I want to try repairing a TV. I have a soldering i...

Your current “cold-soaked” state is valuable. Do not power it yet. Prepare the board so the first successful boot becomes a reference measurement rather than merely confirming the symptom.

What the evidence suggests so far

Several conclusions are already reasonable:

• The LCD panel, T-con path, and backlight are unlikely to be the primary fault. Disconnecting the panel ribbons produced no meaningful change, and there is no operating system activity.
• The resistance readings do not indicate an obvious short. A processor core rail measuring about 30 Ω can be normal; semiconductor loads are nonlinear, so dividing voltage by that resistance does not predict operating current reliably.
• Resistance that rises slowly is usually the meter charging capacitors, not evidence of a faulty component.
• Requiring days unplugged is more suggestive of incomplete discharge, backfeeding, reset/enable timing, or a latched power condition than ordinary cooling. The board reaches room temperature long before 48 hours.
• The audio failure might be independent, but it could also involve a shared supply or an audio IC holding an I²C/control line low.

Modern processors require multiple rails to start in the correct order, with power-good and reset signals coordinated. A rail can measure correctly after startup yet still rise incorrectly, briefly collapse, or release reset at the wrong time. (Texas Instruments)

Important correction about the inductor readings

Do not interpret measurements such as:

11.8 V → 0.96 V across the 1R0 inductor

as regulator input and output.

On a typical buck regulator:

• One end of the inductor is the regulated DC output.
• The other end is the rapidly switching SW node.
• The actual input voltage normally enters the regulator IC directly, not through the inductor.

A multimeter can display misleading averages on the SW end. Measure the rail on the end connected to several output capacitors. Use continuity, while unplugged, to determine which end connects to the regulator’s SW pin and which connects to the load/output capacitors. Output ripple and switching-node behavior require an oscilloscope; a DC meter cannot establish that a rail is clean. (Texas Instruments)

Before the first cold boot

With the set unplugged:

1. Photograph the entire mainboard clearly.
2. Record the TV model, mainboard part number, and markings on:
   • Every regulator beside those inductors
   • The eMMC/NAND device
   • The audio amplifier
   • Any reset-supervisor or PMIC IC
3. Locate the output side of every inductor.
4. Look for test points labeled something like:
   • 12V, 5V, 3V3, 1V8, 1V2, VCORE, DDR
   • RESET, RST, POR
   • PG, PGOOD
   • EN
5. Look for an unpopulated serial header labeled UART, TX, RX, DBG, or similar.

Do not probe the mains/primary side of the power supply. Attach the meter ground lead to verified secondary-side mainboard ground with power disconnected, then use an insulated probe. Large primary capacitors can retain dangerous voltage after unplugging.

Use the successful boot as the reference state

When ready, power it on once and leave it running until all measurements are complete.

Measure each rail on its output capacitor, not the switching end of the inductor. Record it shortly after startup and again after 15–20 minutes.

A useful log is:

Test point	Good, 1 minute	Good, 20 minutes	Failed state	Notes
Main 12 V input
Standby 5 V or 3.3 V
Core rail, approximately 0.9–1.1 V
DDR rail
1.8 V rail
3.3 V rail
Each regulator EN
Each regulator PGOOD
Processor RESET/POR
eMMC VCC/VCCQ
Audio amplifier supplies
Audio MUTE/STBY/FAULT

For each regulator, identify its datasheet and measure:

• VIN
• VOUT
• EN
• PGOOD
• FB

Do not measure SW except with appropriate oscilloscope techniques.

Also examine the audio circuit while it is running

Because Bluetooth audio works, the OS and audio decoding path probably function. The speaker fault is likely farther downstream.

Check:

• Audio amplifier main supply, often called PVDD
• Amplifier logic supply
• MUTE, STBY, SD, or ENABLE
• FAULT
• Speaker connector and speaker resistance
• I²C SDA and SCL idle levels, if the amplifier is digitally controlled

A BTL/Class-D amplifier’s speaker outputs should not be treated as ground-referenced. Measure across each speaker pair, not from a speaker terminal to chassis ground.

Reproduce the failure in a way that separates the causes

After recording the good-state measurements:

1. Turn the TV off with the remote, but leave AC connected.
2. Wait about 30 seconds and try turning it on again.

Interpretation:

• Fails without unplugging: the fault is associated with warm restart, standby transition, reset, firmware state, or a temperature-sensitive component.
• Restarts successfully: shut it down again, unplug for 10–30 seconds, reconnect and retry. If that causes failure, the problem is specifically associated with the AC power-down/power-up sequence.

Record the exact LED pattern with slow-motion video. “Rapid blinking” may actually be a repeated sequence such as reset every 500 ms or a model-specific fault code.

Compare good and failed states

The comparison—not the absolute voltage—is what matters.

If one rail is absent or low

Measure that regulator’s EN:

• EN low: the regulator is probably obeying an upstream command. Trace the enable signal backward toward the PMIC, standby controller, or preceding regulator’s PGOOD.
• EN high but output missing: suspect the regulator, its input supply, or excessive loading.
• Output repeatedly appears and collapses: suspect overload protection, undervoltage lockout, unstable startup, or a load pulling it down.
• Output correct but PGOOD low: suspect excessive ripple, incorrect feedback voltage, a marginal regulator, or the PGOOD circuit itself.

If the meter has MIN/MAX capture, use it while pressing power. A normal displayed average can conceal a brief collapse.

If all rail voltages match the working state

Move to reset and boot activity:

• RESET remains asserted in the failed state: follow the reset supervisor and PGOOD chain backward.
• RESET releases, but no serial output and no eMMC clock activity: suspect SoC clocking, PMIC sequencing, reset integrity, or the processor.
• Boot activity begins and repeatedly restarts: suspect a watchdog/reset loop or unstable rail.
• Boot log or eMMC clock starts and then stops: investigate eMMC power, DDR initialization, and the boot device.

eMMC includes its own controller and communicates with the host processor through command, clock, and data signals. Watching whether the host attempts that communication is much stronger evidence than assuming “bad eMMC” from a black screen. (KIOXIA America, Inc.)

A USB-to-UART adapter is often the highest-value additional diagnostic tool. Verify the logic voltage first—it may be 3.3 V or 1.8 V—and connect only ground, adapter RX, and adapter TX. Do not connect the adapter’s power output to the TV board.

Test the long-discharge hypothesis directly

Once the TV is in its failed state:

1. Unplug it.
2. Immediately measure the identified rails.
3. Continue recording after:
   • 10 seconds
   • 1 minute
   • 5 minutes
   • 30 minutes
   • Longer if a rail remains elevated
4. Include RESET, EN, and PGOOD-related nodes where accessible.

Most logic rails should collapse reasonably quickly. A node remaining at a few hundred millivolts or more for a very long period is important, especially if it feeds reset, enable, standby logic, or eMMC.

First try holding the physical power button for 30–60 seconds while unplugged, then reconnect. If that makes the TV boot immediately, retained charge or incomplete discharge becomes very likely.

Controlled discharge test

On confirmed secondary-side low-voltage rails only, you can use a 1 kΩ, ¼-watt resistor from the rail to ground for 30–60 seconds while the TV is unplugged.

Do this one rail at a time:

1. Confirm voltage is below that rail’s normal value and the AC cord is disconnected.
2. Apply the resistor across the rail’s output capacitor.
3. Confirm it falls near zero.
4. Remove the resistor.
5. Reconnect power and test.

Do not short rails with a screwdriver. Do not apply this procedure to the mains capacitor or an unidentified node.

If discharging one particular rail immediately restores booting, you have localized the fault to:

• That rail’s discharge path
• Backfeeding from another rail
• Its regulator’s power-off behavior
• An associated reset/enable circuit
• A device remaining partially powered through an I/O pin

Then check for reverse voltage coming through signal lines while the regulator itself is off.

Investigate whether the audio fault is related

Disconnect the speaker connector before one restart test. A shorted speaker or amplifier output can trigger amplifier protection, although it would not normally prevent the entire OS from booting.

More importantly, check SDA and SCL at the audio amplifier:

• Both should normally idle high when inactive.
• If one is held near 0 V in the failed state, disable or disconnect the audio amplifier supply and retry.
• If booting returns, the amplifier or its control bus is involved.

If there is a small fuse, zero-ohm link, or separate inductor feeding the audio amplifier, opening that supply is a cleaner diagnostic than removing the IC.

Temperature testing

After reproducing the failed state, cool components individually and retry after each one:

1. Reset supervisor and its small capacitors
2. PMIC and buck regulators
3. Main crystal/oscillator
4. eMMC
5. Audio amplifier
6. SoC area

Use electronics freeze spray sparingly and avoid condensation. This is a localization test, not a repair. Similarly, gentle localized warming can help establish sensitivity. Do not “reflow” the SoC or eMMC based solely on the symptom.

If cooling a component restores booting within seconds, that is meaningful. If the entire board must remain unpowered for days, electrical discharge/backfeed remains more likely than a simple thermal solder fault.

Provisional likelihood ranking

Based only on the present evidence:

1. Power-down/startup sequencing or reset problem, potentially involving a leaky timing component, PGOOD chain, incomplete rail discharge, or backfeeding.
2. Shared supply/control-bus problem, possibly involving the audio amplifier because the internal sound path is already abnormal.
3. eMMC, DDR, oscillator, or early boot failure that is state- or temperature-dependent.
4. Marginal BGA or cracked solder joint.
5. Panel, T-con, and gross power-supply failure are presently less likely.

The most decisive next evidence will be a good-versus-failed rail/control-signal table, the rail discharge timeline, regulator part markings, and any UART boot output.