SMPS to linear voltage regulator for extra smoothing?

jlawley1969

Joined Feb 22, 2021
97
Hi,
I am making a flyback converter circuit switching at 132kHz (using) and on the secondary side for 5V 2.2A the application notes are suggesting a 2200uF/10V filter capacitor. That capacitor is far too big for my design constraints and while I am going to be outputting 1.2A max the comparable downsized capacitor is still too big.

Now the main thought is that I am using a SMPS to output 5V then using a 5V linear voltage regulator(like LDL212PU50R). The idea is that the voltage regulator would provide stabilization and hopefully smoothing the ripple. The issue is I don't know if the voltage regulator does any of the things I said in the previous sentence. I've read that linear voltage regulators actually impart MORE noise to the signal, so am I wasting my time/hurting my self by adding a secondary voltage regulator?

This is going to be for a dedicated USB power supply so the signal will need to be coherent enough so that cellphones can be charged without the phone throwing a fit because I've seen some so bad that phones just refused to charge all together(I know about command resistors and such)

Ian0

Joined Aug 7, 2020
6,668
The ripple Current of the capacitor is much more important than its capacitance. Only 36uF is required to keep the ripple below 0.25V. A larger capacitor may have been specified in order to find one with a high enough ripple current rating.

ronsimpson

Joined Oct 7, 2019
2,519
Try splitting the cap into two. Then make a CLC filter. 470uF 100uH 470uF. You will need a 2.2A inductor.

Last edited:

AlexS4

Joined Sep 8, 2021
5
as very rough estimate of ripple voltage use: C=q/u~=I/(6FU)
C- capacity of ideal capacitor
U-ripple voltage(1 period sag)
F-lowest spectrum frequency
so for 130kHz,1.2A,100mV 1.5uF or even less would be quite enough ...
but in real world with resistance and inductance i d suggest 3.3uF ceramic capacitor X5R,X7R or ~22uF10V tantalium capacitor. or both of them for paranoid design )).

ElectricSpidey

Joined Dec 2, 2017
2,179
You can't output 5 volts from a linear regulator with a supply of 5 volts, because of the dropout voltage so you would need to output more than 5 volts from the SMPS.

Ian0

Joined Aug 7, 2020
6,668
as very rough estimate of ripple voltage use: C=q/u~=I/(6FU)
C- capacity of ideal capacitor
U-ripple voltage(1 period sag)
F-lowest spectrum frequency
so for 130kHz,1.2A,100mV 1.5uF or even less would be quite enough ...
but in real world with resistance and inductance i d suggest 3.3uF ceramic capacitor X5R,X7R or ~22uF10V tantalium capacitor. or both of them for paranoid design )).
I think you’re a little bit on the low side
for C=q/v
t=7.7us
I=1.2A
q=9.24 uC
v=100mV
that gives C=92uF
t is debatable, depending on how much of the time the capacitor is actually being charged

AlexS4

Joined Sep 8, 2021
5
>>> 1.2/(13e5*1e-1)
9.23076923076923e-06
9.2uF for infinite duty factor ?) ok 10uF ceramic. sorry.

>>> 1.2/(1.3e5*1e-1)
9.23076923076923e-05
92uF for infinite duty factor, you are right! sorry.

Last edited:

Ian0

Joined Aug 7, 2020
6,668
>>> 1.2/(13e5*1e-1)
9.23076923076923e-06
9.2uF for infinite duty factor ?) ok 10uF ceramic. sorry.
130kHz =130*10^3 = 13*10^4 = 1.3*10^5

AlexS4

Joined Sep 8, 2021
5
...fixed ideal formula: C=I*D/(F*U)
D= max discharge duty factor = 1-min_charge_duty_factor.
it should be quite accurate for relatively low ripple voltages and reasonable(common) duty factors... but it considers no esr and esr-freq. for al-oxyde capacitors(especially small ones) this correction may be spectacular.

so 33..100uF x10V tantalium capacitor 293D Vishay f.e. would be suitable i suppose.

ps never seen common IC linear regulators capable process well 100+kHz switching ripples (with ~1Mhz significant spectrum)

jlawley1969

Joined Feb 22, 2021
97
Does duty factor=duty cycle?
the switching IC I am using (which I forgot to write down the first time) is TOP257GN-TL
which has a max duty cycle of 78%

also why is there such a discrepancy in the required capacitance values between electrolytic, tantalium, ceramic?

So, I should get rid of the secondary VR and just properly size capacitor?

AlexS4

Joined Sep 8, 2021
5
1. D= 1- _min!_duty_cycle %/100 or D=1 as the worst case.
2. these capacitors have very different common esr and it's frequency dependence.

meaning above ambiguitis it seems just prototyping+maesurements would be better way to prevent remortgaging.

3.yes.

Last edited:

AlexS4

Joined Sep 8, 2021
5
to prevent excessiveness or insufficiency i meant

ronsimpson

Joined Oct 7, 2019
2,519
The ripple Current of the capacitor
Before you get too fare, look to see if the output capacitor is rated for 2A of ripple current.