Power Supply's routing and design

[PovilasKubilius]

Hi,
I am young engineer and have the task to make 200w 36V Half bridge Led lighting SMPS, that can make true EMC Europe standards. Does anyone has favorite ST NXP or other manufacturer chips and can help my with this task. Routing 200w SMPS is tricky, I have experience with 30w Supplies and it was 2 mounth of work to get everything right with no additional cost of product. Any help will be highly appreciated.
PS. I have one week to make prototype.
Thanks Robert Feranec for teaching me Altium Designer thrue Youtube and making me better engineer.

[robertferanec]

Go with a reference board. I am not expert for power supplies, however in your situation I would definitelly try to find an existing reference design of something similar and would adjust it.

Maybe something like this can help you to start (130 W dimmable constant current LED driver using ICL5102 in PFC and LLC topology - I know it is not 200W, but still can be a good start):

https://www.infineon.com/dgdl/Infineon-EngineeringReport_130W_dimmable_constant_current_LED_driver_ICL5102-AN-v02_00-EN.pdf?fileId=5546d4626102d35a01612de7a06b6fb3

[PovilasKubilius]

Funny, I'm doing just that the same chip the same board. Already talking to Infineon about better documentation and development tools. Genius minds think alike. Thank for help!!!

[Steve.Picotest]

You don't provide much in the way of specifications, but you might see my relatively old article on single stage PFC flyback DC-DC converter. You can find it here https://www.powerelectronics.com/tec...tor-correction. One advantage of this design topology is that it is low cost and another is that the gate is ground referenced - no high side driver.

Current FED push-pull design is also very good for this application, as it is current fed, robust and with excellent utilization of semiconductors. This has three MOSFET's, 2 magnetics and a high side drive, so much more expensive, but very nice topology.

[chitransh92]

Steve.Picotest .
The moment I saw Steve "Picotest", dictionary in my brain threw out the name "Steven M. Sandler".
Nice to have an esteemed expert in the club.

[Steve.Picotest]

Thanks for the kind words

[PovilasKubilius]

Hello again, I made the prototype of ICl5102, but can't make the PFC work correctly. I tried cut grounds and reroute with wires, tried filter signals with pF capacitors nothing helps. My PFC just loses control periodically. Interesting that the PFC gets better at higher loads. In those photos you can see 80w load and PFC behavior. I measured it before PFC diode.

[Steve.Picotest]

Looks like the PFC is working, but is either oscillating or needs an input filter.

[PovilasKubilius]

It's probably oscillating, because i have double EMC filter and inductor near input capacitor of PFC. How can I prevent oscillations?
Thanks for response Steve, it's very valuable.

[Steve.Picotest]

You would need to look at the control loop gain either in measurement or in simulation. It would also be good to look at the input current with triggering on the high frequency or looking at the current on a spectrum analyzer to see the frequency it's oscillating at.

I don't know your requirement t, but I do have a few older articles about isolated and non-isolated PFC.

[PovilasKubilius]

I'am not able to measure current at this time, but I did Lisn spectrum analysis. Can it help me to tackle this issue ? It seems that the problems appear at 30 Mhz.

[Steve.Picotest]

You may be right. See if you can get a screenshot of the input current at a faster sweep so we can see the high frequency waveform

[PovilasKubilius]

I don't have current probe. I use chip: ILC5102. Can all of this problems be caused because of not optimized Leakage transformer and resonant capacitor?

[Steve.Picotest]

That's possible, or layout. Without data it's hard to tell. Maybe you can borrow a current probe from a scope company.

[PovilasKubilius]

Hi,
I have another interesting question.
Why is there 285 VAC potential between PE and GND after SMPS (fly-back).
Does it not provide full independent power from mains?
This question came from our testing team who touched the enclosure and power supply gnd and got slightly shocked. I answer it that is because of TN-CS wiring system, so what is happening there truly?