How does TNY266PN work?

[max.wwwang]

I'm trying to repair the faulty main board of a Bosch clothes dryer. It does not power on. Some online research suggests that the culprit is the TNY266PN IC and possibly the associated power resistor.

I have ordered this chip and can simply replace it. But I don't understand how this IC works. I'm not looking at the inside of the chip, but rather how it works in a circuit. Can someone give me some help? Below is a typical application of the chip on the datasheet (https://www.alldatasheet.com/datasheet-pdf/pdf/139805/POWERINT/TNY266PN.html). Thank you.

[SeanB]

Smjall package that has the power switch, the current sense resistor to turn it off, the start up power supply and the controller all in one package, so making it a single part to insert in assembly. Instead of having a dozen parts to put on the board, you now have one, optimised internally for efficiency, but with that compromising design flexibility, but for applications needing a power supply that does up to 5W of power, and wide input voltage range, very convenient to have.

[golden_labels]

max.wwwang:
The primary mode of operation for this entire family is releasing magic smoke.

The operation may not be instantly obvious, because the primary power source is on the right of the functional diagram (Fig. 2) and is also the output. Consider S (source) being the ground reference. D (Drain) initially provides voltage through the primary of the transformer. At this point that is acting as a simple inductor with nearly zero current and hence voltage on D being nearly nothing. Since the chip sucks a tiny current, the inductor charges and voltage on D slowly rises. At some point it’s enough so the initial 5.8V voltage regulator can provide voltage, which charges the bypass (BP) capacitor. That capacitor serves as a reservoir of energy for later operation. That’s needed, because turning off the main MOSFET will cause huge oscillations, which will temporarily make the internal voltage regulator deliver nothing. To this point we have established where the chip is powered from.

Initially the ENABLE signal (EN/UV) is logical TRUE. In absence of current or any fault/overheat detection signal, the internal state machine enables the built-in power MOSFET. On the right in Fig. 2; the huge area with regular lines pattern area in my photos. It allows large current to flow, pulling the transformer’s primary down. That provides current to appear on secondary, charging the DC output capacitor.

At some point the DC output voltage is hight enough to trigger the feedback optoisolator (bottom right in Fug. 1), which pulls EN/UV down (logical FALSE), signaling that the voltage is high enough. That tells chip’s internal logic that it may turn MOSFET off. After which the entire process repeats. Note I used word “may”. It seems that with this family the cycle is still determined by the internal logic and EN/UV is just taken into account while taking decision about starting the next charging cycle and not forcefully turning the MOSFET off instantly. That is because those chips try not to maintain stable voltage, but pull fixed current. You may see that in the charts in Fig. 6 to 9 with a complete cycle being performed, while feedback voltage going down only preventing the next one from starting.

The devices also incorporate a bunch of additional logic related to detecting and reacting to abnormal conditions, like undervoltage and too high current over MOSFET.

[BrokenYugo]

https://en.wikipedia.org/wiki/Flyback_converter

[amyk]

The primary mode of operation for this entire family is releasing magic smoke.

That was my first thought too... these single-chip SMPSes tend to be very sensitive to noisy mains. As for actual operation, they're not too different from the classic self-oscillating flyback, but with most of the components integrated into a single chip.

[golden_labels]

I can’t tell anything about single-chip SMPS controllers in general, but I see many mentions of this particular family failing catastrophically and I even encountered one explosive failure myself. A good joke to insert when mentioning them, in particular since OP is dealing with one that died too.

[max.wwwang]

https://en.wikipedia.org/wiki/Flyback_converter

That's exactly the kind of thing that I was after. Thanks!

[max.wwwang]

The primary mode of operation for this entire family is releasing magic smoke.

Given my case and that case, this is a fair comment. 

The operation may not be instantly obvious, because the primary power source is on the right of the functional diagram (Fig. 2) and is also the output. Consider S (source) being the ground reference. D (Drain) initially provides voltage through the primary of the transformer. At this point that is acting as a simple inductor with nearly zero current and hence voltage on D being nearly nothing. Since the chip sucks a tiny current, the inductor charges and voltage on D slowly rises. At some point it’s enough so the initial 5.8V voltage regulator can provide voltage, which charges the bypass (BP) capacitor. That capacitor serves as a reservoir of energy for later operation. That’s needed, because turning off the main MOSFET will cause huge oscillations, which will temporarily make the internal voltage regulator deliver nothing. To this point we have established where the chip is powered from.

Initially the ENABLE signal (EN/UV) is logical TRUE. In absence of current or any fault/overheat detection signal, the internal state machine enables the built-in power MOSFET. On the right in Fig. 2; the huge area with regular lines pattern area in my photos. It allows large current to flow, pulling the transformer’s primary down. That provides current to appear on secondary, charging the DC output capacitor.

At some point the DC output voltage is hight enough to trigger the feedback optoisolator (bottom right in Fug. 1), which pulls EN/UV down (logical FALSE), signaling that the voltage is high enough. That tells chip’s internal logic that it may turn MOSFET off. After which the entire process repeats. Note I used word “may”. It seems that with this family the cycle is still determined by the internal logic and EN/UV is just taken into account while taking decision about starting the next charging cycle and not forcefully turning the MOSFET off instantly. That is because those chips try not to maintain stable voltage, but pull fixed current. You may see that in the charts in Fig. 6 to 9 with a complete cycle being performed, while feedback voltage going down only preventing the next one from starting.

The devices also incorporate a bunch of additional logic related to detecting and reacting to abnormal conditions, like undervoltage and too high current over MOSFET.

Thanks for the explanation of its working in detail. I get the rough idea but am unable to fully understand it yet. 

[max.wwwang]

For people like me who are new to flyback converters, I found this material extremely good.
https://usermanual.wiki/Document/TheFlybackConverterLectureNotes.120621815

Can someone please help me here - in this material (Fig. 2/b on p 2) I think the arrow for iC should be pointing upwards. Am I correct?