Sonoff TH10 disconnects after a few seconds and needs hard reset

User question

sonoff th10 disconnects after few second. Hang and needs hard reset

Artificial Intelligence Response

Direct answer to the question

A Sonoff TH10 that disconnects after a few seconds, freezes, and only recovers after a power reset is most commonly suffering from one of these four problems:

• Faulty or shorted sensor
• Unstable 2.4 GHz Wi‑Fi / bad network re-pairing state
• Power-supply brownout inside the TH10
• Corrupted or unsuitable firmware/configuration

The fastest diagnosis is:

1. Unplug the sensor
2. Disconnect the load from the relay output
3. Factory reset and re-pair on 2.4 GHz only
4. Test it close to the router or on a phone hotspot
5. If it still freezes, suspect internal hardware power failure, especially aged capacitors or 3.3 V rail instability

If the unit hangs even with no sensor and no load connected, the problem is very likely inside the device, not in the app.

---

Detailed problem analysis

This symptom pattern is very characteristic of an ESP8266-based device entering an unstable state during or just after boot:

• It powers up
• Starts Wi‑Fi association
• Current consumption rises sharply
• Either firmware crashes, a sensor bus locks up, or the internal supply voltage dips
• The MCU stops responding and requires a hard power cycle

For a Sonoff TH10, the root cause is usually in one of the following domains.

---

1. Sensor-related failure is the first thing to check

The TH10 is not just a relay; it is designed to work with external temperature/humidity probes. If the connected probe or its cable is defective, the unit may appear to boot normally and then hang when firmware starts polling the sensor.

Typical mechanisms:

• Moisture ingress into the sensor head
• Cable damage causing intermittent short/open circuit
• Wrong or non-genuine sensor with incompatible wiring
• Sensor bus lockup that stalls the firmware

This is especially plausible if:

• The problem began after changing or moving the probe
• The device disconnects shortly after boot rather than instantly
• The relay hardware itself seems fine

Engineering interpretation:
A bad probe can either:

• electrically disturb the 3.3 V logic rail, or
• cause a communication deadlock on the sensor interface

Test:

• Remove the sensor completely
• Power the TH10 with only mains input connected
• Do not connect any load to the relay output
• Observe whether it remains online for several minutes

Result interpretation:

• Stable without sensor: the sensor or cable is the cause
• Still unstable: continue with power/network/hardware diagnosis

---

2. Wi‑Fi instability and bad pairing state are common on older IoT devices

Sonoff TH10 uses 2.4 GHz Wi‑Fi only. These devices can behave badly on modern routers when:

• 2.4 GHz and 5 GHz share the same SSID with aggressive steering
• Airtime fairness is enabled
• Signal is marginal
• DHCP or cloud registration gets stuck after ISP/router changes
• The device was not properly reset before re-pairing

Typical visible behavior:

• connects briefly, then shows offline
• pairs but drops almost immediately
• app loses it even though power is present

Important correction:
A pure Wi‑Fi problem usually makes the device go “offline,” but a full freeze requiring hard reset points more strongly to either:

• firmware fault,
• sensor issue,
• or internal power integrity problem.

Still, Wi‑Fi should be ruled out early because it is easy to test.

Recommended network checks:

• Use 2.4 GHz only
• If your router combines bands, temporarily create a dedicated 2.4 GHz SSID
• Move the unit close to the router
• Or test using a simple phone hotspot on 2.4 GHz if available
• Reserve a DHCP IP for the TH10 in the router
• Avoid crowded channels if possible

---

3. Internal power-supply degradation is a very strong candidate

From an electronics perspective, your symptom strongly suggests a brownout or rail collapse.

The TH10 contains:

• an AC mains power supply section
• a low-voltage DC rail
• a 3.3 V regulator for the ESP8266 logic

During Wi‑Fi startup, the ESP8266 draws short current bursts. If the low-voltage supply has degraded capacitance or high ESR, the 3.3 V rail can dip long enough to cause:

• lockup
• watchdog resets
• corrupted behavior
• complete hang until power is removed

This is one of the most common aging faults in compact mains-powered IoT devices.

Typical failing parts:

• low-voltage electrolytic filter capacitor
• primary-side bulk electrolytic capacitor
• 3.3 V regulator section
• poor solder joints around regulator or PSU components

Why it freezes after a few seconds:
The device may boot correctly when current demand is low, then fail when:

• Wi‑Fi transmits
• cloud connection starts
• sensor polling begins
• relay state changes

That timing matches your description very well.

Strong indicators of PSU trouble:

• problem happens even with no sensor
• problem happens even without relay load
• unit gets worse over time
• warm environment or older unit
• recovery only by full power removal

---

4. Relay load or mains wiring can aggravate the issue

Although the TH10 relay is rated for switching significant load, the device can misbehave if:

• load current is too high
• wiring is loose
• neutral is poor
• inductive load produces noise/spikes
• line voltage sags under load

Examples:

• heaters with inrush
• pumps or fans
• inductive coils without suppression
• poorly clamped terminals

Test properly:

• Disconnect the controlled load entirely
• Leave only mains input to the TH10
• Check whether the freezing still occurs

If the device is stable with no load but unstable with the load connected, investigate:

• load current
• terminal tightness
• wiring gauge
• snubber/surge suppression if inductive

---

5. Firmware or flash corruption is possible

If the hardware is electrically sound, firmware is the next layer.

Possible firmware-side issues:

• corrupted configuration
• damaged flash contents
• failed OTA update
• incompatible custom firmware
• sensor-driver bug in aftermarket firmware

If you are using stock Sonoff/eWeLink firmware, start with reset and re-pair.
If you are using Tasmota or ESPHome, misconfigured GPIOs or bad builds can absolutely cause freezes.

Important nuance:
Some online reports refer to specific firmware/sensor combinations. Those cases are real, but they should not be generalized to every TH10 failure. For stock devices, hardware and sensor isolation still come first.

---

Current information and trends

Based on recent user reports and manufacturer-style troubleshooting guidance reflected in the provided material, the most relevant current practical trends are:

• Factory reset and re-pairing remain the first official-style remedy
• 2.4 GHz-only network compatibility is still a frequent issue
• Users continue reporting intermittent disconnects in Sonoff TH-series devices
• Sensor-related instability is still observed in temperature-control use cases
• Open firmware can improve reliability, but only if correctly configured

A realistic engineering takeaway is this:

• If the unit merely shows “offline,” suspect network/pairing first
• If it freezes and needs hard reset, suspect sensor or power hardware first

That distinction matters.

---

Supporting explanations and details

Recommended diagnostic sequence

Use this exact order:

Step	Action	Why it matters	What the result means
1	Remove sensor	Eliminates probe/cable fault	Stable now = bad sensor or sensor interface
2	Remove relay load	Eliminates load noise/inrush/wiring issue	Stable now = load or mains wiring problem
3	Factory reset	Clears bad pairing/config state	Stable after re-pair = config/network issue
4	Re-pair on 2.4 GHz only	Eliminates band steering / router issues	Stable now = Wi‑Fi compatibility issue
5	Test near router or hotspot	Separates RF issue from device fault	Still unstable = likely internal fault
6	Inspect internal PSU	Finds brownout cause	Bad caps/regulator/solder fault likely
7	Reflash firmware if applicable	Fixes corrupted flash	Helps only if hardware is healthy

---

Factory reset procedure

For TH10-family devices, long-pressing the button for about 5 seconds typically places the device into pairing mode. Deleting it from the app may also effectively return it to a fresh setup state.

Recommended process:

1. Delete device from eWeLink
2. Power cycle it
3. Hold button about 5 seconds until pairing LED pattern appears
4. Re-add it on a 2.4 GHz SSID
5. Do not connect sensor or load during first test

---

Power integrity explanation

The most useful electrical model here is:

\[ V{rail}(t) = V{nominal} - I{burst}(t)\cdot Z{supply}(f) \]

If the supply impedance rises because electrolytic capacitors age and ESR increases, then during Wi‑Fi transmit bursts:

\[ \Delta V = I_{burst} \times ESR \]

Even a modest ESR increase can create enough droop to destabilize a 3.3 V ESP8266 system.

This is why a unit can:

• boot,
• appear normal for a few seconds,
• then lock when radio activity increases.

---

What to inspect if you open the unit

Only if you are experienced and the unit is fully isolated from mains.

Look for:

• bulged or dried electrolytic capacitors
• brown discoloration near hot components
• cracked solder joints
• regulator overheating
• carbonized contamination or moisture ingress

Key suspect areas:

• low-voltage filter capacitor near logic supply
• primary-side bulk capacitor
• 3.3 V regulator path
• sensor connector solder joints

If you have instruments:

• check the 3.3 V rail
• observe whether it dips during Wi‑Fi startup
• check ripple and transient droop

A multimeter may miss short dips; an oscilloscope is much better.

---

Ethical and legal aspects

Electrical safety

This is a mains-connected device. Opening it or probing the live side without proper isolation is hazardous.

• Disconnect mains fully before opening
• Do not connect a PC USB-to-TTL adapter while the TH10 is on mains
• Use only 3.3 V serial methods with the board isolated from AC power

Regulatory and compliance considerations

If used in a fixed installation:

• repairs may affect compliance or insurance acceptance
• replacement may be preferable to board-level repair in safety-critical use
• enclosures and creepage distances must not be altered

Reliability and safety of control applications

If the TH10 controls heating, pumps, or safety-relevant loads:

• do not rely on a freezing unit
• implement fail-safe behavior
• verify relay default state after reboot
• consider external thermal cut-out or independent protection

---

Practical guidelines

Best immediate actions

Perform these exact tests:

1. Power off the TH10
2. Disconnect the temperature/humidity sensor
3. Disconnect the load from output
4. Power it with only mains input
5. Watch whether it stays alive for at least 10–15 minutes
6. If stable, reconnect sensor only
7. If still stable, reconnect load only
8. Reproduce the failure step by step

This isolates the fault cleanly.

---

If the problem is Wi‑Fi related

Do this:

• dedicated 2.4 GHz SSID
• WPA2-Personal preferred
• disable band steering for test
• keep router nearby
• assign DHCP reservation
• re-pair after deleting old device entry

---

If the problem is hardware-related

Most likely repair paths:

• replace low-voltage electrolytic capacitor with quality low-ESR 105°C part
• inspect primary bulk capacitor
• inspect/regenerate suspect solder joints
• verify 3.3 V regulator stability
• if flash or ESP8266 is damaged, replacing the whole unit is usually more economical

---

If using custom firmware

If flashed with Tasmota/ESPHome or similar:

• erase flash fully
• reflash a known stable build
• verify TH10 GPIO mapping
• disable unsupported sensor configuration
• test with minimal config before restoring automations

---

Possible disclaimers or additional notes

• Not every “disconnect” is a network problem; freezing is a stronger sign of hardware or firmware instability.
• Forum reports are useful but not always definitive.
• Specific firmware-version issues reported by users may apply only to certain sensor models or custom firmware combinations.
• If the unit is old and the fault is getting worse, replacement may be more time-effective than repair.

---

Suggestions for further research

If you want to go deeper, the next useful technical checks are:

• measure the 3.3 V rail during startup and Wi‑Fi association
• inspect capacitor ESR rather than capacitance alone
• collect UART boot logs from the ESP8266
• compare behavior with and without sensor polling enabled
• verify whether the relay switching event coincides with lockup

For a serious engineering diagnosis, the most informative data would be:

• whether the sensor is connected
• whether a load is connected
• whether it freezes on stock firmware or custom firmware
• LED blink behavior before hang
• whether it stays powered but simply goes offline, or becomes completely unresponsive

---

Brief summary

Your Sonoff TH10 most likely has one of three real faults:

• bad sensor / sensor cable
• bad Wi‑Fi pairing or 2.4 GHz compatibility issue
• internal power-supply instability, especially aged capacitors or 3.3 V brownout

The most probable hardware cause for a unit that truly hangs and needs a hard reset is power rail instability.
The first practical test is to run it with no sensor and no load. If it still freezes, the device itself is faulty.

If you want, I can give you a step-by-step fault tree for:

1. stock eWeLink firmware, or
2. board-level repair and measurement points on the TH10.

Disclaimer: The responses provided by artificial intelligence (language model) may be inaccurate and misleading. Elektroda is not responsible for the accuracy, reliability, or completeness of the presented information. All responses should be verified by the user.