Capacitor Ripple Current Ratings on Datasheet in mArms?

Thread Starter

jim0000

Joined Oct 28, 2020
130
For the capacitor with product number HHXA350ARA271MJA0G I am trying to find the ripple current rating to make sure I have enough in parallel. I am stuck since the datasheet has it in mArms and I have no clue what that is. How does that relate to Amps so I can find the Irms limit of the capacitor instead of mArms, and then from there find out how many caps I need in parallel to account for that limit? The circuit and waveforms from LTPowerCAD is below. My LTspice waveforms show the input output capacitor currents which are seperated by capacitor type. I did this to get a better idea of ripple although LTpowerCAD gave expected Irms I am sure it is still helpful. I am very curious why my input current through my caps has such a huge spike? Then there appears to be no input ripple after the current spike.

Input: Up to 4 solar panels in parallel (12V 100W panels)
Output: 100Ah or 200Ah battery (12V)

Example capacitor datasheet:

1667356714206.png

LTspice Waveforms:

Output Current for 10ohm load (I noticed the output ripple is actually slightly less but input current through the caps are the same for a load with small resistance such as 1 ohm or less.)

Also the current is expected to be a little higher since the load will be a battery with only small internal resistance. When I run it at 1 Ohm or 0.02 Ohm or anywhere around there output current is 25A with the same characteristics basically which is desirable since I am limiting the output current by design to 25A.


1667571620894.png



Capacitor current from LTspice:
Cinc1 input ceramic capacitor current: 10 Ohm load
1667570240149.png



Cinb1 Bulk input electrolytic capacitor current: 10 Ohm load
1667570348009.png



Coc1 Output ceramic capacitor current: 10 Ohm Load
1667570740995.png

Cob1 Output Bulk electrolytic capacitor current: 10 Ohm load
1667571141530.png


1667479147269.png

Boost Mode Load Transient:
1667479190497.png

Buck Mode Load Transient:
1667479235225.png


1667479322365.png
 
Last edited:

crutschow

Joined Mar 14, 2008
31,123
RMS current value is used, since that is proportional to the heating of the capacitor due to its ESR, which determines the ripple current limit.
 

Papabravo

Joined Feb 24, 2006
19,578
Keep in mind that the average DC current will not cause the ESR to dissipate power. It is only the AC component, aka the ripple current, that will cause a problem. So, if we look at the 22μF part with a rating at 100 kHz of 1100 mArms, that is equivalent to 1.1 Arms and the power dissipated in the ESR of 45 mΩ will be:


\( P_{D}\;=\;(I_{rms})^2R\;=\;(1.1)^2(0.045)\;=\;54.5\text{ milliwatts} \)

Does that make a bit more sense?

You may have to look elsewhere in the datasheet to estimate the temperature rise you can expect for that power level. Also the ESR is specified at 20 °C. It may be different when the capacitor warms up. I'm not familiar with polymer hybrid aluminum electrolytic capacitors, but the old garden variety aluminum electrolytic capacitors had pretty crappy temperature characteristics.
 
Last edited:

ronsimpson

Joined Oct 7, 2019
2,527
Does that make a bit more sense?
From what we know, that is at 120C. The capacitor is already near max temp.
This is a little 10mm cap.
1667360994064.png
We need to know the circuit. Maybe there is 1A passing by and the cap sees very little current, or, there is a flyback PWM and the ripple current is much higher then the output current.
 

Thread Starter

jim0000

Joined Oct 28, 2020
130
Keep in mind that the average DC current will not cause the ESR to dissipate power. It is only the AC component, aka the ripple current, that will cause a problem. So, if we look at the 22μF part with a rating at 100 kHz of 1100 mArms, that is equivalent to 1.1 Arms and the power dissipated in the ESR of 45 mΩ will be:


\( P_{D}\;=\;(I_{rms})^2R\;=\;(1.1)^2(0.045)\;=\;54.5\text{ milliwatts} \)

Does that make a bit more sense?

You may have to look elsewhere in the datasheet to estimate the temperature rise you can expect for that power level. Also the ESR is specified at 20 °C. It may be different when the capacitor warms up. I'm not familiar with polymer hybrid aluminum electrolytic capacitors, but the old garden variety aluminum electrolytic capacitors had pretty crappy temperature characteristics.
Thank you, I have updated my post to include the circuit and the associated waveforms if that helps. It definitely makes more sense. Based off of the waveforms, about how much current ripple should my caps be prepared to handle? Should only a certain type of caps be used? Right now, I was looking at doing xr5 or xr7 ceramic caps, and the datasheet also called for some electrolytic caps to be used as well on the input and output.
 

Thread Starter

jim0000

Joined Oct 28, 2020
130
From what we know, that is at 120C. The capacitor is already near max temp.
This is a little 10mm cap.
View attachment 279737
We need to know the circuit. Maybe there is 1A passing by and the cap sees very little current, or, there is a flyback PWM and the ripple current is much higher then the output current.
Thank you I updated my post with the circuit, load, and input. I included design curves from LTPowerCAD as well.
 

Papabravo

Joined Feb 24, 2006
19,578
Thank you, I have updated my post to include the circuit and the associated waveforms if that helps. It definitely makes more sense. Based off of the waveforms, about how much current ripple should my caps be prepared to handle? Should only a certain type of caps be used? Right now, I was looking at doing xr5 or xr7 ceramic caps, and the datasheet also called for some electrolytic caps to be used as well on the input and output.
The rating on the 22μF cap is 1100 mArms. That RMS value is equivalent to:

\( 1100\text{ mArms}\;\times\;\sqrt{2}\;=\;1.56\text{ A Peak} \)

This means that the capacitor can handle your average output current ±1.56 A. In a CCM (Continuous Conduction Mode) design your average output current should be in excess of 1.56 A in order to ensure that there is continuous conduction. A practical rule of thumb is to set the peak to peak current ripple at 20% of the average current. Your mileage may vary.
 

ronsimpson

Joined Oct 7, 2019
2,527
In LTSPICE you can plot the current in each capacitor. Probably you should enter values for ESR and ESL first.
You should see a plot like this. (Middle red line=0) I do not care about the step, that is not important. I drew lines showing P-P current. P_P/2=peak Then x0.707 = approx. RMS. (Assuming a sine wave) (a ramp/sawtooth is slightly different) You need to know if you are 10X over current or 1/2 current. The 10% error, about how you got from Peak to RMS is not important.

During start up and during a step the current gets very high but that is only for a short time. You want to know what the RMS current is at the switching frequency steady state.
1667484827023.png
I do not know what condition causes the largest ripple at the switching frequency. Likely heavy load and low input voltage. I do not do buck-boost often. But most PWM have heavy ripple at low input voltage. You should look at both Hi input and Low input.
 

Thread Starter

jim0000

Joined Oct 28, 2020
130
In LTSPICE you can plot the current in each capacitor. Probably you should enter values for ESR and ESL first.
You should see a plot like this. (Middle red line=0) I do not care about the step, that is not important. I drew lines showing P-P current. P_P/2=peak Then x0.707 = approx. RMS. (Assuming a sine wave) (a ramp/sawtooth is slightly different) You need to know if you are 10X over current or 1/2 current. The 10% error, about how you got from Peak to RMS is not important.

During start up and during a step the current gets very high but that is only for a short time. You want to know what the RMS current is at the switching frequency steady state.
View attachment 279818
I do not know what condition causes the largest ripple at the switching frequency. Likely heavy load and low input voltage. I do not do buck-boost often. But most PWM have heavy ripple at low input voltage. You should look at both Hi input and Low input.
Okay thank you I actually do have the buck-boost on LTspice and I noticed the ripple in the output cu
The rating on the 22μF cap is 1100 mArms. That RMS value is equivalent to:

\( 1100\text{ mArms}\;\times\;\sqrt{2}\;=\;1.56\text{ A Peak} \)

This means that the capacitor can handle your average output current ±1.56 A. In a CCM (Continuous Conduction Mode) design your average output current should be in excess of 1.56 A in order to ensure that there is continuous conduction. A practical rule of thumb is to set the peak to peak current ripple at 20% of the average current. Your mileage may vary.
Thank you so I could basically compare that capacitor rating with my simulated I_RMS through my caps from LTpowerCAD thats pictured above, and if the I_RMS current of my cap is much smaller how many should I have in parallel to account the shortcoming?
 

Thread Starter

jim0000

Joined Oct 28, 2020
130
In LTSPICE you can plot the current in each capacitor. Probably you should enter values for ESR and ESL first.
You should see a plot like this. (Middle red line=0) I do not care about the step, that is not important. I drew lines showing P-P current. P_P/2=peak Then x0.707 = approx. RMS. (Assuming a sine wave) (a ramp/sawtooth is slightly different) You need to know if you are 10X over current or 1/2 current. The 10% error, about how you got from Peak to RMS is not important.

During start up and during a step the current gets very high but that is only for a short time. You want to know what the RMS current is at the switching frequency steady state.
View attachment 279818
I do not know what condition causes the largest ripple at the switching frequency. Likely heavy load and low input voltage. I do not do buck-boost often. But most PWM have heavy ripple at low input voltage. You should look at both Hi input and Low input.
Interesting, I have the design on LTspice and I just attached the waveforms if you want to take a look. Really concerned that my input current spikes so high I am not sure why that is happening.
 
Top