Author Topic: Input capacitor in buck converter  (Read 1617 times)

0 Members and 1 Guest are viewing this topic.

Offline Red_MicroTopic starter

  • Regular Contributor
  • *
  • Posts: 122
  • Country: ca
Input capacitor in buck converter
« on: May 04, 2020, 12:51:17 am »
I've chosen this Reference Design as starting point. Here the schematic. I'm using full wave rectifier instead.

I'm designing a buck converter with these requirements:

Input voltage: 120-690Vrms AC (three phase) 50/60Hz
Output: 50V/200mA (max)
To meet the high voltage requirements, the reference design uses a balancing network of three 400V capacitors, which may be expensive.

For my application following the controller datasheet, I calculated a minimum capacitance of 10 uF. However, I'm considering if I really need it. For example, this another Reference Design based on flyback topology doesn't use bulk input capacitors, just ceramic ones or film ones. But I think that applies if I always have three phase, right? If it happens that only two phases are available, then may I run into issues? What do you suggest?
 

Offline graybeard

  • Frequent Contributor
  • **
  • Posts: 443
  • Country: us
  • Consulting III-V RF/mixed signal/device engineer
    • Chris Grossman
Re: Input capacitor in buck converter
« Reply #1 on: May 04, 2020, 05:03:32 am »
My issue with that design is the output is not isolated from the mains.

Offline Red_MicroTopic starter

  • Regular Contributor
  • *
  • Posts: 122
  • Country: ca
Re: Input capacitor in buck converter
« Reply #2 on: May 04, 2020, 05:28:38 am »
My issue with that design is the output is not isolated from the mains.

My application does not require isolation. If I put it, it would be a cost adder. But it will be more safe.
 

Offline tszaboo

  • Super Contributor
  • ***
  • Posts: 8218
  • Country: nl
  • Current job: ATEX product design
Re: Input capacitor in buck converter
« Reply #3 on: May 04, 2020, 07:09:58 am »
What do you suggest?
Buy a 10W 48V isolated DC-DC converter from a reputable brand. Why would you even connect it on 3 phase in the first place. Even if you have a 3 phase connection for other reasons, just leave the other 2 out of the design.
 

Online jbb

  • Super Contributor
  • ***
  • Posts: 1265
  • Country: nz
Re: Input capacitor in buck converter
« Reply #4 on: May 04, 2020, 08:20:21 am »
There’s nothing wrong with a non-isolated converter, but be careful when testing! Your supply design power is small but there might be a lot of fault energy available in your test setup.

The selection of capacitors needs to hit a few things: low impedance against HF switching, high enough ripple current rating, enough energy to keep voltage somewhat stable over 50 / 60 Hz line cycle, and probably other stuff I’ve forgotten.

3x 10uF electrolytic is probably quite cheap. They are in series, so 3.3uF effective. Now derate by 50% for tolerance and ageing, get around 1.7uF effective.

If energy is the limiting factor, a 2.2uF film capacitor would be equivalent but more expensive.

For safety reasons you probably need some bleed resistors in there in any case.

Given the voltages in play, you should try to find operating voltage specifications for your inductor in addition to the usual inductance and current ratings.
 
The following users thanked this post: Red_Micro

Offline Red_MicroTopic starter

  • Regular Contributor
  • *
  • Posts: 122
  • Country: ca
Re: Input capacitor in buck converter
« Reply #5 on: May 04, 2020, 01:40:40 pm »
Given the voltages in play, you should try to find operating voltage specifications for your inductor in addition to the usual inductance and current ratings.

It's hard to find voltage specs for inductors. I chose this one to start with. https://www.digikey.com/product-detail/en/eaton-electronics-division/DR73-102-R/513-1815-1-ND/2666685
 

Offline TimNJ

  • Super Contributor
  • ***
  • Posts: 1720
  • Country: us
Re: Input capacitor in buck converter
« Reply #6 on: May 04, 2020, 02:45:55 pm »
There’s nothing wrong with a non-isolated converter, but be careful when testing! Your supply design power is small but there might be a lot of fault energy available in your test setup.

The selection of capacitors needs to hit a few things: low impedance against HF switching, high enough ripple current rating, enough energy to keep voltage somewhat stable over 50 / 60 Hz line cycle, and probably other stuff I’ve forgotten.

3x 10uF electrolytic is probably quite cheap. They are in series, so 3.3uF effective. Now derate by 50% for tolerance and ageing, get around 1.7uF effective.

If energy is the limiting factor, a 2.2uF film capacitor would be equivalent but more expensive.

For safety reasons you probably need some bleed resistors in there in any case.

Given the voltages in play, you should try to find operating voltage specifications for your inductor in addition to the usual inductance and current ratings.

Please use a 50/60Hz isolation transformer in front of your device under test, if possible. I know they can be a little costly, but may prevent you from getting hurt or worse. When I was younger, and didn't know any better, playing around with mains, GFCI may have saved my life once or twice too. Not a bad idea to have.
 
The following users thanked this post: Red_Micro

Offline TimNJ

  • Super Contributor
  • ***
  • Posts: 1720
  • Country: us
Re: Input capacitor in buck converter
« Reply #7 on: May 04, 2020, 03:04:55 pm »
Usually, the enameled wire used in inductors can withstand 2KV for diameters 0.3mm and up. 1.5KV for 0.2-0.3mm. In my experience, voltage breakdown is usually not so much of an issue, but every manufacturer does lead termination differently and may have varying amounts of isolation from the ferrite core. Maybe you can ask the inductor manufacturer if they have this data.
 

Offline TimNJ

  • Super Contributor
  • ***
  • Posts: 1720
  • Country: us
Re: Input capacitor in buck converter
« Reply #8 on: May 04, 2020, 03:17:51 pm »
Just noticed your 3-phase requirement. 3-phase GFCI is available, but pretty sure they are very expensive, and not usually rated at low enough current. 3-phase isolation transformers also are a little hard to come by, but probably available. I suppose you could just use 3 separate single phase transformers too. (I think that should work fine, though I'm not a 3 phase expert.)
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf