解决方案查找器 技術支援

TOPSwitch output voltage drop

Posted by: Tugay on

Hi friends,

I designed 32V 7A smps using Top261EN. Below I share the test results of the design. There are some issues with smps.

Questions:

1.) Clamp circuit components overheat, especially tvs diode and clamp blocking diode. I reviewed the AN14 document, if I increase the power of the elements in the clamp circuit, will the problem be fixed?



2.) Output diode is overheating, what should I do to reduce heating. I tried DPG60IM400QB and MBR30200FCT_T0 diodes, they both get too hot even though I have a heatsink installed.



3.) The most important question is, when I connect a load between 3A - 7.5A, the smps components get hot, but there is no problem in its operation. Output voltage drops to 30V at 7.5A load. When I overload, for example, 8.5 A, the output voltage does not decrease, it becomes 30V again, but the topswitch explodes.How can I prevent topswitch explode when overload ?  I don't want the voltage to stay constant when overload is loaded, I want the voltage to drop. What should I do to reduce the output voltage in overload?

I attached all documents and test results for Smps desing.

Transformer primary inductance:  250uH

Transformer leakage inductance: 5uH 

Files

Attachment 大小
load test results 32.18 KB
Pıexpert design 440 KB

回應

Submitted by PI-Wrench on 11/11/2021

I cannot see your design results - when I try to open them, I get random characters. Perhaps it would be best to post the design file as a PDF.  A schematic would also be very helpful.

Submitted by PI-Wrench on 11/12/2021

Was the UF4007 in the snubber actually the part called out by PI Expert? This is definitely too small a part for the output power you are trying to achieve. A 3A-rated part like the UF5400 series would be more appropriate. Did PI Expert also call out use of a 400V TVS for the snubber?

Submitted by PI-Wrench on 11/12/2021

Another approach for your application might be to use an LLC converter rather than a flyback, utilizing the Hiper-LCS series of parts. This would likely have higher efficiency and lower peak primary current. 

Submitted by Tugay on 11/23/2021

 " Did PI Expert also call out use of a 400V TVS for the snubber? "

No , Pı expert called out 160V TVS but since I didn't have 160v TVS, I used 200V, 350V and 400V TVS diodes.

When I try to use 200V, 350V and 400V TVS, anything was not change and Topswitch was exploded again when smps overloaded.

What is the importance TVS diode voltage ? Because when I change the TVS voltage in pı expert design, anything does not change.

Is tvs voltage important in topSwitch explosion?

I have read the AN14 and AN16 documentations. I know how to select the TVS clamp voltage but I don't know what will happen or how will it affect if chosen above or below the calculated TVS clamp voltage? Can you help me to understand ?

Submitted by PI-Wrench on 11/23/2021

The TVS clamp is used to limit the drain voltage excursion during the TOPSwitch "off" time so that the drain voltage does not exceed the TOPSwitch drain voltage rating. The clamp voltage for a TVS generally has a positive tempco, so that the clamp voltage increases as the device heats up. Also, the higher the TVS voltage, the more "resistive" the device will appear, increasing the clamp voltage in response to the energy stored in the leakage inductance. Selecting a 400V device will probably allow the TOPSwitch drain voltage to exceed the device ratings. This can be verified by monitoring the drain voltage using a 100X oscilloscope probe. For this design, I would not use a TVS of more than 200V rating. 

In order to prevent the clamp components from overheating, you should use a 3A rated clamp diode, such as the UF5407. If the TVS is overheating, you can take some of the stress away from it by increasing the value of the clamp capacitor and reducing the value of the resistors across the clamp capacitor - the resistors allow the clamp capacitor to discharge while the primary switch is "on", so that when the switch turns off, some of the leakage energy is diverted into the clamp capacitor rather than immediately hitting the TVS. Also note that the schematic calls out 4 X 47k resistor in parallel to discharge the clamp capacitor. If you are using only one 47k resistor, the capacitor will not discharge sufficiently in order to do its job.

In order to reduce the stress on the primary clamp components, it would be a good idea to try and reduce the transformer leakage inductance. It is also a good idea to make the secondary layout as tight as possible, since any stray inductance on the secondary side is reflected back to the primary side by the transformer and acts like extra leakage inductance. The output rectifier and filter capacitors should be as close to the transformer as possible, with short, direct (and wide) traces. This is especially important for high power flyback designs like this one.