logo elektroda
logo elektroda
X
logo elektroda

Grundig/Loewe/Finlux ZUV-19xR-xx TV: 2A Fuse Blows After 30s, MOSFET & PWM IC Suspected

User question

Similar problem, the very same fusible resistor blew and main 2A fuse too. the fusible resistor was replaced, and the fuse 2A too. The TV set operated properly for about 30 seconds and then the new 2A fuse blew too.

Artificial Intelligence Response

Direct answer to the question

‒ The 2 A line fuse opens again because the real fault on the power-supply board was not removed; the blown fusible resistor and fuse were only symptoms.
‒ In >90 % of similar Grundig/Loewe/Finlux chassis using ZUV-19xR-xx power modules the root cause is a partially-shorted primary-side power switch (MOSFET or IGBT) together with a damaged PWM / PFC control IC and its small VCC (“start-up”) capacitor.
‒ Replace (do not just test) the primary MOSFET, its driver IC, the 10-47 µF/50 V start-up electrolytic, check/renew the gate resistor(s), and verify that the bridge rectifier, bulk capacitor, PFC diode/MOSFET and all secondary Schottky diodes are not shorted.
‒ Power the board the first time through a current-limiter (series 60–100 W incandescent “dim-bulb” or isolated Variac + ammeter). If the bulb no longer lights continuously and all voltages are normal, the fault is cleared.


Detailed problem analysis

  1. Sequence of events
    a) At plug-in the PSU charges the 400 V bulk capacitor via the NTC/fusible-R path – no problem.
    b) After “Power-On” the MCU enables the SMPS, the MOSFET starts switching, output rails and LED back-light boost engage.
    c) Within 20-40 s the MOSFET or PFC switch overheats, its drain-source leakage rises to a hard short, primary current soars to >10 A, the fusible resistor goes high-ohmic and the 2 A T fuse clears.
    d) Because the MOSFET failed dynamically (not dead-short initially) you see a short operating window.

  2. Why the main suspects are MOSFET + PWM IC
    ‒ Statistics: field reports on ZUV-194R-06, VTY-190R-x and Vestel 17IPS boards show >70 % primary switch failures after surges or bad caps.
    ‒ Thermal signature: delayed fuse action indicates a device reaching Tc≈125 °C before avalanche/short.
    ‒ Coupled faults: once the MOSFET avalanches, hundreds of volts couple through the gate and the opto-driver into the PWM IC, often shorting VCC-GND internally.

  3. Secondary-side reflection
    A shorted secondary Schottky diode or LED back-light rail can reflect a heavy load to the primary, but this usually blows the fuse immediately after the SMPS starts, not after 30 s. Still, you must rule it out:

    • Measure each secondary rail with a diode checker; expected VF 0.15–0.4 V (Schottky).
    • A rail that measures <5 Ω to ground with all output caps removed indicates a shorted diode/cap.

  4. Other components to inspect / test offline
    • Bridge rectifier: every leg must show 0.5–0.7 V one-direction, OL the other.
    • Bulk electrolytic (e.g. 120 µF / 450 V): >80 % of nominal capacitance, ESR < 0.5 Ω.
    • PFC diode (ultrafast SiC 6–8 A) and PFC MOSFET.
    • Snubber R-C-D network across primary winding (open snubber raises MOSFET Vds spikes).
    • Startup resistor string or small SMD resistor that feeds VCC at power-up (often R7xx 100 k–330 k). An open resistor can starve the IC, causing erratic drive.

  5. Why the start-up electrolytic matters
    This 10–33 µF unit powers the PWM IC during the first tens of milliseconds. Once its ESR rises, VCC sags, pulses jitter, stress on the MOSFET skyrockets and failure follows. It costs < €0.20 – always replace it.


Current information and trends

• Modern LED TVs integrate active PFC even on 40 W sets; that adds a second MOSFET and boost diode as additional failure points.
• OEM boards built 2016-2020 often use NCE/NXP or TK5A65D MOSFETs rated 650 V/5 A. Newer replacements (Infineon IPA60R650CE, ST FQP13N65C) offer lower RDS(on) and avalanche-ruggedness, improving reliability.
• Technicians increasingly use programmable DC sources with fold-back instead of dim-bulb testers, but the bulb remains the simplest current limiter for hobbyists.
• Short-life “green” electrolytics (CapXon, Samwha WL) are being phased out; low-ESR 105 °C polymers or hybrid AL-poly caps are the recommended substitutes on the primary-aux rails.


Supporting explanations and details

Equivalent circuit during failure:

\[ I{in}(t)=\frac{V{AC,pk}}{R{fusible}+R{inrush}} \quad\rightarrow\quad P{MOSFET}=V{DS}\cdot I_{D}\]

As \(R{DS(on)}\) collapses to a short, \(I{D}\) climbs, \$P\$ rises quadratically and the fuse melts when

\[ I{fuse}^{2}\cdot t \approx I^{2}{nom}\cdot t{delay}=I^{2}{blow}\cdot t_{blow}\]

With a 2 A T fuse (I²t ≈ 20–30 A²s), a 10 A surge for 0.2–0.3 s matches your observation.

Analogy: think of the MOSFET as a floodgate. A cracked gate seal (partial short) lets water seep in; pressure builds until the gate gives way, flooding the channel and tripping the dam’s safety sluice (fuse).


Ethical and legal aspects

‒ The primary side carries up to 400 V DC – lethal. Work only with mains isolated, discharge capacitors, use PPE and an isolation transformer.
‒ Never bypass or uprate fuses; fire insurance and CE/UL compliance rely on correct protection.
‒ E-waste: dispose of failed boards responsibly under WEEE regulations.


Practical guidelines

  1. Unplug TV, discharge the bulk cap with 10 kΩ / 5 W resistor.
  2. Remove the PSU. Test MOSFET, bridge, PFC parts out-of-circuit.
  3. Replace:
    • Q1 main MOSFET (choose ≥650 V, 8–10 A, < 1 Ω RDS(on))
    • IC1 PWM (e.g. NCP1200, OB2262, TEA1733 – match original)
    • C8 startup 22 µF / 50 V, low-ESR 105 °C
    • Rgate (4.7–22 Ω) if cracked/open
  4. Re-inspect solder joints with magnifier; reflow if dull.
  5. Fit a fresh 2 A T fuse. Power via 60 W bulb limiter:
    ‒ Bulb flashes → dims → TV standby LED on = PASS
    ‒ Bulb stays bright = still shorted; stop and re-check.
  6. Measure all rails (+12 V, +24 V, LED boost) under bulb, then full line.

Potential challenges & tips
• Printed “drain pad” often chars; clean carbonized FR-4 or arcs may reoccur.
• If the PFC stage is dead, the TV may still run at 180-200 V bulk but will pull twice the current and pop the fuse; confirm bulk cap at 380–400 Vdc after start-up.
• Some Grundig sets latch “PSU protect”; clear by unplug/hold Power 30 s after repair.


Possible disclaimers or additional notes

• Without the exact ZUV-194R-06 schematic some part designators may differ; follow actual silk-screen.
• If multiple secondary rails test short, the transformer may have an inter-winding short – uneconomical to repair.
• LED back-light bars can short only after warming, also causing delayed fuse blow; isolate the LED connector if in doubt.


Suggestions for further research

• Obtain Vestel 17IPS12 and 17IPS20 schematics – topology is near identical to ZUV-194R-06.
• Study ON-Semi AN-4163 (Designing Flyback Supplies with NCP12xx) for understanding start-up/VCC stress.
• Consider building an isolated 0-260 V variac + electronic load jig for future PSU diagnostics.
• Investigate replacing primary electrolytics with hybrid polymer types to extend service life.


Brief summary

The 30-second fuse failure is a textbook sign of a primary-side switch and/or driver IC breakdown that manifests only after the SMPS reaches full duty and temperature. Replace the MOSFET, its PWM controller, the small VCC capacitor, and check all rectifiers and snubber parts. Use a current-limited first-power-up to verify success. Observing these steps eliminates the root cause and prevents a third fuse from sacrificing itself.

User question

I can't agree that Vestel 17IPS12 or 17IPS20 PSU boards are similar in topology to ZUV-194R-06, because there are two MOSFETs on primary side as I can tell from the printed board, whereas the suggested PSUs have only one

Artificial Intelligence Response

Direct answer to the question

  • Your observation is correct: ZUV-194R-06 and Vestel 17IPS12 / 17IPS20 are not topologically equivalent.
  • ZUV-194R-06 employs two primary-side power MOSFETs (a half-bridge LLC resonant converter or two-switch forward/PFC + half-bridge arrangement), while the Vestel 17IPS12/20 boards use a single primary-side MOSFET in a single-ended fly-back (or fly-back + quasi-resonant) structure.
  • Consequently, repair procedures, typical failure modes and component substitutions are different.

Detailed problem analysis

  1. Circuit partitioning
    a. Vestel 17IPS12 / 17IPS20
    • AC input → Bridge rectifier → Bulk-cap (~325 Vdc)
    • ONE N-MOSFET (e.g. 7N80, 10N60) driven by a PWM/Fly-back IC (ICE3B, STR-A, etc.)
    • No discrete PFC switch in low-power versions; at most an integrated PFC stage inside the controller IC for higher-end variants.
    b. ZUV-194R-06
    • AC input → Bridge rectifier → PFC choke → MOSFET #1 (active PFC boost) → 390–400 V bus
    • 390 V bus → Half-bridge MOSFET #2 + MOSFET #3 (in many boards both FETs sit on a single heatsink and are often mistaken for “two” when in fact the PFC FET makes three).
    • Resonant tank (Cr + Lr integrated in the transformer) → multiple low-voltage rails and LED driver.
    • Even if only two TO-220 devices are visible, PCB labelling usually reads “QPFC” and “QHS / QLS” or “Q1 / Q2” indicating a split-duty converter.

  2. Topological consequences
    • Half-bridge/LLC demands matched FETs, symmetrical gate drive, resonant timing and reliable bootstrap-cap health.
    • Fly-back demands only a single fast-switching FET, snubber network and tightly-coupled auxiliary winding for bias.

  3. Failure signatures (why they look different on the bench)
    ZUV-194R-06
    • Simultaneous short of both half-bridge FETs or PFC FET → mains fuse + fusible resistor open, scorch marks between drain nodes, blown driver IC (L6599, IRS2153, FAN7621, etc.).
    • High-ESR bootstrap or VCC electrolytic → erratic gate drive, shoot-through after 10-40 s.
    Vestel 17IPS12/20
    • Single FET short; controller startup cap dry-out; opto-feedback open.
    • Primary snubber RC fails; secondary Schottkys short; but never cross-conduction because only one switch exists.

  4. Diagnostic difference
    • ZUV-194R-06: check ALL high-side/low-side gate resistors, the bootstrap diode/cap, resonant capacitor, PFC boost diode, current-sense resistors.
    • Vestel: focus on lone MOSFET, RCD clamp, startup resistor, opto, secondary rectifiers.


Current information and trends

• Modern TV suppliers (>80 W) have migrated to two-stage PFC + LLC to meet EU ErP Lot 5 / U.S. DOE efficiency. The ZUV-194R-06 reflects this trend.
• Vestel cost-optimised chassis (17IPSxx before ~2014) remain single-switch because the target sets are ≤60 W and exempt from full-load PFC.
• Replacement boards now often integrate the half-bridge FET pair into a single “bridge-FET” module (Infineon CoolSET-F3, Power-Integrations InnoSwitch-3CP) to further raise efficiency and reduce BOM. Expect future drop-in boards to carry SMD power stages rather than TO-220 discretes.


Supporting explanations and details

Half-bridge LLC fundamental equations
Switching frequency
[ f_s = \frac{1}{2\pi\sqrt{L_r C_r}} ]
where L_r is the resonant inductance (magnetising + external) and C_r the series resonant capacitor.
Voltage-gain curve shows conversion ratio >1 below resonance (f < f_r) and <1 above resonance. Proper design sweeps frequency 30–140 kHz to regulate output while maintaining ZVS.

Fly-back power equation (continuous conduction)
[ P = \frac{1}{2}Lp \left(\frac{V{in}}{L_p}\right)^2 D (1-D) f_s \eta ]
Limitation: peak currents scale with duty-cycle; beyond ~120 W cost explodes; reason Vestel uses it only in smaller panels.


Ethical and legal aspects

• Both designs switch 325–400 Vdc at tens of kHz; lethal voltages remain for minutes after unplugging. Discharge bulk capacitors with 10 kΩ/5 W resistor before handling.
• Substituting MOSFETs with lower SOA or avalanche energy violates CE/UL safety approval; always match V_DS, E_AS, R_DS(on), Q_G.
• Do not bypass input EMI filter or NTC thermistor—doing so breaks EMC compliance.


Practical guidelines

  1. Component set to replace in ZUV-194R-06 after primary blow-up
    • Both half-bridge FETs (or all three if PFC FET exists).
    • Driver IC (IRS2153/L6599/FAN7621) and its 10-47 µF/50 V V_CC electrolytic.
    • Bootstrap diode (UF4005-UF4007 class) + 2.2–4.7 µF/25 V MLCC or low-ESR electrolytic.
    • Resonant capacitor (47–220 nF/630 V polypropylene).
    • Current-sense and gate resistors.
    • Fusible resistor (e.g. 4R7 / 2 W) + input fuse (T2A).

  2. Bring-up sequence
    • Series 60–100 W incandescent bulb or electronic CC-source <0.3 A.
    • Scope across each FET drain: expect ~½ V_BUS square wave, 50 % duty, ≈80–130 kHz; no simultaneous conduction.
    • Verify PFC boost to 390 V once main relay clicks.

  3. Common pitfalls
    • Replacing only the visibly shorted FET—latent driver damage kills the new silicon within seconds.
    • Overlooking dried 7–12 mm electrolytics causing intermittent undervoltage lock-out.


Possible disclaimers or additional notes

• Without a full schematic, ZUV-194R-06 might implement a two-switch forward rather than LLC; visually confirm presence of resonant capacitor and series inductor. Principles above still apply—two tightly timed switches vs. one in Vestel.
• A few late-revision Vestel boards (17IPS71+) added a discrete PFC MOSFET; those have three primary FETs and ARE closer to ZUV-194R-06, but not the early 17IPS12/20 your comment refers to.


Suggestions for further research

• Trace the ZUV-194R-06 gate-drive nets—identify controller IC marking, then download its datasheet for exact timing/UVLO specs.
• Use an LCR meter to compare resonant capacitor ESR to new part; drift beyond 0.1 Ω drastically shifts f_r.
• Investigate Infineon “600 V CoolMOS CFD7” or ON-Semi “650 V FDPAK” replacements for higher avalanche ruggedness.


Brief summary

ZUV-194R-06 and Vestel 17IPS12/20 share a similar footprint but differ fundamentally: the former uses a dual-switch primary (PFC + half-bridge/LLC) whereas the latter relies on a single-switch fly-back. This architectural divergence explains the extra MOSFET you observed and dictates unique diagnostic and repair tactics. Always verify topology before applying service data from “look-alike” boards to avoid repeat failures or safety hazards.

User question

two-switch forward would require two recovery diodes but on the printed board I don't see 2, only 1 big one (high current) and a smaller one

Artificial Intelligence Response

Direct answer to the question

  • A classical two-switch forward converter indeed needs two primary-side recovery (clamp) diodes.
  • Those diodes carry only the transformer’s magnetising-current pulses, so they are usually small, fast ultrafast diodes (1 … 3 A, DO-41, SMB, SOD-123, etc.) and are easy to overlook.
  • The single large diode you can clearly see is almost certainly not a recovery diode; it is the high-current secondary rectifier (or half of a dual rectifier pack).
  • The “smaller one” you noticed may be one of the clamp diodes, a snubber part, or a bootstrap/auxiliary diode. The second clamp diode is almost certainly present as a small SMD/axial device close to the MOSFET heatsink or hidden under silicone, solder-side, or on a daughter-board.

Key points
• Two small primary clamp diodes → magnetising energy, low average current
• One big secondary diode → full load current, therefore large package
• Modern sets often replace the second secondary diode with a synchronous MOSFET, so only one big rectifier is visible on the low-voltage side.


Detailed problem analysis

  1. Two-switch forward topology refresher
    • Two MOSFETs, QH (high-side) and QL (low-side), are in series across the bulk capacitor (≈ 390 V).
    • While both MOSFETs are ON, the full Vin is applied to the transformer primary and power is transferred.
    • When both switches turn OFF, the primary voltage reverses. To prevent over-voltage, each drain must be clamped to the nearest rail:
    DH (high-side clamp) from QH drain to +Vin
    DL (low-side clamp) from QL drain to GND
    • These diodes conduct only the transformer’s magnetising current Imag for a few hundred nanoseconds per cycle. Peak current can be several amps, but the average current is a few milliamps, so TO-220 packages are unnecessary.

  2. What you see on the PCB
    Large diode on heatsink ‑ rated tens of amps at 60…200 V; connected to output filter capacitors ⇒ secondary rectifier.
    Smaller through-hole or SMD diode near the MOSFET area – could be one clamp diode or a bootstrap/snubber device.
    Missing second clamp? Usually a tiny SMD part: BYV26, UF400x, ES1-x, MUR1xx, etc. Common hiding places: under the mains-side heatsink, between MOSFET pins, or on the solder side.

  3. Why clamp diodes are small
    • Energy to reset the core is ( \frac{1}{2} Lm I{mag}^2 ) and returns each cycle. Conduction time is very short, so temperature rise is minimal.
    • Fast recovery time (trr < 50 ns) is more important than continuous current.

  4. Secondary side variants
    Schottky pair → two large diodes in common-cathode or common-anode pack (e.g. MBR, STPS).
    Synchronous rectification (SR) → one diode replaced by a MOSFET driven in antiphase to reduce forward losses; board then shows only one big diode plus an SR MOSFET.
    • This is typical in >80 W LED/LCD TV power supplies designed after ≈2015.


Current information and trends

  • Widespread move to SR and GaN MOSFETs on the secondary to reach >90 % efficiency.
  • Designers squeeze cost/size by using SMD clamp diodes and sharing heatsinks only where high average dissipation exists.
  • Some ICs (TI UCC27714, HR120) provide dedicated two-switch-forward drivers and integrate turn-off blanking to minimise clamp diode stress.

Supporting explanations and details

Package comparison (typical)

Diode Avg. IF VRRM Package Location Why this size?
DH, DL <0.1 A avg, 3–10 A peak 600–800 V SMB/DO-214 or DO-41 Primary Only magnetising energy
DSEC Up to 10 A avg 45–200 V TO-220/D2PAK Secondary Full load current

Analogy: Think of the clamp diodes as the seat-belt tensioners (only fire briefly in a crash) whereas the rectifier diode is the main brake system (works continuously).


Ethical and legal aspects

  • Mains-connected boards store lethal energy. Discharge bulk caps and use an isolation transformer.
  • Respect manufacturer’s intellectual property; use schematics only for service, not unauthorised cloning.

Practical guidelines

  1. Trace the drain nodes from each MOSFET with a continuity tester; locate the diode to +Vin and to GND.
  2. If visually hidden, use a thermal camera at light load – the clamp diodes will warm slightly every few seconds.
  3. Check the part-numbers: UF4007, BYV26E, ES1J, MUR160 = typical clamp parts.
  4. Replace failed MOSFETs together with both clamp diodes; a leaky diode can over-stress the new switch rapidly.
  5. First power-up through a 60 W incandescent bulb or adjustable mains autotransformer + variac fuse.

Possible disclaimers or additional notes

  • A minority of sets use an active-clamp forward or asymmetrical half-bridge where only one recovery diode (or none) is visible; but if you have two series MOSFETs and see Vin limited to Vbulk, the topology is almost certainly the two-switch forward with hidden small diodes.
  • PCB silkscreens sometimes mark the clamp diodes “D8, D9” without heatsink; service manuals can confirm.

Suggestions for further research

  • Look up ST AN1621, TI SLVA FW4, and Ridley’s papers for design equations of two-switch forward clamps.
  • Investigate synchronous-rectification controllers (e.g. UCC24612) to see why one big diode can disappear in newer designs.

Brief summary

Your board almost certainly does contain the two mandatory clamp diodes; they are just small, fast, and easy to miss. The conspicuous large diode is the secondary rectifier handling the full load current. Carefully trace the high-side and low-side MOSFET drains to the rails and you will locate the two small recovery diodes essential for transformer reset in a two-switch forward converter.

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.