[Solved] VIPer53 chip (PSU) blown in an air conditioned unit

[nuno]

Hello all,

I'm trying to fix an air conditioned unit which is totally silent. I opened it up and found this VIPer53 chip from ST (datasheet) blown. This is a "PSU controller", on the AC-DC converter. I see no other components with fault signs nor blown fuses.

My question is, do you think something else may be blown? I've tried to fix another AC-DC converter once, that had the controller chip blown, but it kept blowing as soon as I plugged the "fixed" board to mains (in this converter I think the problem was the transformer, the ferrite was cracked, which probably changed the transformer characteristics).

Many thanks for any input you may have.


Update: corrected chip's name
Update: added "[Solved]" to the post's title and here's the link to the closing post.

[RayRay]

My question is, do you think something else may be blown?

Probably.
Looks can be deceiving, and you should never presume that just because a component looks okay, then it really is!
Everything needs to be properly tested.

[nuno]

Yes... I've been poking around the chip, and there's a diode that seems to be open. SMD in the most tight place, of course, and with that "hard yellow epoxy" hiding more components. Humm :\ ... There are also some "diode-looking" components that kind of behave like diodes, maybe they have something in parallel, but I've never seen such diodes... on the forward direction they drop only a few 10 tenths of mV ... I'm going to try to find the schematics.

[strawberry]

Check Viper53 Drain pin RCD clamp network capacitor for droped capacitance ( this fault will blow it instantly) if there is not capacitor then ~180V zener clamp.
Check if VDD bias voltage circuitry is ok Dried electrolitics or damaged small rectifier diode
Check mains electrolytic, may cause Viper chip to overheat at nominal load
If that pasesd and there is clicking noise/pulsating voltage present, check secondary capacitors and feed back circuitry.

[nuno]

I removed the dead VIPer53 chip and did some measurements and reverse eng. on the converter block, with the help of VIPer53's datasheet application example; schematics below.

Except for the electrolytic caps which I haven't measured yet for ESR (don't have an ESR-meter, but have fungen & scope so will try next), the only part I think is broken (besides the Viper chip) is D34. This diode isn't in the example application and is never suggested by the datasheet. It's a SOT-23 part with the marking WT3, then 98 in vertical text; some sites say it's a BZX84-C2V4 zener, but the datasheets say a different marking for this part in this package... any suggestions?

The RCD clamp network looks good, all 3 parts measure well.

The VDD circuitry is more complex, with a snubber for D38 (a fast diode), alternate path through D45/R157 (power the Viper at startup while the big 220uF C22 charges?) and a 7805 to power something else in the board. D44 and D45 measured funny (0.6V fw, 1.9V reverse), D38 measured shorted; then I removed D38 from the PCB and it measures good, as well as D45 which also measures good now. I suppose the loop through the transformer and 7805 affects the diode test; D44 still reads funny... I was now looking again at the schematics as I write this and I don't see a reason for D44 to read funny...
By Vfw and physical aspect, D44 and D45 are probably 1N4148 or similar.

Thanks!

update: replaced the schematics by a "tidier" one, same functionality
update: corrected chip's name

UPDATE: See the updated schematics further down in this thread: schematics
2019/06/17: Schematics updated by biank88, see his post below.

[noidea]

Post a couple of photos of the board, component and solder sides please

[nuno]

Everything is crammed in tightly and conformal coated and there's "hard epoxy" holding the caps and almost embedding some of the SMD comps. Had a hard time trying to reach the SMD components with the DMM's probes 

[nuno]

Ok, so I found another faulty component. I measured ESR of the electrolytic caps around the transformer's primary (T3) and found C91 (2.2uF 100V) of the compensation network with 21 Ohm ESR. If it were this alone, it wouldn't probably matter as there's a 2K resistor in series, but then I remembered that the capacitance is also probably reduced, which was confirmed: it measured 1.2uF. Same area as the busted (supposed) 2.4V zener diode. I compared the readings on this capacitor (already out of the circuit) with a 25yo unused 2.2uF 50V electrolytic, which gave 2.2uF and 3.2 Ohm ESR (both tested at 200KHz with a fungen and a scope).

By biggest doubt now is the diode D34; can it really be a 2.4V zener diode? VIPer's working is a bit over my head, but  the datasheet says there's a short-circuit / overload protection mechanism that kicks in when COMP reaches 4.35V, which it wouldn't reach with a 2.4V clamp in there.

COMP pin - Input of the current mode structure, and output of the internal error amplifier. Allows the setting of the dynamic characteristic of the converter through an external passive network. The useful voltage range extends from 0.5V to 4.5V. The Power MOSFET is always off below 0.5V, and the overload protection is triggered if the voltage exceeds 4.35V. This action is delayed by the timing capacitor connected tothe TOVL pin.

On the other hand, my experience with small value zeners is that they don't have a very sharp knee.

[nuno]

Except for the electrolytic caps which I haven't measured yet for ESR (don't have an ESR-meter, but have fungen & scope so will try next), the only part I think is broken (besides the Viper chip) is D34. This diode isn't in the example application and is never suggested by the datasheet. It's a SOT-23 part with the marking WT3, then 98 in vertical text; some sites say it's a BZX84-C2V4 zener, but the datasheets say a different marking for this part in this package... any suggestions?

Ahh!!... I misread the datasheet, there's in fact a marking WT3 on SOT-23 meaning it's a BZX84-C2V4 (China made, 5% tol 2.4V zener).

[nuno]

Check Viper53 Drain pin RCD clamp network capacitor for droped capacitance ( this fault will blow it instantly) if there is not capacitor then ~180V zener clamp.
Check if VDD bias voltage circuitry is ok Dried electrolitics or damaged small rectifier diode
Check mains electrolytic, may cause Viper chip to overheat at nominal load
If that pasesd and there is clicking noise/pulsating voltage present, check secondary capacitors and feed back circuitry.

After replacing the chip and 2 damaged components around it, it's no longer "dead"... now it's clicking
Next stop is to look at the sec. caps and feedback circuitry then. Another PCB reverse eng. task

[stj]

if it's pulsing, check the output rectifier diodes.
also any small electrolytics on the primary side.

[nuno]

if it's pulsing, check the output rectifier diodes.
also any small electrolytics on the primary side.

Electrolytics on the primary are covered. So adding the output rectifiers to the check list, thanks... This board is a big PITA to work on, no schematics and this part of the circuit is a mix of SMD with big electrolytics, I sometimes don't even have space to put a DMM's probe nor the soldering iron, had to remove a cap or 2 just to reach certain parts   

[stj]

i made probe extensions for that scenario using large medical needles.

you can of course buy long thin probes - but who wants to wait for the post! 

looking at your photo, you need to remove the brown glue - it slowly becomes conductive!!
(it's also acidic)

[nuno]

i made probe extensions for that scenario using large medical needles.

Well insulated, I hope !

you can of course buy long thin probes - but who wants to wait for the post! 

looking at your photo, you need to remove the brown glue - it slowly becomes conductive!!
(it's also acidic)

Jeeez.... any safe alternative to mechanical removal?

[stj]

unfortunatly no, small flat screwdriver + long-nose pliers is my method. 

[nuno]

Humm... not sure if I'm going to take the risk... the stuff is rigid and some of it is hiding SMD parts

[wraper]

and found C91 (2.2uF 100V) of the compensation network with 21 Ohm ESR.

[/quote]
for 2.2 uF it's OK-ish ESR.

[nuno]

Do you think so? My 25 year old unused 50V 2.2uF electrolytic has 3.2 Ohm... and besides, the 2.2uF cap measured 1.2uF, 45% down from nominal value, way out of spec. I just replaced it. The 2V4 zener in "the same area" was also fried (open).

[nuno]

It's not clicking in my bench (according to A/C technician it was clicking when installed at the A/C unit...).
The circuit has no "explicit" feedback from secondary to primary side (uses the transformer). There doesn't seem to be any ruined capacitor nor rectifier. Measurements at the secondary after the rectifier show that when I plug it in, it looks like it's trying to raise the voltage but stays only at about 3 - 4V (should be at least 12V), then after some time it kind of quits and the voltage slowly drops to zero; viper53 gets quite hot, so I suppose it's thermal protection kicking in. I think my next step is to remove the transformer in order to better check all components around... any other suggestions?

[wraper]

Do you think so? My 25 year old unused 50V 2.2uF electrolytic has 3.2 Ohm... and besides, the 2.2uF cap measured 1.2uF, 45% down from nominal value, way out of spec. I just replaced it. The 2V4 zener in "the same area" was also fried (open).

Yes I do. It's on the high side but even some new, especially ultrasmall capacitors are like this.
3.2 ohm is on the low side. What I do think, you have a cheap LCR meter from ebay. They don't use 100kHz sine wave to measure capacitance, as it is in decent meters. But some sort of charging/discharging IIRC. It may result in wrong measurements, especially if capacitor is leaky or you try measuring in circuit.

[nuno]

I measured ESR (on both caps, and others too) with a function gen and a scope, outside of the circuit.

[nuno]

I think I'm starting to get somewhere. The DC-DC where this VIPer53 goes supplies power to several sub-circuits, both with the converter's secondary but also with his primary. Looks like one of the primary fed sub-circuits has a reverse power supply polarity protection diode in parallel (still only a suspect), and this little guy is shorted (had to cut a few PCB traces to find it). I repaired the converter but this short is still preventing it from working, having the behavior I described above (unable to keep output voltage, then thermal protection kicking in). This is going to take me time... the little b@st@rd is under a huge heatsink where a bunch of stuff is attached

[noidea]

I think I'm starting to get somewhere. The DC-DC where this Viper53 goes supplies power to several sub-circuits, both with the converter's secondary but also with his primary. Looks like one of the primary fed sub-circuits has a reverse power supply polarity protection diode in parallel (still only a suspect), and this little guy is shorted (had to cut a few PCB traces to find it). I repaired the converter but this short is still preventing it from working, having the behavior I described above (unable to keep output voltage, then thermal protection kicking in). This is going to take me time... the little b@st@rd is under a huge heatsink where a bunch of stuff is attached

Have you had a good look at the components loads that are connected to the board? you are way more advanced than I am as far as component level fault finding and repairs but I also live breath eat sleep inverter driven AC units and 99 times out of a 100 the SMPS is the innocent victim and it's what's connected to it that has caused the failure.

[nuno]

Man, this is by far the most complex thing I've ever tried to repair!
Indeed... even with "thermal protection", "overcurrent protection" and all that... the guys still fail under a bad (shorted) load.
I couldn't find a schematic for this board (an ELECTRA outdoor DC inverter controller from an Airwell DCI 50 A/C, with just a few years! They don't sell the board anymore), so I've been reverse-engineering the PCB as I go (and finding some mistakes on my part, the partial schematic I left above is not 100% correct, I'll have to update it later). There's a few loads fed by the auxiliary winding (from C22), spread all over the PCB. Resistance from C22+ to C22- is below 0.5 Ohm. To find the responsible I did a binary search cutting the positive rail until the culprit zone was isolated. I isolated one leg of the diode, but since I can't see the other side of the board yet because of the heatsink, I don't know if there's something else connected there...
After all the work, I hope it's just the diode... I've been thinking and I may be able to cut it out without removing the heatsink, then I'll know... and being a small diode, I'll put a replacement on the unobstructed side of the PCB (nice that the designer put silkscreen on both sides of the PCB, at least for most TH components; on the other side... there's things like vias going to nowhere ).

There's one other "little thing" that worries me a bit, but I don't know if it's normal. The VIPer53 runs a little hot even with a very small load... it stabilizes near 60ºC. The PCB all around is darkish, even the box cover on the copper side of the PCB has an isolation plastic that is darker in that area, so I'm not sure this really is a "bad" design or there's another problem waiting for me to fix the diode.

[noidea]

Please post a photo of the whole board component and solder sides as well as the block wiring diagram that was on the unit.

As per previous post what have you done to check the loads connected to the board?