Hi Jaikumar,

If you are sure that the 100Hz ripple is reflected from the inverter side (we can't be too sure, as we have not seen your detailed schematic or waveforms), then I doubt if anything can be done about it by controlling the DC-DC converter. Rather, if we don't take adequate care, the ripple can lead some issues for the controller !!

My suggestion will be to install a 100Hz tuned filter on the intermediate link. That should eliminate the ripple midway, and prevent it from messing up both the DC and the AC side. The design should involve the tuning frequency of 100Hz and adequate quality factor to preferably accommodate the 200Hz harmonic as well.

-Sanjay

Hi Jaikumar,

The 100Hz ripple is inevitable. Note that the AC-side power pulsates at 100Hz. Let us assume a pure resistive load:

P = Vcos(2pi50t).Icos(2pi50t) = 0.5VI(1+cos(2pi100t)), therefore, the AC-side requires a 100Hz component and it has to be supplied by the DC source. This 100Hz component is present in both the inductor current (DC input current) and the DC-link voltage. Based on the source's vulnerability to providing 100Hz component, you may design controllers to trade off ripples in input current with the DC-link voltage.

You may refer to the work below for ideas on achieving the trade-off:

https://arxiv.org/pdf/1604.03573.pdf

Regards,

Mayank

Hello Jay,

I agree that you need to deal with the 100Hz ripple at a more serious level but try also to figure out proper grounding in your multi-stage design. Sometimes, improper grouds make a huge difference in those ripples. Filters alone may not be as effective as we expect so you will have to address the Ground currents as well.

Hope this helps you in positive way.

Best wishes!

I assume that you dont have any noise issue, the two converter, taken as single unit, are working correclty.

I assume also that the two converters are well designed (input filter output filter, ecc)

For the first stage (dc-dc):

First to be considered, you have two cascaded converters. The dc-dc unit sees a negative restance as a load: if its output voltage goes up, the load current will go down, as the second converter will keep power constant ; this may generate oscillation in the dc link. I dont know how to mitigate it ( if its possible maybe you should link the two controls), once I had a similar problem on a real sistem, I accepted a slight oscillation and everythig worked well.

For the second stage (dc-ac)

If you cannot remove the 100 Hz from the dc-dc output, use feedforward: in your modulator correct the pwm value by a factor  K/Vdc, where K is the amount of feedforward. Feedforward moves the PWM with no intervention of the regulator

Feedforward can be also used in the dc-dc. 

I agree with Guido Botto. Feedforward is a good technique to deal with DC voltage ripple.  

First understand that what you describe is called LOAD REGULATION which is simply the ratio of source to load impedance for modulated load currents or step loads.  We know negative feedback gain reduces output and input impedance at the each respective port. You also have to manage the choice of low RdsOn or Rce of your switch (IGBT, Mosfet diode etc), the topology  and it's on/off duty cycle as well as apply the Z^2 effects of any voltage transformation whether it is DC-DC or AC-AC or DC-AC or AC-DC.  So manage your Zout to be lower and you will see improvements on load regulation on source. I refer to Zout also as ESR and ESL(f). depending on the source. Transformer inductance and coupling are also factors as well as switching rates and their implied impedance as well cap bulk Cap ESR  and more. Each contributes to the result. of ESR. Spectral admittance is often used to measure it or simply ESR depending on where/how you test it.  Panasonic's 600V X-GaN FETs are pretty great!  eg https://www.semicon.panasonic.co.jp/rd/?id=3010-PGA26E08BV_E

You can reduce the ripple in the DC input (at the cost of higher DC link voltage ripple). We had the same issue in our converter. We used resonant control.

Conference Paper Current ripple reduction in 4kW LLC resonant converter based...

Kalhana as much as I like your primary series resonant step-up transformer with duty cycle controlled secondary, is DC supply current constant at full load? If not then pulse load regulation effects apply for source/load impedance ratio. Active PFC is often the preferred way to stepup to create a linear primary load to eliminate AC secondary ripple on source,  then a step-down DC-AC bridge is used to  instead. It may be not practical to expect no battery ripple with non-linear or switched reactive AC loads on secondary as this would imply inductive energy storage of ~>8 cycles at rated power for

Just to clarify, it was a resonant controller to control a resonant converter.

The resonant controller provides high controller gain at the 100 Hz frequency thereby nearly eliminating the 100 Hz current ripple on the DC side. So you can have high gain at the desired frequency only and still have loop stability.

most likely you have underestimated the storage time constant using LEDs which may have 10% of your linear load line thus insufficient C*ESR=T relative to 10ms.