Author Topic: Cleaning the output of a cheap SMPS?  (Read 1780 times)

0 Members and 1 Guest are viewing this topic.

Offline Pack34

  • Frequent Contributor
  • **
  • Posts: 653
Cleaning the output of a cheap SMPS?
« on: November 07, 2015, 04:24:33 am »
These SMPS power supplies are all over ebay and picked one up for some projects. However, the output of the unit is rather noisy and it seems to permeate through the input linear regulators and can be seen super-imposed on some analog signal lines that I'm sampling.

Power supply in question:
http://www.ebay.com/itm/Mean-Well-RT-65C-AC-to-DC-Power-Supply-Triple-Output-Enclosed-Switching-65-5W-/111702973357?hash=item1a02041bad:g:WA4AAOSw-vlViwGS

The Noise:
See the "scope_4.png" attachment to see the noise on the power rail AFTER the input linear regulator.

A clean input:
See the "scope_5.png" attachment as a reference for the rail using a benchtop power supply.

Now, it's not something that's absolutely terrible but it's something that I want to learn how to address.

My first plan-of-attack was to add an LC filter between the SMPS and the electronics. It can be seen in the first reply in this thread: http://www.eevblog.com/forum/beginners/lc-filter-basics-for-smps/

Unfortunately, it didn't seem to do anything. The most likely culprit could be the values I used (3.9uH inductors and a 470uF cap). However, the cut-off frequency is 1/(2*pi*sqrt(LC)). This should give me a cut-off of about 3.7kHz. If I'm reading it right, the datasheet link below shows a switching frequency of 60 kHz. It's marked as fosc next to the block diagram.

http://www.meanwell.com/webapp/product/search.aspx?prod=rt-65

Taking some measurements on my scope. The most noticeable peaks are roughly at:
244kHz
181kHz
118kHz
116kHz
6.82kHz

Then a really small peak at around the 120Hz mark (8.3ms period) as seen on the scope screenshot.

Considering that all of these are substantially higher (sans the 120Hz) than my input filter I hacked in ( 3.7kHz ), do I just need an inductor with a greater rating? How should I move forward to try to clean up the switching output to mitigate those pulses every 8.3ms?


« Last Edit: November 07, 2015, 05:16:49 am by Pack34 »
 

Offline miceuz

  • Frequent Contributor
  • **
  • Posts: 371
  • Country: lt
    • chirp - a soil moisture meter / plant watering alarm
Re: Cleaning the output of a cheap SMPS?
« Reply #1 on: November 07, 2015, 05:07:05 am »
both of your attachments are 0 bytes

Offline Pack34

  • Frequent Contributor
  • **
  • Posts: 653
Re: Cleaning the output of a cheap SMPS?
« Reply #2 on: November 07, 2015, 05:17:38 am »
both of your attachments are 0 bytes

Fixed. Thank you.

EDIT:// Now that's interesting. It appears to be ground noise. I can have the probe attached to ground and probe the ground pin on the input linear regulator and it has the same noise as when I probe the power rails.
« Last Edit: November 07, 2015, 08:45:09 am by Pack34 »
 

Offline Pack34

  • Frequent Contributor
  • **
  • Posts: 653
Re: Cleaning the output of a cheap SMPS?
« Reply #3 on: November 10, 2015, 04:22:13 am »
I think I have it all cleaned up.

1. Added a small resistor and inductor ferrite bead in series with the incoming power and ground lines.
2. After the LDO I hacked in an RC low-pass filter targeting 30 Hz.

Considering we're only talking a maximum of 200mA on the voltage rail, I think the RC filter will work just fine. an LC filter would probably be better but the necessary inductor would be incredibly large and expensive.
« Last Edit: November 10, 2015, 06:54:35 am by Pack34 »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf