Filter for buck converter output

[I14R10]

Hi

I need to have 5V voltage source. I'm looking into buying either AZ 34063 (https://www.tme.eu/hr/en/details/az...egulators-dc-dc-circuits/diodes-incorporated/) or AP6503 (https://www.tme.eu/hr/en/details/ap6503sp-13/voltage-regulators-dc-dc-circuits/diodes-incorporated/).

AZ34063 recommends to have additional LC filter on the output, AP6503 doesn't.

I'm planning to have this used as 12V to 5V converter from my solar powered lead acid battery. I'll be powering Raspberry pi, RTL SDR receiver and RFM98 LoRa module. SDR is listening to 1420 MHz, and LoRa is transmitting at 433 MHz. My question is - do I need to do some special low pass filtering after buck converter output? I don't know if their switching produces noise up to, and over 433 MHz.

I'm assuming it's wise to have the buck converter shielded into metal enclosure and grounded, but that won't do anything if the noise is present on the output wire.
I even considered using buck to lower the voltage to around 7V and then using linear voltage regulator to bring it down to 5V. All options are still open.

Any advice is welcome.

 

[KlausST]

Hi,

I'd say it depends on schematic, device selection and (maybe most of all) PCB layout.
In all my very previse industrial measurement tools I use swich mode supplies (good quality, own PCB layout). No metal shield.
I never had problems with 16 bit ADC next to them, they always worked at least as good as their datasheet says.

Now I can not speak about the influence to HF devices.

But I guess a proper design and maybe an RC filter at sensitive nodes should do.
The Raspberry will have no problem. So spend the other parts (SDR and LoRa) individual RCs.
Use really solid GND plane, maybe splitted for the individual circuits.

I used so called "shielded" inductors, and parallel Cs (tantal, 1u ceramics, 10n ceramics)

But I can not give any guarantee for your design.

Klaus

 

[cupoftea]

I don't know if their switching produces noise up to, and over 433 MHz.

Yes they produce some at 433MHz, but not that much, unless you really mess up the pcb layout.
Radiated emissions testing goes up to 1GHz usually.

You can buy PCB mounted metal shields to go over your bucks if you want......and you then shield the pcb side with pcb copper planes connected to quiet nodes. (eg vcc gnd)...but at such low voltages, i doubt it would be needed.

....Your voltage is too low i think for needing shields.......now if it was an offline flyback smps say, feeding off the 240vac mains, then it would need shielding...or at least it would need it to pass en55032 radiated emissions......generally speaking no offline 240vac SMPS that is hard switched can pass en55032 radiated emissions without it being in a metal shield.

When ive done radiated emc, the strong emissions peaks go up to about 250MHz max.

Of course, one thing you can do to reduce emissions at 433MHz is damp the switching transistion by adding gate resistance to slow up the mosfet switching...but not too much as it causes fet heating

 

[D.A.(Tony)Stewart]

Can you define your SDR sensitivity to DC ripple or noise threshold vs f? Sometimes a DC Balun on output helps where stray RF ground currents interfere with unbalanced impedance. Then a load C load acts as a DM filter or "Pi" filter.