PSU: Noise decoupling and bypassing (Hobby project)

[sowlaki]

Hello, I am currently doing a project which consists of an 3v3 and 5V output PSU inside an 3 plug power cord supplied by 240 VAC outlet. The 3 plugs are connected in series with a coil relay each and the relays are driven by a rf bluetooth module (AMS001/S).

When the RF module is driven by the PSU (3v3) inside the extension cord it acts extremely random and sends out arbitrary data. I believe this is because the 3v3 power line is extremely noisy due to the 240VAC powerline right next to it. Other rf modules work perfectly when connected to another PSU isolated from the 240 VAC powerline.

I recently read this article (https://www.eetimes.com/document.asp?doc_id=1272268&page_number=2 )

Should decouple the 3v3 line with a parallel elyt.capacitor and a inductor with a specific value to reduce this powerline noise?

The 3v3 voltage is driven by a TLV1117-3v3 IC which works perfectly in other projects.

 

[Harald Kapp]

The methods discussed in that article are more related to noise that is created by the time-varying load current through the finite output impedance of the voltage regulator.
This is very likely not an issue here as you note that the same regulator works fine in other projects. Nevertheless you can try adding output capacitance. For details about capacitor selection lookup section 9.1 in the regulator's datasheet.
You can also try the inductor, but I suggest you use a ferrite. A ferrite reduces the risk of resonance as it incurs way more loss than an inductor.

The issue you see is more likely to come from capacitive coupling between the low voltage side and the mains side. Usual remedies are:

1. Put more distance between primary side and secondary side - obviously not suited to your space limited arrangement.
2. Put a shield between primary and secondary side. Some copper foil connectd to the earth terminal of your mains wire should work. Take care to ensure proper insulation between the shield and any other voltages, especially mains.

 

[sowlaki]

Thanks for the two tips, the PSU happens to be reversed engineered from some samsung phone adapter because I needed it to be different dimensions. I guess am not very good at PCB layout to reduce noise as much as possible. I will try to make some custom EMI shield like you said to protect from primary side. Might add more bypass ceramic capacitors next to the RF module's Vcc to minimize the increased negative effect big capacitors could give, Inductance etc.

 

[Harald Kapp]

I guess am not very good at PCB layout to reduce noise as much as possible.

Good layout is indispensable for a switched mode power supply.

I sincerely hope you have observed all required safety measures such as minimum creepage distances and air gaps as required for the safety of that psu? Typically you'll have to have > 6 mm between primary side and secondary side for a mains powered unit and > 3 mm within the primary side. Exact values depending on voltage levels and applicable standards.
If there is any chance that metallic parts that can be touched (e.g. case, connectors etc.) could come in contact with mains power, a proper earth connection is required for safety reasons. As you used a phone adapter as base for your design, there most likely wasn't a protective earth connection in the first place. If your mechanical design allows touching conductive parts (as described above) you need to consider adding protective earth unless you can ensure proper safety by other means of insulation.

 

[sowlaki]

I actually read about those safety distances before constructing the PCB so I think that wont be a problem and the current drain is below 500 mA . The case is hard plastic and the 3 plugs are grounded, I am from Sweden, Europe so they are 240 VAC with 2 ground plugs next to the VAC connectors. Still haven't tested the advise you gave me but since it is weekend I will give it a go.The only cords sticking out of the case are 3 relay (5V) which I will pull to ground to turn them on and 2 other cords 3v3 and GND.

 

[Harald Kapp]

 

[sowlaki]

[SOLVED] I found out that the problem I had did not have to do anything with a noisy PSU, in fact the problem occured from any particular power source. The problem was that when I left the UART (TX, RX) pins floating on the BLE module they acted as random bit transmitters/recievers. Since I am new to the UART protocol I never thought this would be a problem. This problem occured while switching the module to remote COMMAND (see picture below) and it looks as if the UART pins are bypassed completely and shouldn't interfere with the transmitted BLE data. But I guess they're in the same multiplexer or something. However I am glad that it worked out nice and my PSU won't need any corrections.

 

[Harald Kapp]

I left the UART (TX, RX) pins floating

Never leave any input pins floating unless explicitly directed so by the ic datasheet.
This goes for any input, not only UARTs.