Small vs larger SMT ceramic capacitors and internal inductance?

[FlapJack]

30mV is a low ripple voltage. What spec are you trying to reach.

 

[D.A.(Tony)Stewart]

Hello,
We are doing a 65A, 1V5 SMPS using cascade of Vicor Power PRM48JH480T250A00 module and VTM48ET020T080A00 module. Vin=48V.
We are seeing 30mV of ripple at the output voltage of the VTM48’ module, and even if we increase the external output capacitance from 329uF to 2329uF the ripple stays pretty much at 30mV.
(As you’d expect, the ripple frequency is 3MHz, which is the switching frequency)
There is 310uF of output capacitance internal to the VTM48’ module.
So we conclude that we cannot appreciably reduce the output ripple voltage by increasing the output capacitance, because this ripple voltage is mainly caused by the 1 milliohm of internal resistance inside the VTM48’ module. (page 11 of VTM48’ datasheet shows this).
Therefore, we believe that we should only add some high frequency capacitance at the output of the VTM48’ module. As such we wish to use ceramic capacitance with low internal inductance.
Can you confirm that generally the lowest inductance ceramic capacitace is gotten by using the lower capacitance values and the larger case sizes?

PRM module datasheet:
http://www.vicorpower.com/documents/datasheets/PRM48BH480T200B00.pdf
VTM module datasheet:
http://cdn.vicorpower.com/documents/datasheets/VTM48E_020_080A00.pdf

The ripple frequency should be 6MHz with a bridge output.
The spec ESR is 1 mΩ or so which is about as good as you can get.

Consider a 1Oz square of copper is ~ 2/3 mΩ regardless if the square is 1mm or 100mm
remember that.

Inductance is purely a L/W ratio as well for straight conductors, which also affects ESR so it has to be bulked up a lot more than 1 oz copper.
This is why I said you need a thick busbar but a long aspect ratio to get the series inductance to the shunt caps.
A Power plane won't do this so if you use a power plane, then you need to drive it with ZL > 100mΩ @6MHz and Zc=10mΩ in order to get 20 dB attenuation of ripple. That can be done with a Murata low ESL X7R cap around 0.1uF with a SRF > 30MHz from impedance ratios.

If your worst ripple is 30mV at 65A, that's an ESR of 0.5mΩ, which is better than I expected.


With more effort on coil design you might be able to design a bigger L/R [nH/Ω] choke to achieve 40 dB or more at 60Amps.with Series loss <= 0.5mΩ to keep distribution ripple down.

But a far easier solution is a small 50mΩ ESR LC filter for each target processor.

 

[treez]

It is indeed likely a bridge inside the VTM48 and the ripple frequency is 3MHz.
The strange issue of sometimes lesser external output capacitance giving less Vpkpk ripple than higher value ceramics is possibly due to parasitics of the vtm48 eval board or the resistive dummy load, and the situation is complicated in that there may be power stage resonances popping up with certain external Cout values, and also noise issues.
Certainly the VTM48 output cannot be scoped with a "tail" on the probe, whereas with other power supplies, as you know, you can get away with it for a genral look.
I am wondering if anyone has experience with this vtm48 module, they are extremely common worldwide.

- - - Updated - - -

30mV is a low ripple voltage. What spec are you trying to reach.

thanks, yes, its low but its 2%.
We would really like 1% or less...but maybe our chip will tolerate the 2%...as Sunny said, we can have LC filters local to each chip and address the problem like that.

- - - Updated - - -

Certainly not. Just look up the SRF of the two from a manufacturer. This is the reason you'll probably never find a 1206 capacitor in a broadband RF layout. What's the MPN of the capacitors you're using?

Theyre just standard murata ceramic caps from one of their "book" thingys. Ill get the MPN.

I agree now the smaller packages have lower ESL, though i will pull the datasheets for relationship of ESL with capacitance of a given case size.

 

[treez]

I have reported here that when using a single 4u7,0603 capacitor at the output of the VTM module, we got LESS pkpk ripple voltage than when using a 47u, 0603 capacitor. The attached two datasheets seem to explain why this is. Because at 3MHz, (the output ripple frequency of the VTM module), the impedance of the 47u capacitor is actually greater than the impedance of the 4u7 capacitor.
3MHz is in fact beyond the self resonance frequency of either capacitor.

Do you agree that this explains this?

 

[FlapJack]

All of the dimensions of the 4.7uF capacitor are smaller than the 47uF capacitor. Therefor less inductance.

The "impedance vs ESR Frequency" chart seems to confirm this.

 

[FvM]

3MHz is in fact beyond the self resonance frequency of either capacitor.

According to the datasheet it's not. 4.7 uF capacitor resonance is very near to 3 MHz which makes the impedance lower for this particular frequency.

Neither of the datasheets is for a 0603 capacitor. In so far the considerations can't refer to it.

 

[treez]

Neither of the datasheets is for a 0603 capacitor. In so far the considerations can't refer to it.

woops yes, the attached are now for 0603 sized caps, and indeed the 47u one does still have greater impedance than the 4u7 one at 3MHz.

 

[FvM]

Halfway there...

 

[FlapJack]

Third time is the charm, the 47uF is still not an 0603.

 

[treez]

Thanks!....woops again, here is the 4.7uF 0603 and the 47uF 0603.......indeed, the 4u7 has LOWER impedance at 3MHz.
So this explains why the ripple was less with the 4.7uF one.

 

[FvM]

I see same |Z| of about 15 mOhm, differences are surely below expectable type variation.

 

[treez]

I see same |Z| of about 15 mOhm

Theyre both not too far away from that, but the 47u cap is definietly got higher impedance than the 4u7 at 3MHz.

differences are surely below expectable type variation.

Thanks, by this you mean you would expect the impedance of them to be more different at 3MHz?

 

[FlapJack]

I see that the 4.7uF cap is showing it's peek at 3 Mhz and the 47uF has it's peak about 900 Khz. After the peaks each capacitor looks more inductive than capacitive. But i am not sure if this explains the difference in mV or not.

Also, it is very easy to ***** SMD components when hand soldering them, maybe you should try another 47uF or try to measure it.

- - - Updated - - -

Ummm, why was the word c-r-a-c-k stared out in my post #33.

 

[D.A.(Tony)Stewart]

if f transformer is 3MHz and rectifier has no 6Mhz ripple then the residual noise is not coming from the bridge caps but from coupling of inverter either CM or DM.

Ground /supply plane coupling between inverter and rectifier may cause this. Consult with CSE.

Either way 10 mΩ is possible at 3MHz with PU caps and Murata Special X7R's and in some cases 1mΩ by cascaded arrays side by side so low ESL with high W/gap ratio. ============== like this