Author Topic: MLCC for high voltage decoupling in SMPS  (Read 4964 times)

0 Members and 1 Guest are viewing this topic.

Offline aniketTopic starter

  • Contributor
  • Posts: 27
  • Country: in
MLCC for high voltage decoupling in SMPS
« on: August 08, 2016, 05:59:23 am »
Hi ,
Is it possible to use MLCC capacitors in 2220 case size in place of polyester film for decoupling purpose . I am designing a 1200W full bridge SMPS running at 45Khz and i need to have some decoupling capacitance at the 400V rail very near to the mosfets .
As size is a constrain will be possible to use MLCC ????
schematic attached ....
 

Online T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22434
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: MLCC for high voltage decoupling in SMPS
« Reply #1 on: August 08, 2016, 06:03:03 am »
2220's are prone to cracking. I don't recommend it.  They're also rather expensive.

You could save the space on the coupling capacitors, which should only need to handle < 50V in operation, with a rating of maybe 160 or 250V.

Why not put it closer to the PFC cap, and why a CMC and fuse?  The fuse won't save anything...

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline aniketTopic starter

  • Contributor
  • Posts: 27
  • Country: in
Re: MLCC for high voltage decoupling in SMPS
« Reply #2 on: August 08, 2016, 06:23:23 am »
As the PFC is running at 66khz and the full bridge is running at 40-45khz the CMC is used to reduce the interference between them both ....
and yes the Fuse doesnt save any thing for sure, its used as safety precaution ......
Find attached layout of the entire SMPS layout  , the full bridge part is not complete yet, it shows the PFC layout ... any comments on the that will also be help full .....
 

Online T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22434
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: MLCC for high voltage decoupling in SMPS
« Reply #3 on: August 08, 2016, 07:30:25 am »
If you'd like help with schematic and board layout, we can take this to IM and talk about work. :)

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline aniketTopic starter

  • Contributor
  • Posts: 27
  • Country: in
Re: MLCC for high voltage decoupling in SMPS
« Reply #4 on: August 08, 2016, 10:42:39 am »
If you'd like help with schematic and board layout, we can take this to IM and talk about work. :)

Tim

Thanks TIM for the offer , but I would like to try it out on my own.

regards
Aniket
 

Offline MagicSmoker

  • Super Contributor
  • ***
  • Posts: 1408
  • Country: us
Re: MLCC for high voltage decoupling in SMPS
« Reply #5 on: August 08, 2016, 11:45:40 am »
MLCCs larger than 1206 (3216 metric) can survive okay if they feature "soft terminations," and/or are specified for automotive use, though if you have to use lead-free solder then all bets are off.

Another thing to watch out for with SMT capacitors in this application is violating minimum creepage distance. Sure, the part is *functionally* capable of withstanding 400 or 450VDC, but that doesn't mean a safety ratings agency will accept it, and once you have to start resorting to putting capacitors in series you might as well go with a through-hole part, especially since a metallized film type is the best for this application.

That said, there are some other issues with your circuit as drawn (as well as the way it is drawn - like those funky angles) but as implied by Tim's offer, we aren't here to do your job for you (not for free, anyway).

 



 

Offline aniketTopic starter

  • Contributor
  • Posts: 27
  • Country: in
Re: MLCC for high voltage decoupling in SMPS
« Reply #6 on: August 08, 2016, 12:01:42 pm »
MLCCs larger than 1206 (3216 metric) can survive okay if they feature "soft terminations," and/or are specified for automotive use, though if you have to use lead-free solder then all bets are off.

Another thing to watch out for with SMT capacitors in this application is violating minimum creepage distance. Sure, the part is *functionally* capable of withstanding 400 or 450VDC, but that doesn't mean a safety ratings agency will accept it, and once you have to start resorting to putting capacitors in series you might as well go with a through-hole part, especially since a metallized film type is the best for this application.

That said, there are some other issues with your circuit as drawn (as well as the way it is drawn - like those funky angles) but as implied by Tim's offer, we aren't here to do your job for you (not for free, anyway).


yep got your point , i didn't ask any one to do my job for free, I just asked for a suggestion. anyways thanks for the valuable comments ... 
 

Offline aniketTopic starter

  • Contributor
  • Posts: 27
  • Country: in
Re: MLCC for high voltage decoupling in SMPS
« Reply #7 on: August 08, 2016, 04:20:48 pm »
Also keep in mind that 1uF 630V MLCC is not 1uF AT 630V. Their capacitance decreases very quickly as biasing voltage goes up.
You only get 1uF at zero bias, at 100V it could be well below 0.5uF, and at 300V it could be well below 0.1uF.
For high power, high voltage and high reliability applications (which is, what you expect for a SMPS DC bus cap), film caps are more appropriate unless you want to save absolute every cent.
Another viable way is to use several MLCCs in series-parallel connection with balancing resistors, generally I would more likely to trust 250V or less MLCCs operating at under 160V each in such an app.

Thanks, i was thinking of using  several 600V MLCC for a 380V DC bus in parallel &  I will be using a case Size of 2220 or bigger to main clearance and creepage ....
 

Offline Ice-Tea

  • Super Contributor
  • ***
  • Posts: 3180
  • Country: be
    • Freelance Hardware Engineer
Re: MLCC for high voltage decoupling in SMPS
« Reply #8 on: August 08, 2016, 07:16:20 pm »
Just wondering... IPC2122A requires only 2.5mm? Wouldn't you have a case if you'd stick to that?
 

Offline bktemp

  • Super Contributor
  • ***
  • Posts: 1616
  • Country: de
Re: MLCC for high voltage decoupling in SMPS
« Reply #9 on: August 08, 2016, 08:06:38 pm »
Thanks, i was thinking of using  several 600V MLCC for a 380V DC bus in parallel &  I will be using a case Size of 2220 or bigger to main clearance and creepage ....

2220 capacitors have 3.55mm creepage between pads, which will never pass UL60950-1 300V~600V class, pollution level 2 requirement, unless you pot the unit in some potting material or super thick conformal coating.
Does it really require >3.5mm for functional isolation? If you use TO220 transistors there are less than 1mm creepage between the pins.
 

Online T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22434
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: MLCC for high voltage decoupling in SMPS
« Reply #10 on: August 08, 2016, 08:56:48 pm »
I don't have 60950 handy anymore, but isn't that only for mains class wiring, with a respective transient overvoltage rating (2.5 to 4kV depending on environment and reliability level)?

Even if you aren't using a MOV, once you connect mains up to a FWB and cap, the differential transient requirement goes away, AFAIK.

It is another concern about ceramic caps, though.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf