Converting ~320V DC down to ~12V DC (not much power required)

[axero]

I've been troubleshooting a faulty power supply that is very small, perhaps 5W or less. The heart of this circuit is a small 8-pin SOIC that has no label on it and I suspect that it is broken.

I want to know what type of circuit/chip this is to begin with. It must be some kind of a standard chip.

On pin 5 comes a about 320V that is rectified and buffered by a 2.2 uF capacitor. Out from pins 1 and 3 comes about 12V. None of the other pins seem to be connected though I might be wrong.

That's when things are working, as of now there is no voltage at all on the 12V side.

Does anyone have any clue?

[T3sl4co1l]

Probably a Chinese clone of a TOPswitch or such.  There are a few likely candidates.

You can order new /assembled/ adapters for much less than you can order the chips, let alone install them (and troubleshoot further faulted components).

Tim

[axero]

Installation is no issue for me. I looked into the datasheets of the TOPswitch circuits. The TOPswitch HX appears to be the only series with 8-pin SOICs. However, they are not pin compatible so this must be something else.

Are these circuits really that expensive? They appear to be in almost every high-voltage LED out there. In for example G5 "corn LED" (becuz it looks liek a piz of corn) you have the rectifier IC and this 8-pin IC.

As for pics, here is one:

https://ae01.alicdn.com/kf/HTB16o_GPXXXXXcPapXXq6xXFXXXz/433-MHz-2-Ch-Wireless-RF-Transmisor-Teledirigido-Interruptor-de-Rel-DC-12-V-220-V.jpg

It's the small 8 pin SMD chip right between 4R7 and 471.

[T3sl4co1l]

Expensive, no, but the manufacturers have million-piece contracts that you don't have access to.  It's ridiculous to talk about repairing a 10W charger -- any meaningful amount of talk has already spent more combined labor than the device is worth!

Tim

[Mechatrommer]

Expensive, no, but the manufacturers have million-piece contracts that you don't have access to.  It's ridiculous to talk about repairing a 10W charger -- any meaningful amount of talk has already spent more combined labor than the device is worth!
Tim

it aint 10W charger in the picture he linked, its mains powered non-isolated rf transmitter, probably not easily found lying about.

On pin 5 comes a about 320V that is rectified and buffered by a 2.2 uF capacitor. Out from pins 1 and 3 comes about 12V. None of the other pins seem to be connected though I might be wrong.
Does anyone have any clue?

As for pics, here is one:
https://ae01.alicdn.com/kf/HTB16o_GPXXXXXcPapXXq6xXFXXXz/433-MHz-2-Ch-Wireless-RF-Transmisor-Teledirigido-Interruptor-de-Rel-DC-12-V-220-V.jpg
It's the small 8 pin SMD chip right between 4R7 and 471.

from your picture and description, its probably a soic8 high voltage fet. i suspect nfet because thats what people usually did. the other soic8 on top (maybe high voltage dc-dc converter) providing charge pump to the fet gate. or... it could be a pfet, tracing the routing may give the clue. if its indeed a fet, you may find any compatible pin fet with the same rated voltage as replacement. ymmv.

[T3sl4co1l]

it aint 10W charger in the picture he linked, its mains powered non-isolated rf transmitter, probably not easily found lying about.

That was an example, and indicated what looks like the chip in question.  Original post says "power supply", "5W or less"...

from your picture and description, its probably a soic8 high voltage fet.

It's an integrated regulator and switch, like the example I gave (TOPswitch).  There is no external controller or driver.  There are many types.  I just don't remember the most common Chinese versions.

Tim

[axero]

from your picture and description, its probably a soic8 high voltage fet. i suspect nfet because thats what people usually did. the other soic8 on top (maybe high voltage dc-dc converter) providing charge pump to the fet gate. or... it could be a pfet, tracing the routing may give the clue. if its indeed a fet, you may find any compatible pin fet with the same rated voltage as replacement. ymmv.

I don't understand what "providing charge pump to the fet gate" means. Ok, I understand the word "FET" but the technicals of this I don't. I also don't know how to find a proper "FET" for this application. I don't even know where to look.

Is there a book or a source on how to design a FET chip so that it turns 320V DC to 12V DC or something?

Tbh, I find it truly impressive that they can turn such a strong voltage into 12V with such small a chip.

I may or may not consider replacing the chip if it isn't worth it. But at the very least I have learned something in the process.

It is an RF receiver circuit and I would say that the load is isolated from the controller circuit. The two blue plastic boxes are mechanical relays and the breaking circuits have nothing whatsoever to do with the rest of the PCB.

[PartialDischarge]

Look online for “offline buck converter”, thats what that is. It could be a Viper from ST, a top switch, a texas Instruments part or something from any shady manufacturer.

I like the viper013 for its 800V rated mosfet. Here’s a dev board
http://www.st.com/en/evaluation-tools/steval-isa178v1.html

[Mechatrommer]

It can be topswitch it can be other topology.. but it also can be a simple buck converter http://www.learnabout-electronics.org/PSU/psu31.php



recently i came across high voltage dc dc converter in a psu circuit i cant recall what chip. But it transformered/isolated type. Thats why i'm thinking such possibility..

[axero]

That diagram is really easy to understand thanks a lot for posting it! I like the name "flywheel circuit".

I guess you can have different "resolutions" at which those pulses are delivered. If the resolution is 20 kHz then a brief calculation yields that the pulse width should be $$1.875\mu s$$ for the conversion from 320V down to 12V.

I wonder how one can design such a chip and how one can control that pulse width. The power supply accepts voltages from 85V to 240V, how come it can accept such a wide range of input voltages and yet maintain 12V of output?

Perhaps it isn't the converter chip that is broken. If the flywheel circuit is broken somehow, would a multimeter register an output voltage from those short output pulses?

[Mechatrommer]

Perhaps it isn't the converter chip that is broken. If the flywheel circuit is broken somehow, would a multimeter register an output voltage from those short output pulses?

if the top ic is a dc dc converter, one of its pin should produce square wave (PWM), if no PWM, then its suspected dead. if there is, the square wave will be feed to fet gate pin so the fet will act like a on off switch. if the fet is ok, you should be able to read 320V square wave voltage on its drain pin that will create 12V output. if no square wave, fet is dead.

[axero]

I have measured the flywheel circuit. I suspected that the coil would be broken, i.e. open. That turned out to not be the case.

If it were, then the output voltage would be short spikes of 320V that an ordinary multimeter perhaps would be unable to register. I don't have a scope.

[Beamin]

Installation is no issue for me. I looked into the datasheets of the TOPswitch circuits. The TOPswitch HX appears to be the only series with 8-pin SOICs. However, they are not pin compatible so this must be something else.

Are these circuits really that expensive? They appear to be in almost every high-voltage LED out there. In for example G5 "corn LED" (becuz it looks liek a piz of corn) you have the rectifier IC and this 8-pin IC.

As for pics, here is one:

https://ae01.alicdn.com/kf/HTB16o_GPXXXXXcPapXXq6xXFXXXz/433-MHz-2-Ch-Wireless-RF-Transmisor-Teledirigido-Interruptor-de-Rel-DC-12-V-220-V.jpg

It's the small 8 pin SMD chip right between 4R7 and 471.

Great care was taken to make the parts straight.

I like fixing things rather then just buying a new product even if it's more money or takes time.