Previous in Forum: PLC   Next in Forum: Peak Shaving Methods
Close
Close
Close
10 comments
Rate Comments: Nested
Active Contributor

Join Date: May 2008
Posts: 12

Noise Filtration of DC / DC Converters

08/01/2009 12:36 PM

hi

i was studing a filtring circuit of a dc to dc converter. it drowned me into several questions.

1: why ceramic and electrolytic both sort of capacitors are usaully placed in parallel in the filter circuits of dc to dc converter?

2: how should we select a particular value of capacitor?

3: how should the design proceed? for example i have a converter delivering 15v, 3A power signal with 100mV pk to pk noise. it has its capacitive load of 1000uf. then what should be my first step? placing a capacitor? electrolytic or ceramic? or an lc filter?

can anyone reply these quiries and share any information in this regard

Register to Reply
Interested in this topic? By joining CR4 you can "subscribe" to
this discussion and receive notification when new comments are added.

Good Answers:

These comments received enough positive votes to make them "good answers".
Guru
Popular Science - Weaponology - New Member United Kingdom - Member - New Member

Join Date: May 2007
Location: Harlow England
Posts: 16512
Good Answers: 670
#1

Re: noise filtration of dc to dc converters

08/01/2009 12:59 PM

1. The small value ceramics are better at decoupling high frequency noise. The electrolytics work better at lower frequencies. Thus the combination of the two is better than either on it's own.
2. Experience, trial and error, looking at manufacturers data sheets, or calculation and subsequent adjustment by trial and error.
3. It's not just what you put in but where...see #2. Layout, especially of earth/ground tracks is critical.
Del

BTW... If you read the instructions on posting questions you wouldn't have posted twice. This could be indicative of your design problems....

__________________
health warning: These posts may contain traces of nut.
Register to Reply
Guru

Join Date: Oct 2008
Posts: 42355
Good Answers: 1693
#2

Re: Noise Filtration of DC / DC Converters

08/01/2009 1:56 PM

Claws gradually retract. Cat, naps.

Register to Reply
3
Guru

Join Date: Sep 2006
Posts: 4513
Good Answers: 88
#3

Re: noise filtration of dc to dc converters

08/01/2009 2:40 PM

student1 queries: "1: why ceramic and electrolytic both sort of capacitors are usually placed in parallel in the filter circuits of dc to dc converter?"

Ceramic capacitors bypass transients and other high-frequency components in the output far more effectively than do electrolytic capacitors. That is why both are connected in parallel.

"2: how should we select a particular value of capacitor?"

See below.

"3: how should the design proceed? for example i have a converter delivering 15v, 3A power signal with 100mV pk to pk noise."

Commercial DC-to-DC converters are effectively switching (as opposed to linear) power supplies. The switching frequency is typically very high - often in the MHz region. DC-to-DC converters can radiate significant levels of harmonic-rich RF that can interfere with other parts of the system. You can minimize this problem by installing the filter (and ceramic) capacitor as close to the converter as possible.

You mentioned that 100 mVpp 'noise' is riding on the DC output. Is this truly noise, which is random, or is it ripple, which is at the converter's switching frequency?

To compute the size of the filter capacitor you will need to determine the ripple frequency, the maximum expected output current, and you must also decide the peak ripple amplitude you are willing to tolerate. Keep in mind that smaller amplitudes require more capacitance. You cannot eliminate ripple entirely, as this would require (theoretically, at least) an infinitely-large filter capacitor. These can be a bit difficult to find.

To determine the minimum size of your filter capacitor, let's begin with the definition of capacitance and use that to determine the size of capacitor we will need; thus:

I = C·dv/dt (eqn 1)

where,

C is the capacitance, in Farads,

I is the maximum expected current, in Amperes,

and

dv/dt is the instantaneous rate-of-change in ripple voltage, in volts per second.

Thankfully, we can approximate dv/dt by substituting Δv/Δt. This approximation greatly simplifies the calculations without incurring significant loss in precision.

Solving for C, we get

C = I·Δt/Δv (eqn 2)

where

Δt is the period of the ripple voltage and is computed as

Δt = 1/fS (eqn 3)

where

fS is the converter's switching frequency,

and

Δv is the maximum ripple amplitude our design will tolerate, in Vpp.

Let's work an example. Say we have:

fS = 300 kHz,

Δv = 10 mVpp,

and

I = 3 A.

From these the minimum capacitance necessary to meet our requirements is:

C = (3 A) * (1 / 300e3 Hz) / (10e-3 V)

C = 1e-3 F

= 1000 μF

Hope this helps.

-e

Register to Reply Good Answer (Score 3)
Guru
Popular Science - Weaponology - New Member United Kingdom - Member - New Member

Join Date: May 2007
Location: Harlow England
Posts: 16512
Good Answers: 670
#4
In reply to #3

Re: noise filtration of dc to dc converters

08/01/2009 5:36 PM

Yeah..that's wot I said

__________________
health warning: These posts may contain traces of nut.
Register to Reply Off Topic (Score 5)
Guru
Engineering Fields - Electrical Engineering - Been there, done that. Engineering Fields - Control Engineering - New Member

Join Date: Dec 2008
Location: Long Island NY
Posts: 15600
Good Answers: 981
#7
In reply to #3

Re: noise filtration of dc to dc converters

08/03/2009 9:40 AM

Nice job europium, I'd give you a GA but you're well past the GA threshold.

One thing though that makes designing switching supplies interesting. The accurate equations you cite imply that increasing the switching frequency will permit smaller value filter components to obtain the same ripple response. This is true as long as the circuit remains a lumped element design for the higher switching frequency and its harmonics. In other words as the switching frequency increases, the parasitic inductors, capacitors and resistances may no longer be trivial components. Add to that the complication from possible coupling to free space, at higher frequencies short wires become antennas, and you'll see that a low noise switching supply can seem like an oxymoron.

This detail is getting beyond the original question, but you did such a superb job europium that I just wanted to dot a few of the remaining i's and cross a few t's.

Again, nicely done.

__________________
"Don't disturb my circles." translation of Archimedes last words
Register to Reply
Guru

Join Date: Sep 2006
Posts: 4513
Good Answers: 88
#8
In reply to #7

Re: noise filtration of dc to dc converters

08/03/2009 5:57 PM

Thanks.

Btw, I'm building a nanoswitcher that uses optical waveguides exclusively. To hell with microstriplines. Gawd, what a bother!

Seriously, though, I didn't want to bog student1 down with details he/she won't encounter for a few years yet. Best that student1 take a couple of courses in RF circuit design before engaging in discussions about distributed impedance and all that other crap you need to know about just to design a simple motherboard these days (remember the S100 bus? I hope not). By then student1 will probably be a CR4 Mage answering our questions!

Register to Reply
Guru
Popular Science - Weaponology - New Member United Kingdom - Member - New Member

Join Date: May 2007
Location: Harlow England
Posts: 16512
Good Answers: 670
#9
In reply to #8

Re: noise filtration of dc to dc converters

08/04/2009 6:50 AM

Those optical transformers are a pain to wind, and finding optically permeable ferrite for the cores is tricky, or are you machining some out of Unobtainium? (optical grade of course)
Del

__________________
health warning: These posts may contain traces of nut.
Register to Reply Off Topic (Score 5)
Guru

Join Date: Sep 2006
Posts: 4513
Good Answers: 88
#10
In reply to #9

Re: noise filtration of dc to dc converters

08/04/2009 8:30 AM

When an ignoble atom (argon, for example) is hit by an intense laser pulse that has an E-field strength comparable to the Coulomb field of the atom, the atom's outermost electron is stripped away. Once freed the electron is accelerated by the E-field, first moving away from the ion, then back toward it as the field reverses polarity. The electron recombines with the ion and releases its energy as a photon. The light is emitted at every peak of the E-field component, producing a series of attosecond light flashes. The photon energies generated by this process can extend past the 800th harmonic order - up to 1300 eV - which makes for some real kick-ass optical transformers. The laser, of course, must be linearly polarized so that the electron returns to the vicinity of the parent ion instead of going round and round and round like some bloody drunken quantum sailor.

Unfortunately, argon crystals are in short supply these days and so I've abandoned the optical approach in favor of a much simpler technology which I am developing: DC transformers.

Know where I can find some magnetic monopoles? Radio Shack is fresh out.

Register to Reply Off Topic (Score 5)
Guru

Join Date: Sep 2006
Posts: 4513
Good Answers: 88
#5

Re: Noise Filtration of DC / DC Converters

08/02/2009 3:33 AM

"... and electrolytic ..."

student1, if you are actually planning to build this thing, and if your switching frequency is in the 100s of kHz (or more), it is highly recommended that you use tantalum capacitors instead of electrolytic ones. Tantalum capacitors do a much better job of filtering at high frequencies than do their sluggish, electrolytic cousins. You will still need the ceramic cap for the reasons Del* gave earlier.

(*Happy now? )

Register to Reply
Guru

Join Date: Feb 2006
Posts: 1758
Good Answers: 6
#6

Re: Noise Filtration of DC / DC Converters

08/02/2009 6:40 PM

Regards

"

1: why ceramic and electrolytic both sort of capacitors are usaully placed in parallel in the filter circuits of dc to dc converter?"

electrolytics are Hi capacity storage & ceramics are "Low-Series-Resistance" to pass Hi freq specially pulses [containing hi-harmonices]

Register to Reply
Register to Reply 10 comments

Good Answers:

These comments received enough positive votes to make them "good answers".
Copy to Clipboard

Users who posted comments:

Haajee (1); lyn (1); redfred (1); user-deleted-1105 (3); user-deleted-13 (4)

Previous in Forum: PLC   Next in Forum: Peak Shaving Methods

Advertisement