Input Capacitor selection For 450W Synchronous Buck

[biswaIITH]

Hello All,
Currently working on a Synchronous Buck Converter
Vi= (40-60) V
Vo=14 V
Io=30A
Fs=130Khz

I have calculated the input capacitor rms current rating .to be around 14.5A.
I don't want to use only Electrolytic Cap(100V El Cap) as it will increase the BOM cost to put 4 or 5 of them to meet current requirement

I am planning to use a combination of both ceramic as well as Electrolytic capacitor..5 Ceramic capacitors with required ripple current rating(2.5 A each) + 1 Electrolytic Capacitor(~200uF & 2.5 A).

Will the ripple current rating by Ceramic cap Varry from its datasheet value due to any external conditions other than temp??

 

[Easy peasy]

haha - what size & volt rating are the ceramic caps? - if the ripple current really is 14.5A rms you are going to have some fun if you run at full power for long periods ...

check out the C vs VDC applied rating for the smaller caps ...

 

[biswaIITH]

The Ceramic Cap is 100V,4.7uF .. Each Can carry 2.5 A rms with an ESR of around 0.006 Ohm. It will give rise in temp of around 15 degree as per the datasheet.
The size of cap is 1210 .

What problem do you think can happen?

 

[FvM]

I would use some more ceramic caps and drop the electrolytic, which is almost useless in the combination at 130 kHz. As mentioned by Easy peasy, did you look at actual C at 60V? 1210 is minimal available size for 4.7uF/100V, the C/V characteristic might be rather extreme. For Murata X7R capacitor in 2220, I see about 50% of rated capacitance at 60V. A Murata X7S capacitor in 1220 has only 25% capacitance remaining.

 

[biswaIITH]

Yes the ceramic capacitor % of reduction is around 66% at Vin=60V which results in a capacitance of 1.598uF from 4.7 uF.

I thought of adding some more ceramic cap & remove the Electrolytic Cap but then again input capacitance will not be sufficient to achieve a target input ripple voltage.

Any suggestion?

 

[FvM]

If the electrolytic capacitor reduces the ripple voltage considerably, it will also carry most of the ripple current, e.g. at least 10A rms.

 

[treez]

I doubt input ripple will be major with a decent ceramic cap bank.........you could always use an LC input filter.....and add a diode across the L so that you dont suffer high switch_on overvoltage ringing. (cathode to input connector)
As long as the ripple V at the input to the Buck isnt more than say 3V pkpk then it should be ok....and as long as they dont overheat......spray some black matt spray on them and view them through a thermal cam to quickly check temperature.

You could always use two linear.com bucks in parallel. and simply tie their error amplifier outputs together (they are gm type)...they will then share in parallel, and it cuts down your current as you have two channels...and can interleave them to reduce the input ripple current.

im pretty sure you can get a single chip with 2 buck controllers on it...and it automatically interleaves them and does the sharing for you.

Or use two LT8705s and sync them.

 

[KlausST]

Hi,

2.5A RMS at 0.006 Ohms gives a dissipated power of 33mW... how can this give a 15°C temperature rise?
This means an R_th of about 450K/W.

Klaus

 

[biswaIITH]

Hi,

2.5A RMS at 0.006 Ohms gives a dissipated power of 33mW... how can this give a 15°C temperature rise?
This means an R_th of about 450K/W.

Klaus

It is given in the datasheet

 

[BradtheRad]

Simulation of interleaved dual buck converters (mentioned post #7).

Notice reduced current levels, reduced stress on components.

The simple control circuit produces interleaved signals with adjustable duty cycle.

I omitted capacitors in order to clarify waveforms. Load ripple is not large even when smoothing capacitor is absent.

 

[KlausST]

Hi,

It is given in the datasheet

What brand and exact type? Are they suitable for your application? Datasheet?

Klaus

 

[biswaIITH]

Hi,


What brand and exact type? Are they suitable for your application? Datasheet?

Klaus

 

[treez]

That is interesting
The datasheet of #12 indeed shows a 15degc rise for 2.5a of ripple current....thats just 37mw of power.
A thermal resistance of over 400degc/watt.

We have discovered a different phenomina of heating in ceramic caps here.
I'd say this is due to eddy currents in the ceramic....or just the terribly poor thermal resistance of ceramic material.
Or maybe skin effect in the thin foil of the conductors in the cap.
And some proxmity effect in there too.

...or even all of the above.

It shows how MLCCs are pretty lousy with ripple current.

Then again, who cares about a 15degc rise....after all ceramic is very resilient to very high temperatures...unlike electro caps, ceramics can run at well over 125degc internal temp and still be fine for long life

 

[biswaIITH]

That is interesting
The datasheet of #12 indeed shows a 15degc rise for 2.5a of ripple current....thats just 37mw of power.
A thermal resistance of over 400degc/watt.

We have discovered a different phenomina of heating in ceramic caps here.
I'd say this is due to eddy currents in the ceramic....or just the terribly poor thermal resistance of ceramic material.
Or maybe skin effect in the thin foil of the conductors in the cap.
And some proxmity effect in there too.

...or even all of the above.

It shows how MLCCs are pretty lousy with ripple current.

Then again, who cares about a 15degc rise....after all ceramic is very resilient to very high temperatures...unlike electro caps, ceramics can run at well over 125degc internal temp and still be fine for long life

Yes.. I have seen some other parts from TDK, Murata Etc which are having worse Temp rise characteristics than the above

So i am considering 7-8 of above ceramic capacitors in parallel in the input side..

Should it be sufficient?

 

[treez]

Yes i would have thought so...as attached shows max 13arms of input ripple ( i assumed 10uh inductor at 130khz).....but beware if your ambient is very high, the capacitance falls dramatically with temperature above 125degc, so the ripple will get much higher

 

[biswaIITH]

Yes i would have thought so...as attached shows max 13arms of input ripple ( i assumed 10uh inductor at 130khz).....but beware if your ambient is very high, the capacitance falls dramatically with temperature above 125degc, so the ripple will get much higher

As it is EV for Application in India, You can expect External Ambient Temp as high as 55 Degree Celcius. Internally it should be limited to 100 degree celcius at the worst case scenario.

 

[treez]

Actually
Ripple voltage on output capacitor of SMPS.
The ripple voltage is the sum of the squares of the two constituent parts.
One part is from the actual peak current magnitude which goes into the capacitor and its product with the ESR of the capacitor.

V1 = I(PK) * ESR
……………………………………………………………
Other part is the actual charge that goes into the capacitor each switching cycle (this obviously equals the charge that goes out of the capacitor each switching cycle).

The voltage increase from this constituent is V2 = dQ/C
……………………………………………………………………..

And the overall ripple voltage is….
SQRT (V1^2 + V^2)
__________________________------------------------________________________-----------------________________

Now supposing that your seven paralleled 4.7uf caps, at high temperature become 1uf each
Then the dv of the input ripple voltage will be about 7V.
That may actually be ok, because that gives you a vin of 40+/-3.5v and 3.5v is less than 10% of 40v.

So yes i believe you will be ok.

Just make sure that any upstream filter L doesnt have a resonant frequency with the cap bank of 130khz.

 

[biswaIITH]

Hello All,

I am trying to simulate the above converter In PSPICE. I am using the LM5116 IC PSPICE Model & CSD19531Q5A Mosfet PSPICE Model downloaded from TI Website

I am facing the below error

"ERROR(ORPSIM-16362): Name on .ENDS does not match .SUBCKT"

Is it due to model files are corrupted???

 

[mrinalmani]

Just to share some experience, I've never actually had the need to use electrolytic capacitors for any buck converter at such frequencies.
And also, an X7R cap in 0805 package can easily handle over 3 amps of current. With a 1210 size cap, I will be expecting over 6A capability to begin with.

 

[biswaIITH]

Just to share some experience, I've never actually had the need to use electrolytic capacitors for any buck converter at such frequencies.
And also, an X7R cap in 0805 package can easily handle over 3 amps of current. With a 1210 size cap, I will be expecting over 6A capability to begin with.

Hii..
Yes 6A Can be drawn but also associated temp rise will also be higher