Hi i have designed the power supply of 500W(400VDC input and 48VDC out put) ,with phase shifted full bridge topology by using ISL6754 IC. and my transformer turns ratio is 16:3 but i am seeing the primary current ringing with 10MHz frequency. , please help how make my current wave form ringing free?

Please take a look on the secondary side during the ringing - you will probably see high ringing voltage across the rectifiers.

thanks for reply,
yes secoundary voltage is following primary current ringing with peak voltage spike of 200V

Can you post some waveforms?

For the phase-shifted bridge experts out there, I have a question:

This ringing happens on the secondary of the phase-shifted bridges, and there are various fixes for it. Can anyone comment on why it doesn't appear in the papers that talk about the phase-shifted bridge with synchronous rectifiers? Is there something in the timing that eliminates the ringing or clamps it somehow?

In phase shifted full bridge topology:
Ringing causes on secondary side is just because of transformer's leakage inductance and rectifier diodes body capacitor.in short they are resonating at some frequency like MHZ. to suppress this ringing we have to put saturable reactor in this path.it will block some ringing generated at secondary side .and thats why no ringing transfer at primary side.

I guess, it is not necessary to explain that the ringing occurs due to difference between the primary current and secondary filter inductor current. It happens because of parasitic charge associated with Qrr of the secondary diodes. There are several methods of addressing this issue and voltage clamping on the secondary side is one of them.

When people work with sync rectifiers the ringing is still present if special precuasions (including "soft switching") are not taken. The presence of soft switching is necessary but not sufficient condition to prevent the ringing. All solutions are either described in the literature or require some off line discussions.

Either Qrr or trr of the diode and the Leakage of the transformer play main role in this ringing, it can be suppress by using saturable reactor or a tinny ferite bead/ ring on the diode lead/leg, or an RC snubber but it is lossy. The amount of ringing or peak-to-peak of ringing can be manupulated in such a way that by the speed of your main mosfet driving, means slower it down to a manage-able level by ferrite bead or snubber.

By rules of thumb, first, select a low trr diode, like 200V SBD diode, 2nd, design the PCB with shortest current circulating path either both Primary & Secondary switching path, so that the trace contributed inductance is small, 3rd, design the transformer with lowest leakage inductance, then quantify the best speed of driving for ON and OFF moment of mosfet, lastly, design a best snubber or select the best available ferrite bead or saturable reactor.

It is catch 22 . ForZVS you need leakage in the transformer and leakage cause the ringing. You may use extra inductor series with primary winding for proper ZVS. Then you could make lower leakage winding design. But this will bring extra cost . The best way to fight this to utilise adaptive dead time control for zvs operation where dead time is function of the load .TI has ıntroduced new generation of controller for this purpose UC28950 as I remember

To achieve ZVS We should have some leakage inductance so for generating leakage inductance in-spite of adding one inductor in series we should have to wound transformer side by side ( not disperse) .with the use of this we can get lakage inductance and we can get ZVS.

I tried with shim inductance of 10UH with EI core , my ringing is came down , but the inductor is heating and getting saturated, once it get saturated again ringing will starts , what kind of inductor will support for this type of application????

Leakage is hard to kill into zero, is a parasitic very hard to avoid, ZVS make use of it, that's the beauty innovated by genius, that's why I said get it lowest to the best not eliminating it, achieving ZVS shall not be an issue for hand wound transformer, I believe you have more than enough if your transformer structure is not so good in design.

The added choke get heated it is the fact ringing energy is dump actually, the more you "block" the ringing the more energy you dumped, 10uH is too big and EI is not saturable type we are talking, suggest to drop by to Toshiba website to look for suitable ferrite bead.

Anyhow, aim to get zero ringing by fact it would be over-design and may not optimized in terms of cost, efficiency, meeting the EMI, output noise and component stress, I would think that the ringing just get it enough lower so that the output diode not over stress, output noise is meeting the spec. and there is no contribution of EMI that cause "Fail", with such compromise you could get lower in power losses.

you have to use powdered core or vitrovac types which are specially used for magnetic amplifier .your core will be saturates for that you have to decide some blocking time and this time is depends on H, turns N,core area & applied voltage.so normally i prefer w-464core with 3 turns to suppress ringing. Don't use EI core its Bsat = 5000gause.
so it will saturate soon.

The following questions arise:
1) Under what load conditions was the measurement made?
2) How was the primary current measured?
3) Is this problem noticed with no load?
4) What if the primary of the transformer is disconnected?
5) How is the DC bus voltage under these conditions?

Yes, waveforms will also be definitely useful.

Hi Rajesh,

I'm trying to understand your waveform in details; the waveform looks like is Secondary rectifier voltage but as per your turns ratio is should be 75V (400/16*3) instead of 150V, and it is not Primary main mosfet switching waveform, this make me confuse.

And, the current looks like it is not the main mosfet or output rectifier current, likely, it is both branch current from both ON cycles from both rectifier, in which is refer at the nod after both rectifier diode and before the output choke. OR, it is the critical point of minimum load ?

As per my understanding this ringing is not the culprit of not achieving ZVS. and roughly estimate your current ringing frequency is lower than 112nS, could be as small as ~50nS, which is somewhere most of the none SBD diode trr is, in such a 400V (I guess what you use now) is hard to find SBD, may can try SiC diode to check will this ringing gone !

About not able to achieve ZVS over the whole loading range, an adaptive delay time may help you, as recommended by Kadir.

Saturable reactors on the secondary are a non-starter. They add cost and will get extremely hot. I've seen them glowing! Use a shim inductor with a MPP core to add leakage on the primary. It'll still burn some power but not as bad as the SR's and certainly not as bad as attempting to add it to the transformer itself. Make sure you add FAST, 600V clamping diodes from the AC node between the shim and the primary to the 400V bus and ground. Your efficiency and thermals should dramatically improve.

I am assuming that the ringing is due to reverse recovery of the secondary diode rectifiers, while turning off. In that case, the spike seen in the primary current is the reverse recovery current reflected to the primary. This can be confirmed by observing the said spike on the secondary side as well. Further, the ringing frequency and current spike should decrease if the primary (or secondary) lead is passed through a small toroid.
If all this is true, then the ringing frequency can be reduced by connecting an R-C snubber ( say, 10 ohms – 2.2 nf ?) across the secondary diode. Here, the snubber resistor power = ½*Csnb*V^2*f. Ideally, the snubber resistor requires knowledge of the leakage inductance value as well.

you could also increase the gate drive resistor to reduce the current spike. Again, I am suggesting this only to gain insight into this problem.