John, The usual failure mode for these type of supplies is the filter capacitor fails short and destroys itself spectacularly. It certainly looks like that is the case here if you look carefully at the hole near its end. When it went short the fault current would have then caused that resistor to fail as a consequence.
That would be my expectation as well. A severe short in the capacitor (too severe to self-clear) can dramatically increase the current flowing through the limiting resistor causing it to overheat and fuse. And judging from the amount of heat damage it looks to have been a rather severe short leading to plasma arcing and carbon tracking exacerbating the damage. Fortunately the resistor did its job by acting as an engineered weak link and disconnecting from the mains.
Hmmm. Proper fusible resistors are designed to fail in a more controlled manner, not light up like a lamp and char everything around them. You usually just see a pinhole in the glaze at the point of failure. Looks like the damage may have been caused by prolonged arcing across the gap in the resistor foil where it "blew". So were either of the drive transistors short circuit?
Dr. Lecter I've dismantled numerous CFLs and found failed fusible resistors but they have been controlled failures, not like the one shown here. On CRT TVs they were commonly used in the secondary side of power supplies, voltages and fault currents significantly lower though. Even where used on the primary, I can't remember encountering anything like that. I think I would class this as a product fail. Wrong choice of component for the job. In my opinion the fault took way too long to disconnect and was a bit too close to causing a fire. Also while they've used an isolation slot with barrier between mains and SELV, they seem to have forgotten about that fusible resistor again. I'm surprised that arcing hadn't occurred between it's PCB traces during the failure.
It does look like that resistor has been running hot for a prolonged period, if it were to blow suddenly, it wouldn't have caused such excessive damage to surrounding components. I'd say either a rectifier diode or that very charred cap has gone short. Or the secondary side has been drawing excess current for a prolonged period causing everything that resistor to get toasty slowly cook everything around it. Those diodes have been running pretty warm too by the looks of them.
Sorry, I could not spot the point in the video where you identified the failing component. All the charring seemed to be the effect of the fault (a short circuit down the line) rather than the cause. Did you actually find the failing component?
thats what i would think,....i've seen and repaired a fair few switch mode psu blow ups in tvs in the 1990's, usually the switching device goes short, taking out a resistor or two, rarely actually popping a fuse!
Hi John. I don't particularly like the idea of a resistor that blows and incinerates the adjacent components as opposed to a fuse that goes OC. It looks like there was a fire on board, which luckily was contained by the enclosure. Are there any fire-proof safety ratings on the enclosure?
Usually these special resistors fail without drama, you normally have to closely scrutinise them to see that they have failed (or neasure then). Often though with nore expensive devices repairers will use conventional resistors not realising that these are a special type.
+Spoif I don't like it either! I really hate them! Because, they can act like a fire starter on gross overload, but not extreme overload (short circuit)!!! Even if the enclosure is flame retardent (won't burn on it's own), the hot resistor can make it burn and catch other things on fire. This is the same deal with those CFLs. You see sometimes they have a hole burnt in the side? That is from the resistor "fuse" ... very dangerous!
Interesting vid (I realise it is 4 years old now) but I am investigating failures of an SELV for LED panels. What does Ta and Tc refer to? I ask because I have had two SELVs fail in 12 months and I believe they are just getting too hot (environmentally - not much air flow and no heat sink). I can send you one John if you would enjoy ripping it apart to determine root cause? It is a 220v AC to 27-42V DC 1.15A ( PMax 48W)
I agree with some comments below that the short circuit protection method used in this unit represents a fire hazard as amply demonstrated, in principle these units should be housed in fireproof containers either metal (which are earthed) or polyimide material. I have worked on several of these units that have not failed catastrophically and interestingly it is always the bottom transistor of the half bridge totem pole that fails. The circuits are invariably as shown in my link, self oscillating converters using two bipolar high gain high voltage transistors running at about 15KHz frequency and a low power low voltage (BC337) starting circuit transistor. The small rectagular device in the sleeve on top of one of the transistors is actually a thermal fuse and can be readily bought through RS components or Elelement 14 etc. The following link provides typical circuits and photos of electronic ballast units. s1189.photobucket.com/user/Gantos1/library/ELECTRONIC%20BALLAST%20CIRCUITS I'm all for conventional steel core transformers of the past that also had thermal overaload protection devices (soft soldered spring). The unit shown in John's video is a particularly nasty design because the low voltage SELV output is adjacent to the mains input wheras better designed untis have the input & output at opposite ends of the package, albeit isolation is still predicated upon the transformer insulation technique and material. The OSRAM units have the toroidal core in an insulated container and the low voltage secondary wound outside.
+Andrew Joy Older installations did use a single transformer, usually a large toroidal type. The problem with that is the transformer was too large and heavy to pass through the hole in the ceiling, so had to be located elsewhere. Cables from the transformer to each light had to be of substantial size to avoid volt drop, typically 2.5mm² or 4mm², and the transformer was an expensive item. Once the smaller electronic power supplies were available, it was easier and cheaper to fit one per lamp. All irrelevant now as LEDs would be used instead.
mikokuijn If it were that much of a gap either way, Wouldn't you see switchmode transformers for EPS power distribution? No, Its no where as near reliable as a good'ol fat transformer.
i doubt if a resistor caused the blowing of the lamps, more like something else failed making it push out higher voltage, with the resistor then burning up due to excess current
I had one do the same, it was the 12v AC type for halogen lamps. It had no real control circuit and was completely self oscillating, a single 150K resistor failed open circuit and made it go haywire.
Hi John good video, do you know if the led drivers are constructed in the same way, Ie:- the same separation before the conditioning of the output into DC, constant current, Kind regards Warren
I find it impossible to believe that a safety measure involves a requirement for a component to over heat, cause charring and detach itself from the board by melting its solder pad. you seem to think that this is an acceptable design. in my view this is a very dangerous design and the owner was lucky his house did not burn down.
if it was a proper 'fusible' resistor they are supposed to blow in a more safe and controlled manner..i bet it wasnt a proper fusible type, just a normal metal film or small wirewound type...
or it may have been just slightly overloaded, enough to make it hot enough to melt the solder causing arcing for a while, but not enough to pop it quickly
John, The usual failure mode for these type of supplies is the filter capacitor fails short and destroys itself spectacularly. It certainly looks like that is the case here if you look carefully at the hole near its end. When it went short the fault current would have then caused that resistor to fail as a consequence.
That would be my expectation as well. A severe short in the capacitor (too severe to self-clear) can dramatically increase the current flowing through the limiting resistor causing it to overheat and fuse. And judging from the amount of heat damage it looks to have been a rather severe short leading to plasma arcing and carbon tracking exacerbating the damage. Fortunately the resistor did its job by acting as an engineered weak link and disconnecting from the mains.
Hmmm. Proper fusible resistors are designed to fail in a more controlled manner, not light up like a lamp and char everything around them. You usually just see a pinhole in the glaze at the point of failure.
Looks like the damage may have been caused by prolonged arcing across the gap in the resistor foil where it "blew".
So were either of the drive transistors short circuit?
Simon Parkinson Usually depends on the kind of fault. Fusible resistors fail spectacularly against a short circuit.
Dr. Lecter I've dismantled numerous CFLs and found failed fusible resistors but they have been controlled failures, not like the one shown here.
On CRT TVs they were commonly used in the secondary side of power supplies, voltages and fault currents significantly lower though.
Even where used on the primary, I can't remember encountering anything like that.
I think I would class this as a product fail. Wrong choice of component for the job. In my opinion the fault took way too long to disconnect and was a bit too close to causing a fire.
Also while they've used an isolation slot with barrier between mains and SELV, they seem to have forgotten about that fusible resistor again. I'm surprised that arcing hadn't occurred between it's PCB traces during the failure.
It does look like that resistor has been running hot for a prolonged period, if it were to blow suddenly, it wouldn't have caused such excessive damage to surrounding components. I'd say either a rectifier diode or that very charred cap has gone short. Or the secondary side has been drawing excess current for a prolonged period causing everything that resistor to get toasty slowly cook everything around it. Those diodes have been running pretty warm too by the looks of them.
I may have cringed at the screwdriver levering . been there, done that, got the scars
I've stabbed my hand with a screwdriver too many times like that.
Darian Cabot Who hasn't? Same with getting stabbed by that one strand of wire when twisting stranded wires together.
same thing seeing screw driver levering under pressure in hand getting stabbed (not good)
Awesome shirt, John!
Fuse resitors generally make a lot less smoke
they probably used a standard type of resistor...?
Sorry, I could not spot the point in the video where you identified the failing component. All the charring seemed to be the effect of the fault (a short circuit down the line) rather than the cause. Did you actually find the failing component?
thats what i would think,....i've seen and repaired a fair few switch mode psu blow ups in tvs in the 1990's, usually the switching device goes short, taking out a resistor or two, rarely actually popping a fuse!
Why cant you use a single transformer to step down? I didn't see a control chip?
superb videos i watch them every day keep up mate
Hi John. I don't particularly like the idea of a resistor that blows and incinerates the adjacent components as opposed to a fuse that goes OC. It looks like there was a fire on board, which luckily was contained by the enclosure. Are there any fire-proof safety ratings on the enclosure?
Usually these special resistors fail without drama, you normally have to closely scrutinise them to see that they have failed (or neasure then).
Often though with nore expensive devices repairers will use conventional resistors not realising that these are a special type.
How exactly does that relate to fire tolerance of the enclosure ?
Sorry, my reply does not relate to the flammability of the enclosure.
+Spoif
I don't like it either!
I really hate them!
Because, they can act like a fire starter on gross overload, but not extreme overload (short circuit)!!!
Even if the enclosure is flame retardent (won't burn on it's own), the hot resistor can make it burn and catch other things on fire. This is the same deal with those CFLs. You see sometimes they have a hole burnt in the side? That is from the resistor "fuse" ... very dangerous!
Interesting vid (I realise it is 4 years old now) but I am investigating failures of an SELV for LED panels. What does Ta and Tc refer to? I ask because I have had two SELVs fail in 12 months and I believe they are just getting too hot (environmentally - not much air flow and no heat sink). I can send you one John if you would enjoy ripping it apart to determine root cause? It is a 220v AC to 27-42V DC 1.15A ( PMax 48W)
''Meant to act like a fuse, which it has done'' Not!
I don't like how the mains wires are stripped back far, I could see how if they were moved in a certain way, they'd short out! O_O
Hi John. Did that come out of a loft or ceiling void?
Simon Portillo Yes, loft above an upstairs room.
+John Ward Wow, easily could have burnt the whole building to the ground!
I agree with some comments below that the short circuit protection method used in this unit represents a fire hazard as amply demonstrated, in principle these units should be housed in fireproof containers either metal (which are earthed) or polyimide material. I have worked on several of these units that have not failed catastrophically and interestingly it is always the bottom transistor of the half bridge totem pole that fails.
The circuits are invariably as shown in my link, self oscillating converters using two bipolar high gain high voltage transistors running at about 15KHz frequency and a low power low voltage (BC337) starting circuit transistor. The small rectagular device in the sleeve on top of one of the transistors is actually a thermal fuse and can be readily bought through RS components or Elelement 14 etc.
The following link provides typical circuits and photos of electronic ballast units.
s1189.photobucket.com/user/Gantos1/library/ELECTRONIC%20BALLAST%20CIRCUITS
I'm all for conventional steel core transformers of the past that also had thermal overaload protection devices (soft soldered spring).
The unit shown in John's video is a particularly nasty design because the low voltage SELV output is adjacent to the mains input wheras better designed untis have the input & output at opposite ends of the package, albeit isolation is still predicated upon the transformer insulation technique and material. The OSRAM units have the toroidal core in an insulated container and the low voltage secondary wound outside.
The high frequency output causes a lot of interference!
looks like there has been a small fire in there.
Is it not terribly inefficient to have one behind each lamp ? would it not be better to have one large one running several ?
+Andrew Joy Older installations did use a single transformer, usually a large toroidal type. The problem with that is the transformer was too large and heavy to pass through the hole in the ceiling, so had to be located elsewhere. Cables from the transformer to each light had to be of substantial size to avoid volt drop, typically 2.5mm² or 4mm², and the transformer was an expensive item.
Once the smaller electronic power supplies were available, it was easier and cheaper to fit one per lamp.
All irrelevant now as LEDs would be used instead.
i'd also assume voltage fall of is a problem?, i mean, it is only 12v.
I'm not an expert, but I'd assume that the smaller toroid allows a higher switching frequency to get more efficiency.
mikokuijn If it were that much of a gap either way, Wouldn't you see switchmode transformers for EPS power distribution?
No, Its no where as near reliable as a good'ol fat transformer.
That looks like the resistor didn't do it's job as a fuse, that is spectacular failure. Magnetic transformer would last 10 of these smps easily._
I had 7 of this 12v electronic transofrmers blow up, even one made the lamp explode on the inside! The problem was on all of them the resistor...
i doubt if a resistor caused the blowing of the lamps, more like something else failed making it push out higher voltage, with the resistor then burning up due to excess current
I had one do the same, it was the 12v AC type for halogen lamps. It had no real control circuit and was completely self oscillating, a single 150K resistor failed open circuit and made it go haywire.
Hi John good video, do you know if the led drivers are constructed in the same way, Ie:- the same separation before the conditioning of the output into DC, constant current,
Kind regards Warren
could of caused a fire, it was well and truly burned. Good video!
I find it impossible to believe that a safety measure involves a requirement for a component to over heat, cause charring and detach itself from the board by melting its solder pad. you seem to think that this is an acceptable design. in my view this is a very dangerous design and the owner was lucky his house did not burn down.
if it was a proper 'fusible' resistor they are supposed to blow in a more safe and controlled manner..i bet it wasnt a proper fusible type, just a normal metal film or small wirewound type...
or it may have been just slightly overloaded, enough to make it hot enough to melt the solder causing arcing for a while, but not enough to pop it quickly
I can smell it.
MainsOnTheOhmsRange you know what's scarier? *Mains on the amps range.*
A mistake you only make once.
An ac smps they can't even get a regular TX