Testing components in-circuit saves an enormous amount of desoldering, but the reading you get is never about one part. Your meter measures the whole network between its two probes, so a resistor reads low, a capacitor reads high, and a diode may appear shorted by a winding you cannot see. Knowing which direction the error runs is what makes the technique usable.
Everything in parallel joins the measurement
When you place two probes across a component on a board, current from the meter can take any path that exists between those two points. Other resistors, coils, capacitors, semiconductor junctions and the traces themselves all participate. Because parallel resistances always combine to something lower than the smallest one, a resistance reading in circuit is equal to or lower than the true value, never higher. That asymmetry is genuinely useful. A resistor reading higher than its marking is almost certainly faulty, because nothing in parallel can raise a reading. A resistor reading lower may be perfectly fine. Applying that logic first prevents most unnecessary component removal, and it works without knowing anything about the surrounding circuit.
Stored charge and live circuits come first
Before any in-circuit test, the equipment must be disconnected from its supply and the energy storage discharged. Large capacitors in power supplies hold a dangerous charge long after the equipment is switched off, and some hold it for a long time with no bleed path. Charge can also reappear on a capacitor that was discharged quickly, as the dielectric relaxes. Verify with a meter that the voltage has actually fallen and stayed down rather than assuming a discharge worked. Testing components while a circuit is energized is a different activity entirely, requiring appropriate training, personal protective equipment and instruments rated for the installation; if that is the situation you are in and you are not trained and equipped for it, the correct decision is to stop rather than improvise.
Semiconductors and low-value paths that mislead
Diode test mode applies enough voltage to forward-bias a junction, which also forward-biases every other junction in a parallel path. A transistor tested in circuit frequently reads as though its junctions are shorted, when in reality a winding, a low-value resistor or another junction is bridging the measurement. Ohmmeter ranges with higher test voltages can turn nearby junctions on and produce readings that change if you reverse the probes, which is itself a clue that a semiconductor is participating. Inductors and transformer windings read as near shorts to a DC ohmmeter regardless of health, so a low reading across anything with a coil nearby is close to meaningless. In all these cases the answer is to lift one lead of the component or remove it entirely.
Capacitance and inductance in circuit
A capacitance reading taken on the board includes every other capacitor on the same rail, so the figure returned is higher than the part alone and is often dominated by the bulk capacitance elsewhere. That makes in-circuit capacitance testing nearly useless for confirming a specific small capacitor, though it can reveal a total that is wildly wrong. Inductance measurement suffers a related problem: nearby windings couple, and shorted turns elsewhere in a magnetic assembly change the reading. Meters that inject a small signal, including many ESR testers, avoid turning semiconductors on and therefore give more trustworthy in-circuit results than a general-purpose multimeter, but they still cannot separate parallel components of the same type. Treat any in-circuit reactance reading as a screening result rather than a verdict.
Working out when a part must come off
The practical rule is to use in-circuit testing to eliminate parts rather than to confirm them. A reading that is impossible for a healthy part, such as a resistor above its marked value or a truly open connection where a path should exist, is a finding you can act on. A reading that merely looks suspicious is a candidate for removal. Lifting one end of a two-terminal component is usually enough and is faster and gentler than full removal. Where a board is dense and removal risks damage, comparing against the same measurement point on a known-good identical board is often the quickest way to decide. Record what you measured and where, because the comparison you make next week depends on it.
What to check before you act
- Disconnect the supply and verify stored charge has actually fallen.
- Treat a low in-circuit resistance reading as inconclusive, a high one as a finding.
- Lift one lead before judging any semiconductor or reactive part.
- Compare against an identical known-good board where one is available.
Common questions
Why does a good transistor test as shorted on the board?
Parallel windings, low-value resistors or other junctions carry the meter’s test current. Lift a lead and retest before replacing anything.
Is in-circuit testing ever conclusive?
Yes, when the reading is impossible for a healthy part, such as an open path or a resistance above the marked value. Anything else needs the part isolated.
The practical takeaway
In-circuit measurement is a filter, not a verdict. Discharge the equipment first, use readings that can only be caused by a fault to act, and lift a lead whenever a result merely looks odd. Applied that way it removes most of the guesswork without removing most of the components.
References and further reading
- NFPA 70E: standard for electrical safety in the workplace
- CDC/NIOSH: electrical hazards and safe work practices
- OSHA: Subpart S, electrical requirements and safe work practices
Numerical scenarios are illustrative unless identified otherwise. Follow the exact product instructions; component ratings and local installation requirements can differ.
