Mains hum in SMPS

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Disabled Account
Joined 2010
Does mains hum propagate in switchmode supplies?
They're typically using low value caps on the primary side, due to size and cost. So I'm thinking that the noise might be able to get to the secondary side.
Are there any special requirements for audio?


Primary caps are big in reality. A 100uF primary cap charged to 300V= stores the same energy as a secondary cap with 10.000uF charged to 30V. So primary hum ripple is similar to a conventional transformer supply. Additionnally output voltage of smps is regulated so there is very little hum at the DC output terminals.
 
But then regulation adds noise in the audio range...
Primary caps are big in reality. A 100uF primary cap charged to 300V= stores the same energy as a secondary cap with 10.000uF charged to 30V. So primary hum ripple is similar to a conventional transformer supply. Additionnally output voltage of smps is regulated so there is very little hum at the DC output terminals.
 
I don't agree. A well done regulated supply will be orders of magnitude quitter than a simple CLC supply. For most switchers the noise is measured in millivolts and generally at a frequency that doesn't effect the audio circuits.

The 60 / 120 hz of the AC supply is blocked by the way a switch mode supply works. The DC on the input side of the switcher is used to power a high frequency power switching circuit the drives a transformer "tuned" for the frequency of the switching circuit. That frequency can be 30 kHz to 100 kHz. The 60 / 120 Hz can't cross the transformer. The high frequency is much easier to filter requiring smaller inductors and caps to get a very clean DC if done correctly. Much less cost and better performance.
 
The 60 / 120 hz of the AC supply is blocked by the way a switch mode supply works. The DC on the input side of the switcher is used to power a high frequency power switching circuit the drives a transformer "tuned" for the frequency of the switching circuit. That frequency can be 30 kHz to 100 kHz. The 60 / 120 Hz can't cross the transformer. The high frequency is much easier to filter requiring smaller inductors and caps to get a very clean DC if done correctly. Much less cost and better performance.
I was thinking about that as well, but I wasn't sure if the low frequencies would pass through the transformer.
What about the amount of capacitance in the secondary? For such high frequencies, less capacitance is needed, but then the purpose of big caps is to have enough energy storage for transients.
 
Disabled Account
Joined 2010
I was thinking about that as well, but I wasn't sure if the low frequencies would pass through the transformer.
What about the amount of capacitance in the secondary? For such high frequencies, less capacitance is needed, but then the purpose of big caps is to have enough energy storage for transients.


The conventional supply do have the bulk storage caps behind the rectifier on the secondary side.
The smps have the bulk storage caps behind the rectifier on the PRIMARY side.


That makes the difference.
 

PRR

Member
Joined 2003
Paid Member
...bulk storage caps behind the rectifier on the PRIMARY side....

A side-point I had not appreciated before:

Energy Storage (what we really want) goes with the *square* of voltage.

A higher voltage cap tends to store more energy per pint or per buck.

Doing the energy storage at 320V (line-side) may be far better than on the 5V or 12V side.

The 100/120Hz ripple really should NOT "come through", because the switcher typically also Regulates. Instead we have 40KHz ripple, but this may be easier to filter or ignore.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.