Snubber circuit design problem and more

8 posts / 0 new

Hi

  I have made a 30W LED Driver with 30V o/p voltage. I have followed the design made by PI Expert. But snubber resistor dissipates  much more power than its specified value in Excel sheet. While measuring the clamping voltage it's showing 110V, instead of 180V( as per calculation it should come 180V). Vor is 100V. I have connected the same R and C as per the suggestion.

    Another thing is after designing the Drive, Rfb required is 280K where in design sheet it's 210K.Bias winding resistance is  coming about 28V instead of 20V.

   In my design,I have used 0.47uF as Peak detector Capacitor and hope that driver would sustain atleast 4KV surge. Is there any problem with this configuration? Should I include a TVS diode in parallel with the peak detector Capacitor? Please suggest what to do. Its very urgent

Hi,

Snubber resistor dissipates too much heat if the snubber capacitor is not optimized.  Try reducing the capacitance value without too much sacrificing the stress across the drain to source pin.

 

Bias voltage may not be the same as the calculated value if the transformer construction is not optimized.  Make sure to have a good magnetic coupling between secondary and bias winding by making the two adjacent to each other.

 

Another reason that the bias voltage is higher than the expected if the actual LED voltage is different from the calculated.

 

Anyway, you are in a good track.  You can control the output current by adjusting the RFB resistor.

 

0.47uF peak detect capacitor is too small to clamp the energy for 4kV differential surge.  If it is a 4kV common mode surge then it should be okay.

 

Assuming a 4kV differential surge then you need to have more clamping capability.  Make sure that you have MOV in the input side to protect your bridge rectifier.  You can provision a TVS across the bulk to protect the input capacitor and the switching device.  Note that during surge, the output diode is also subjected to stress so make sure proper clamping approach to begin with for this kind of condition.

 

Regards,

PI-Jono

Thanks for your reply. I have couple of other questions. Is MOV across DC link capacitor enough to protect the switch or  another MOV should be connected across peak detector capacitor? Another question is which one is better to use for surge purpose? TVS or MOV?

 

I was facing another problem like after continuous switching ON and OFF the driver, switch is getting blasted. Is it because of switching transient/?

Hi,

MOV is generally used to clamp the input voltage to avoid breaking the Bridge rectifier. Since your requirement is 4kV differential line surge then use a 1kV bridge rectifier diodes plus an MOV sufficient to clamp the energy but make sure your fuse can handle this transient power.  You can use slow blow fuse to avoid unwanted fusing.

 

During chatter test, random on-off; unit might fail if the inductor EMI filter inductance is too high that creates inductive kick in the bulk capacitor. This can be avoided by using a series resistor in the input to minimize the inrush current.  If it is low power application, you can use fusible resistor in place of the normal fuse.  Use the right surge power rating.

For high power application, you can series a thermistor to minimize the inrush current. 

 

TVS clamping voltage is more accurate than MOV.  A 275VAC MOV will clamp around 670Vpeak which is too high to use across bulk.  Use TVS across bulk to guarantee that the spike will not exceed 450V to protect input capacitor, switching device, and output diodes.  You can put TVS across the peak detect capacitor but you need to consider the peak current rating of the diode.  So for simplicity TVS is used across bulk capacitor.

 

Your project is interesting because it looks like it will be used in industrial application.....Surge is one of the challenging parameter to meet. 

 

Regards,

PI-Jono

 

 

Thanks for your reply. But I have couple of other problems like if I will switch ON and switch OFF continuously then sometimes LEDs take sometimes to glow. But if it is OFF for longer time then it becomes ON instantly. Can you tell me the reason behind this and how to solve it?

 

Another thing is when  I am designing a driver for 45V, 30W driver then with multiple switch ON-OF, MOSFET is going bad.Same thing happens with 50V output irrespective of the wattage. Again, driver efficiency also varies from 85 to 87%. Again power taken from the input as well as the output current also changes. Is it possible to make the driver stable?

Hi,

Would you send some waveforms so that we can easily analyze.  Probe the timing of the AC input and the output voltage plus the BP pin.

 

Send also the schematic.

 

Send also the schematic of your 30W design.  So that we can easily review it.

 

Regards,

PI-Jono

Hi, I have similar problems.

 

1 - Could you explain mor about size (400W, 600W or 1,5KW) and voltage of TVS to put across de bulk capacitor? 

2 - In some designs, theres is a 10uF x450V electrolitic for transient surge capability. Is possible to use a TVS in place of this capacitor?  

 

Thanks. 

 

Hi,

1- This refers to the peak power the TVS can handle for a certain time.  Refer to this sample datasheet to understand its capability: http://www.bourns.com/data/global/pdfs/CD0603_1005-T.pdf

2- If pcb space is available then using a capacitor with a diode is more practical.  For LNKPH and LYT-4 you use the peak detector in two applications.  As a sampling signal for the IC and a surge protection at the same time.

Yes, a TVS in place of a capacitor is acceptable. Make sure the proper power rating and peak clamping voltage is considered in the selection.

Regards,

Jono