Tiny switch protection

Hary,


Startup produces the most stress on a primary MOSFET, so it's not surprising this is when you're having your greatest problems. Our internal MOSFETs are rated for 700 VDC. If the voltage across the MOSFET is greater than this at any time, the device may fail. To prevent this from happening, we recommend you install a clamp circuit across the primary winding of your transformer. The clamp will catch any energy stored in the leakage inductance of the transformer and either dissipate it or recycle it back into the circuit. Either way, it prevents this energy from being discharged onto the Drain of the MOSFET, which may cause unsafe voltages for the device.


Does your design currently use a clamp circuit? If not, please review page 8 of the TinySwitch-III datasheet for an example of a clamp circuit, and page 10 for more information about the design. If you're using PI Expert for your design, it will suggest a design and component values for your clamp circuit. (http://www.powerint.com/sites/default/files/product-docs/tny274-280.pdf)


If you are using a clamp, please let me know what kind you are using and what the component values are. It will also be helpful for you to measure the primary leakage inductance of your transformer.


Cheers

Dear Jordi,

In my circuit I am using one clamp circuit in the primary side of the transformer

Its an RCD clamp circuit where the values of the components are

1. R clamp1 : 510K / 1/2 watt, Axial leaded
2. R clamp2 : 100R / 1/4 watt, Axial leaded
3. C clamp : 1.2nF/ 1 KV, Ceramic disc capacitor ,through hole
4. D clamp : UF4007 , Axial leaded

Hary

Hary,


Those values sound reasonable. To make sure that your clamp is completely optimized, we'll need to know the primary leakage inductance of your transformer. If you have an LCR bridge, the primary leakage inductance is measured across the primary winding with all other windings shorted together.


If you don't have an LCR bridge, you can measure the leakage inductance from the drain voltage waveform while your supply is running at full load. For more information on how to do this, watch the AppsTV(tm) episode titled "Estimating Transformer Parasitics". (http://www.powerint.com/en/design-support/appstv). After you have this information, we'll walk through the calculations necessary to double check your clamp design.


-Jordi

Dear Mr. Jordi,

I have measured the primary leakage inductance as per your suggestions. I was able to observe about 25uH.The primary Inductance of the same Tx is 260mH.

I used an LCR bridge for doig the measurement.

Rgds
Hary

Thanks Hary,


Your leakage inductance is pretty low. What frequency are you testing at? Can you please send me your design specification (output power, voltage and current) along with a schematic and layout for your design? I'll look it over to see if there is anything unusual. Also, what is the peak primary current you measure in your design at low line while the supply is running at full load?


Thanks.

Dear Jordi,

Frequency in the sense I/P supply frequency or switch freq,

I/P Supply freq is 50Hz and switching freq is 132KHz

Pls see below specs

I/P Voltage : -32VDC to -72VDC & 85VAC to 265 VAC
Primary current @1Amp load and 36VDC : 220mAmp
O/P Volt : 5VDC
O/P Load Current : 1Amps max.

Here I am attaching 1. PCB layout 2. Schematic diagram 3. DI-152

Note : Its a sigle layer PCB , Also there is a mistake in my previous mail in the measurement of leakge indcutance, iTs actually coming 55uH .

Rgds
Hary

Hary,


Actually I'm curious what frequency you set your LCR to measure inductance at. We usually measure and specify our transformer inductance at 100 kHz. I've looked over your design and simulated some of the circuits and everything seems to look good. Does the TinySwitch always blow up at startup? Or only sometimes?


Also, what is your application? And are you feeding the design with negative DC voltage, or is that a typo?


I think this problem might need some attention from our field staff. I'll send you a private message so that we can get in touch with you directly.

Dear Jordi,

1. Our LCR measures Inductance of Tx at 100Hz frequency

2. Our application is mainly Telecommunication equipments like Modems , LAN extenders, converters( E1 to ethernet , E1 to V.35 ) etc.

We use Interfaces like E1, V.35, COD , Ethernet & DSL.

Our power requirement is max. 5W ( 5V & 1Amp)

Normally in Telephone exchanges we avail negative DC voltages only for power transmission ( From -32VDC to -76VDC ). But irrespective of the polarity of I/P supply Tiny switch gets a good positive DC voltage as we are using a bridge rectifier in our primary circuit.

3. Not all Tiny switch is getting blasted but a considerable qty failed so far, I can say 25 out of 60 of our R&D prototype design.

4. We are already in touch with Mr. Muthu Kumaran & Mr. Pongalaban, both are FAEs working in Spectra Innovations India (P) Ltd on behalf of Power Integration Bangalore, India. We are getting very good response from them,
One week back I had send two of my units to them for evaluation and further testings.
As per their testings the drain current is shooting beyond the limit , Around 1.56Amp ,and they are telling its because of Tx design . Is the leakage Inductance creates all these issues or some other reason should be ther to sorted out

Hoping your reply at the earliest

Rgds
Hary