Aluminum electrolytic capacitor ripple current issue

[NyaaX_X]

Hello everyone.

I am working at a wireless power transfer circuit. And there we full-bridge rectifie the Rx ~128kHz AC signal. Pass through a P-mos ideal diode circuit, then we have 1x 100uF aluminum elelctrolytic capacitor parallel with 3x X5R--22uF mlcc, 1x X5R--10uF mlcc, 1x X7R--0.1uF mlcc. As the filter capactors @ ~16V. Then power the after stage.

We measured the ripple current of the E-cap. With extra lead wire and clamp probe. It seems the ripple current is over the rated current. So, we change that E-Cap to a hybrid polymer, 100uF Aluminum E-Cap as replacement.

The old one has a larger capacitance decay and larger ESR at this frequency compares to the new hybrid one. But the new one shows the more higher temperature increase when working with SMT two leads soldering on board pads. And more ripple current when we clamp measure it (ripple current freq. is 256kHz)

So, I am considering if it is going to the self-resonance or anti-resonance situation, caused that larger ripple current and temp-increase.(The on board ripple voltage peak to peak is ~150mV, the clamping test is ~410mV, ripple current ~800mA,rms)

Is that possible? Or the other reasons there? If so, How can I determine them?

Thank you~

[Twospoons]

One more thing to check is the switching of your pmos ideal diode - if the phasing is wrong then you would see higher than expected ripple current.  As a test, you could replace it with a schottky diode, then recheck your ripple current.

[NyaaX_X]

Thank you Twospoons for the replying.

But what is causing the different phasing as you say? We use two type of E-Cap and only one shows the obvious larger ripple.

Ah- I might describe not clearly. The wireless ~128k AC signal --> MOSFETs full-bridge rectifer --> P-mos ideal diode (P-mos, current mirror with discrete components, not ideal diode controller) --> E-CAP 100uF//MLCC 22uF//MLCC 22uF//MLCC 22uF//MLCC 10uF//MLCC 0.1uF --> post-stage (MCU/ D-D power/ etc.)

[davekni]

Likely several issues. Wire loop for current measurement probably alters circuit operation.  Current probes are likely to be susceptible to fields from wireless charging coil.  Usually best to measure voltage and calculate current based on capacitor impedance (C and ESR).  Typical ESR for new electrolytic caps is about half of worst-case specification.  Measuring voltage accurately also requires good technique.  Use small-diameter coax or twisted pair soldered to bypass capacitor.  Connect to scope (or scope probe) only once far from charging coil.

At 256kHz ripple current should be mostly in ceramic caps.  Bulk (electrolytic) cap is mostly for loop stability and to handle load transients.  I'd guess your ceramic caps aren't very good at operating voltage.  Many drop by 90% of rated capacitance at rated DC voltage.  Some large suppliers such as Murata and TDK publish at least typical curves for their ceramic caps.  Takes work searching, but find caps that are good at your use voltage.

Circuit layout could also be contributing to excess ripple.

Run LTSpice simulation.  If ripple voltage is far from simulated value, then look for PFET cross-conduction or other such issues.

[NyaaX_X]

Thank you Dave,

We tested the ripple voltage on PCB. For the Hybrid E-Cap, the Ripple Vpp is lower then the classic one... The main issue is.. If we just check the Temp. of them. The new one have the more Temp. increasing compared to the old one. Even the less V ripple of it. But as Spec. the ESR and high frequency capacitance decay of the new one is better than the old one.

Here is the compare. Vpp I use twisted OK wire to measure them.
And yes, the 22uF and 10uF MLCC now using the X5R...@~15V, rated 25V

Old (Chemicon MZR 100uF, ESR 260mohm@100kHz)
Final Temp ~52°C, Vpp ripple on PCB ~250mV, and Vpp ripple when clamping is ~300mV, Ipp ~400mA

New (Chemicon HXJ 100uF, ESR 50mohm@100kHz)
Final Temp ~65°C, Vpp ripple on PCB ~150mV, and Vpp ripple when clampling is ~410mV, Ipp ~800mA

Well.. I guess our circuit is not quite healthy.. For load transient changing like starting up... Sad

[klugesmith]

Is this for a hobby project or a commercial design?

[Twospoons]

But what is causing the different phasing as you say? We use two type of E-Cap and only one shows the obvious larger ripple.

I meant the relative phase between the switching of your MOS rectifiers and the signal you are rectifying. For example, slow turn off would result in current flowing back to the pick-up coil, and therefore higher ripple in your caps.

As I said - try it with schottky diodes and see what happens.

Also your 25V MLCC caps running at 15V will be at about 40% of rated value.   Murata have a useful tool for this stuff https://ds.murata.co.jp/simsurfing/index.html?lcid=en-us

[davekni]

Murata have a useful tool for this stuff https://ds.murata.co.jp/simsurfing/index.html?lcid=en-us

Yes, nice tools from some of the larger suppliers.

Also your 25V MLCC caps running at 15V will be at about 40% of rated value.

I've seen much worse even from reputable suppliers.  Just looked at that tool for 22uF 25V X5S parts.  Only one showed up: GRM188C61E226ME01.  It drops to 10% of rated value by 15V.

[Twospoons]

Yeah, its a "gotcha" that traps fresh graduate EEs all the time.

[NyaaX_X]

Yes,
Like this mlcc, the TDK products page shows the frequency/bias relation measurement plot in for each part.
https://product.tdk.com/en/search/capacitor/ceramic/mlcc/info?part_no=C2012X5R1E226M125AC

But I think the issue of different is between the two type of aluminum electrolytic capacitors, they have same package, same capcitance (but diffierent decay under frequency changing), same voltage rated. And the only changing to the circuit is these capacitors for these days question.

The two problems are:

(1)_I measured the ripple voltage cross to the E-CAP when they were soldered on PCB on pads. The hybrid one shows a lower ripple voltage compared to the original one, but the ΔTemp after few hours running of the Hybrid one is greater than the original one.

(2)_And if I connect a lead wire and clamp-measure them (same lead wire, and both of these two types are soldered with (-) on PCB pad. The clamp probe did affect the behaviors a few. And the result is hybrid one Vripple increase is more bigger compares to the original one.

I am only curious about the changing of this one E-Cap make the issue.

Thank you everyone~

[davekni]

Look back at reply 1 from Twospoons and first paragraph of reply 3 (mine), considering that ceramic caps are ~2uF rather than 22uF.  There is little value in providing more advice until you try the good advice already provided.

[NyaaX_X]

Actually, I do have apply a Sch diode SS34-1n5822 at that p-mos placing. The ripple voltage on that E-cap change from 150mV-->~137mV. And the E-Cap still heating up significantly. I am looking for the using experience of this kind of capacitors.

And for MLCC, I need to find the replacement later...

Thanks

[davekni]

Actually, I do have apply a Sch diode SS34-1n5822 at that p-mos placing.

Did you replace both PMOS devices?

The ripple voltage on that E-cap change from 150mV-->~137mV.

What current did you calculate from that ripple voltage?

[NyaaX_X]

OK, using 10X probe, I apply SCH diode to original one and hybrid one,

MZR series max ESR = 0.3ohm at 100kHz,
I get AC rms of 0.06824 mV --> / 0.3 = 0.2274 Arms Iripple


And HXJ series max ESR = 0.05ohm at 100kHz, but 256kHz is nearing its self-resonance Freq.
I get AC rms of V ripple = 0.02468mV --> / 0.05 = 0.4936 Arms Iripple

[davekni]

OK, using 10X prove, I apply SCH diode to original one and hybrid one,

Did you remove FETs when adding diodes?  If not, any possible cross-conduction issue may still exist.

MZR series max ESR = 0.3ohm at 100kHz,
I get AC rms of 0.06824 mV --> / 0.3 = 0.2274 Arms Iripple

As I'd mentioned before, typical initial ESR is around half of specification.  Current is likely about twice above value.

What current do you expect (from simulation or a rough estimate based on DC load current)?

[NyaaX_X]

Hello Dave,

Yes I have removed the p-mos.

These days I am simulating the simplified circuit in LTspice, some of the result ia close to the real circuit. But I still trying the different parameters. Later I will put the file here.

I was considering the lower ESR, the larger capacitance of capacitor could decrease the ripple. But it seems the appropriately of ESR could suppress them better? I have no idea...

[NyaaX_X]

I made a simulation. Here is the LTspice file.

For the bridge FETs, I picked one kind in the lib randomly. And the rectifier FETs I use Schottky diodes as replacements. So do the ideal diode.

I try to make a guess of the two kinds of electrolytic capacitor, parameters for them. So, please have a look in your free time~

Thanks!!

[davekni]

Update design in simulation until capacitor currents are all within part specifications with margin (with currents a ways below max rating).  Measure ceramic capacitor values at operating DC voltage to obtain accurate values to use in simulation, or at least use the manufacturer's c/v curves to estimate.