acceptance class and role (10.1). A hobby board or a simple consumer repair may be fully verified by the hand sequence: meter for continuity and shorts, bring it up current-limited, and confirm it functions. A product that must give reliable service — a Class 2 assembly — may warrant more thorough functional testing across its operating range. A high-reliability, Class 3 board — medical, aerospace, life-support — may require documented, repeatable, often automated electrical test, because a technician's "it worked when I tried it" is not an acceptable standard of proof for a product that must not fail. The principle runs both ways: under-testing a critical board leaves a dangerous fault uncaught, and over-testing a trivial one wastes effort and adds handling risk to a good board. So before you verify, ask what the product requires: what class is it, what does it do, what happens if this repair fails in service? The answer sets how far you take the verification — from a quick meter-and-power check to a full documented test suite. Test to the stakes: enough to prove the repair is sound for what the product must do, and no less on the boards where it matters most.

Documenting the Result

The last step turns a verified rework into a defensible one: record what you tested and how it came out. For anything beyond casual hobby work, a repair that was tested but not documented is, to anyone who comes after, a repair whose verification cannot be trusted. So note the essentials: what the fault was, what you reworked, and what you verified — the continuity and short checks, the current draw on power-up, the functional test result, and any automated test that was run — together with the outcome and the date. This record matters for several reasons. It makes the repair traceable: warranty, safety, and quality all depend on being able to show that the work met a standard, not just that it was done. It helps the next technician — often you — who works on the board and needs to know what was already touched and tested. And for a product held to an acceptance class*, documentation is frequently part of the requirement itself* (10.1). The record does not have to be elaborate — a line in a repair log, a filled test sheet, a saved automated-test report — but it does have to exist for work that must be defensible. A rework is finished when it has been inspected, verified electrically, and recorded: the documentation is what lets anyone, later, trust that the repair was proven and not just performed.

Common Mistakes

• Calling a repair done because it looks good. Appearance is not function — a clean-looking joint can be open, high-resistance, or the wrong part; meter and power it (10.2).
• Powering before the unpowered checks. A short found with an ohmmeter cannot start a fire; one found on power can — always meter for shorts first (9.6).
• Powering up without a current limit. An un-limited supply lets a fault surge and burn the board — bring a fresh rework up current-limited and watch the draw.
• Assuming a scope or a meter can reach everything. You cannot hand-probe the nets under a ball-grid array — those need boundary scan or X-ray (10.5).
• Testing every board the same way. Rigor should match the acceptance class — a critical board needs documented test, a trivial one does not (10.1).

Troubleshooting Guidance

Verification problems trace to a connection, a short, the current, or reach. If a new joint reads open on continuity: it did not truly connect — reflow it and re-check (9.6). If two pins read shorted: a solder bridge — wick it and re-meter (10.2). If a rail reads a short to ground: a bridge or a backwards part — find and clear it before any power (Chapter 2). If the current draw is far above expected on power-up: cut power immediately — a short, a wrong part, or a backwards component; re-inspect and re-meter (smoke test). If it powers but does not work: a functional fault — a wrong value, a marginal joint, or the original problem not fixed; re-diagnose (functional test). If an in-circuit reading looks wrong but the part is fine: other components on the net may be influencing it — interpret in-circuit with the schematic. If you cannot reach a net to test it: it is under a part — boundary scan or X-ray is the route (10.5). If the board must meet a standard: run and document the verification its acceptance class requires (10.1). The throughline: meter cold, power up current-limited and watch the draw, prove the function, and reach the rest with automated test.

Verification & Testing Methods

Use this as an electrical-verification check after a rework:

• [ ] I confirm the joint passed visual and microscope inspection first, then verify electrically because appearance is not proof of function (10.2; 10.3).
• [ ] I check continuity of the new joints, pin-to-pin shorts, and every power rail to ground with the board unpowered (9.6).
• [ ] I bring the board up on a current-limited supply, watching the current draw against what the circuit should take, and cut power at any sign of trouble (9.6).
• [ ] I run a functional test to confirm the repaired circuit actually performs its job, not just that it powers up.
• [ ] I know that automated in-circuit test (bed-of-nails or flying probe) and boundary scan (JTAG) verify what a hand cannot, including nets hidden under a ball-grid array (10.5).
• [ ] I match the verification rigor to the product's acceptance class and document what I tested and the outcome (10.1).

Then try the practice exercises below — verification practice; scenarios differ from the quiz.

Practice Exercises

1. Do the unpowered checks (7 minutes, applied). On a reworked board, meter continuity of the new joints, probe for pin-to-pin shorts, and confirm no rail is shorted to ground — all with the board off (9.6).
2. Cautious current-limited power-up (7 minutes, applied). Bring a repaired board up on a current-limited supply, compare the current draw to what you expect, and practice cutting power the instant it climbs too high.
3. Choose the verification (5 minutes, reasoning). For three products — a toy, a home appliance board, and a medical device — decide what electrical verification each rework should get and why, using the acceptance class (10.1).
4. Explain the automated methods (6 minutes, reasoning). In your own words, describe what an in-circuit test and a boundary scan each verify, and give one net or fault each catches that a handheld meter cannot.

These core ideas — why electrical test is needed after inspection, the unpowered checks, the cautious powered check and functional test, in-circuit test and boundary scan, and matching rigor to the class — are tested in the Chapter Quiz at the end of this chapter, where a score of 80% is required to continue.

Key Takeaways

• Inspection confirms a joint looks right; electrical verification confirms the board works right — and some faults (a marginal high-resistance joint, a wrong-value part, an internal open, a hidden solder bridge*) are electrically real but visually invisible* (10.2; 10.3).
• Verify unpowered first: check continuity of the new joints, pin-to-pin shorts, and every power rail to ground with the board off — a short found with an ohmmeter cannot start a fire (9.6).
• Then power up cautiously: bring the board up on a current-limited supply, watch the current draw against what is expected, cut power at any sign of trouble, and finish with a functional test that the circuit does its job (9.6; Chapter 2).
• Beyond the hand meter, production uses in-circuit test (a bed-of-nails or flying probe checking every node) and boundary scan (JTAG testing the pins and interconnects of complex ICs through a test access port) to verify nets you cannot physically probe, such as under a ball-grid array (10.5).
• Match the verification rigor to the product's acceptance class — a hobby board needs a meter and a power-up, a Class 3 board needs documented, often automated test (10.1) — and record what you tested and the outcome.

Skills Learned

• You can now explain why a visually perfect joint still needs electrical verification.
• You can now run the unpowered continuity, short, and rail checks before power.
• You can now do a cautious current-limited power-up and functional test safely.
• You can now describe how in-circuit test and boundary scan verify a board.
• You can now choose the verification rigor a product's class requires.

Glossary Additions

• in-circuit test — an automated electrical test that probes the individual nodes of a populated circuit board to verify each component is present, of the correct value, correctly oriented, and properly connected, going beyond what a handheld meter checks by hand; abbreviated ICT. It makes contact through either a bed-of-nails fixture (a board-specific bed of spring-loaded pins that touch every test point at once) or a flying probe (a few fast-moving probes that visit each node in turn without a custom fixture), and it catches wrong, missing, or misconnected parts automatically across a whole board. In-circuit test is complemented by boundary scan for the nets a physical probe cannot reach.
• boundary scan — a method of testing the pins and interconnections of complex integrated circuits through a dedicated test access port built into the chips, standardized as JTAG (IEEE-1149.1); a small register cell at each pin, with the devices chained together, lets a tester drive and read every pin through a few dedicated signals rather than a physical probe on each net. Boundary scan verifies connections that cannot be physically probed — most importantly the solder balls hidden underneath a ball-grid array — and works alongside in-circuit test, which probes the physically accessible nodes of a board.

Suggested Next Sections

Must read next:

• X-Ray Inspection — When It's Needed — electrical test proves the board works and boundary scan reaches nets you cannot probe, but neither shows you the hidden joint itself; the final section is X-ray, which sees through the assembly to image the balls under a BGA, the fill inside a barrel, and the voids no light or probe can reach.

Recommended:

• Using Microscopy for Inspection — the visual side of verification that confirms a joint looks right, which this section pairs with proving it works.
• Post-Installation Inspection — the meter-before-power, smoke-test, and functional-test discipline this section builds directly on.