Can't read the print on the PCB too well but that's the 8 pin chip near Q4?
The design is a flyback type which is much simpler than that LG one, the chip is probably something like a UC3842 and I see another primary side 8 pin chip on the left side of the board picture - is it the same?
Not going to be easy as, just going by the pad connections, it's not the same chip as the OB5269CP.
The power circuit itself as fairly simple and also is where all the damage occurred so, as long as the destruction of the FET didn't get HV onto the gate circuit, the IC is likely to be OK. The chip could be verified the same way we used on the LG power supply with an external 12V power supply. Just need to identify the ground and Vcc pins first.
The typical arrangement is to get the power to start chip up via a large value resistor (about 50-100k) off the 340V rail, switching starts and energises an auxiliary winding on the transformer which has one end grounded (same ground as the IC) and the other end rectified via a fast diode then filtered by a smallish electrolytic around 100uF 25V or so to provide the running Vcc supply. With no mains on the board, use external supply to make the chip run and check output (there's only one) to the FET gate for pulses. That auxiliary Vcc supply is also used as the voltage reference for regulating the output which removes the requirement for optos to feed back from the secondary but that simplicity comes at the expense of less precise regulation of the main secondary outputs.
The 340V supply is connected to the transformer primary, the other end of the primary goes via the FET with a resistor between Source and ground. The voltage drop across that resistor provides cycle by cycle current sensing to provide overcurrent protection for the FET (by shortening the PWM pulses). That resistor will be something like 0.2-0.5ohm 2-5W and the fault current had to go through that as well so check it out carefully.
There's a fair bit of snubbing circuitry across the primary and the FET to absorb switching spikes but it should be ok in this case.
Can't be certain of anything going by the board pics but it looks like the FET that failed is for a standby supply, the OB5269 chip has opto feedback suggesting it's the main supply switched on by the standby which has self destructed. The main reason for that deduction is there's only a single secondary winding on the transformer near the failed FET so you could, potentially, fake the standby supply if you knew what voltage was required on its secondary side to see if the set functions.
EDIT to previous post: there's a small transformer (the yellow taped one) across the isolation boundary which could be the feedback for voltage control of the failed supply/
2 photos to compare how the 2 are set up.
The one with the unknown chip is also the blown FET.
The second photo is turned around for a better direct comparison.
My guess is that the chip is ok simply because it wasn't burnt. A lot of the components have a varnish on them that looks like it was just brushed thick and lumpy. Originally I could read some of the numbers on the chip but not enough to identify it and when I started chipping off the varnish, the numbers went too.
From memory I worked out the transistor (Q1) is a 2N3906. The gate resistor (R7) appears to be 2R something I think, and I know what the FET is....The zener is prob the big question, can't read it at all.
Last edited by loopyloo; 26-08-17 at 12:51 AM.
Still a bit hard to sort out from a pic but the gate drive appears to be off pin 7, Vcc on pin 8, current sensing appears to be via what appears to be a 1W resistor on the component side (colour bands uncertain and could be heat damaged). Would need to be buzzed out with a continuity tester to be sure of the connections there and that resistor R7? on the gate looks to be a mess. Is that a smd transistor near the gate (just above R7)?
The thin tracks near the FET look possibly damaged too
That zener (ZD2) will give an idea of what the max Vcc is
Last edited by Skepticist; 26-08-17 at 12:42 AM.
loopyloo (26-08-17)
The more I look at the pics, the more Q4, U4 and associated circuit is looking like a PFC arrangement.
Hard to say via this remote viewing method
loopyloo (26-08-17)
The tracks are ok just a bit scratched up from me cleaning off the carbon.
I can see now the zener is marked " WJ " so it's a 300mW 15v.
I notice the other 3 FETs have a resistor marked 000 or 0000 in the line to their gate, so if it's the same that resistor (R7) is only a fuse.
In this photo I've marked it to show components on the top side. I've left it full size so you can zoom in.
Yellow = wire link.
Blue = 2 power diodes.
Pink = large resistor at least 2 watts. Have to look closer yet cos it's burnt, might be 0.1 or 0.15 ohms.
Red = KBL408, bridge rectifier with both AC leads cut off.
You might be right about PFC.
So to summarize, assuming the chip is ok, I reckon the parts I need to replace is the 2 power caps ((120uF 450v exchange 100uF) and 47uF 450v), the FET (24N60M2), the zener (15v), the transistor (2N3906, could replace with BC327) and the resistor (R7, gamble the size, zero ohm).
What do you think ? ... That series lamp will be handy for testing.
In case you missed it, I added more info to post #25
Last edited by loopyloo; 26-08-17 at 08:12 PM.
Skepticist (26-08-17)
Those 2 power caps (450V ones) are actually in parallel so you could test with just a 100uF cap and fit the smaller one later if all looks good and ready to try it with load connected.
R7 on the FET gate: anything under 100 ohms will work fine there (I'd expect it's 22 or 47) as there's very small current flow through there (uA) and the resistor just limits the inrush current charging the FET's capacitance.
Pin 6 looks like the GND connection but need to verify that with continuity test to the supply -ve on the power cap
If the zener is either shorted or open cct the chip is likely to be damaged so definitely check that for diode action at least, if there's a short between the chip Vcc and gnd check out the electrolytic near the current sensing resistor (it's close to a heat source so its ESR could have gone high).
Plus: try to trace backwards from the chip Vcc to find its source
And: there are components under the heatsink that need to be exposed for tracing purposes
Wondered what was going on with that bridge rectifier when only 2 pins are soldered down - very odd to use series/parallel diodes like that
I'd be verifying that the chip is working with the external 12V supply before taking a chance on applying the mains to the board. Mushroom smoke clouds are a bad thing
Last edited by Skepticist; 26-08-17 at 09:23 PM.
If the U4/Q4 circuit is actually the PFC it could be taken out of circuit the same as with the LG board by simply lifting a few jumpers then checking that the HV bus still sits at 340V.
From what I can see, the Vcc appears to be coming from the other SMPS chip area (need verification with a buzzer)
loopyloo (26-08-17)
U4 looks like it's a OB6561 active PFC controller just going by the pin assignments on the PCB
data
Just a hunch that those chips are both from On-Bright
ST equivalent L6561 datasheet
Last edited by Skepticist; 26-08-17 at 10:07 PM.
loopyloo (26-08-17)
Yeah I removed the heatsink this afternoon which means removing all 4 FETs at the same time (they are screwed from underneath).
Only other thing under there is a 2 or more watt resistor 0.39ohm and 3 wire links.
There are a few sample application circuits in the ST datasheet which will make tracing it out a lot easier to follow
Those wire links are the Vcc track as far as I can tell
The large resistor is the current sensing for the SMPS FET (between source and gnd)
Last edited by Skepticist; 26-08-17 at 10:17 PM.
loopyloo (27-08-17)
Been sketching out that chip (not finished) but you can have a look at where I'm up to.
Never been good at drawing out the circuits. Might be enough info so far for you to work out what's going on.
Last edited by loopyloo; 29-08-17 at 11:22 AM.
Definitely a PFC and almost certainly the chip is the 6561
The source-gnd resistor is selected to produce a 1.7V drop across it at a current well within the rating of the FET so the chip can shorten the gate pulses. O.1 ohm appears to way too low as the trigger for overcurrent would be 17A. 0.33ohm would make more sense there (current spikes limited to 5.2A), 0.39ohm would be even safer.
The gate resistor would be something like 10 ohm or slightly more.
The chip's current demand is so low that the Vcc can be derived from the HV supply via a resistor (or string of resistors), Vcc should be around 13-18V and is internally limited by a 20V zener within the chip (12-13V is the chip's turn-on threshold, turn-off about 10V).
If the PFC was completely disabled the HV voltage would be around 340V. The PFC alllows a much wider operating mains voltage range _eg 100-265V~ but since we're already on 240V it's not doing much apart from cleaning up some harmonics so you can get by without it.
Last edited by Skepticist; 29-08-17 at 12:25 PM.
I'd say in this case removing the PFC sounds like a good idea, I like it, and no need to buy any parts except for those 2 mains caps.
Is it just a matter of cutting some tracks ? Leave the FET space as empty, remove the sense resistor and cut pin 8 receiving from the other chip ?
Last edited by loopyloo; 29-08-17 at 02:09 PM.
Just remove either of those 2 double jumpers you've coloured in yellow and the only power to the main caps will then be untouched rectified mains at 340V DC through that modified bridge coloured in red.
The PFC chip can be disabled by connecting pin 5 to GND or just breaking that Vcc track.
loopyloo (29-08-17)
Original caps were 120uF and 47uF 450v.
I've got 2 100uF 450v caps I could put in so should be good.
I'll get into that now.
Has the other SMPS chip & circuit checked out OK (prior to attempted power-up)?
Just being cautious
Bookmarks