I have seen a few articles that has mentioned an ideal LPF has to filter high frequency signal to 0V as an ouput. However, IMO this is not what it is meant to be.
Let say for Arduino high frequency PWM signal is switching between 0V and 5V with 50% duty cycle, an ideal LPF should filter as a result of steady 2.5V instead of 0V.
Did I messed up the definition of a LPF?
Yes you are correct - the DC value is not affected by a low pass filter, only the higher frequencies drop to zero
for an ideal low-pass filter. Ideal low-pass filters don't exist though, all analog low pass filters reduce high frequencies, not completely eliminate. But the reduction can be by extremely large factors for some filters.
So should I interpret LPF in a DC circuit as a voltage smoother. While it acts as a high frequency filter in AC circuit?
If the circuit is truly just DC there is no need to low pass filter! Filters just knock down some frequencies, whatever they are used with. Its a low-pass filter. "High frequency filter" is ambiguous phrase, stick to the conventional name to avoid confusing people. "high-pass", "low-pass", "band-pass", "notch", "all-pass" are
the basic types.
But yes smoothing is one application for low-pass filtering.
MarkT:
If the circuit is truly just DC there is no need to low pass filter! Filters just knock down some frequencies, whatever they are used with. Its a low-pass filter. "High frequency filter" is ambiguous phrase, stick to the conventional name to avoid confusing people. "high-pass", "low-pass", "band-pass", "notch", "all-pass" are
the basic types.But yes smoothing is one application for low-pass filtering.
@MarkT, I see. Sorry for messing things up... because i was gathering information from different online resources...
Let say for Arduino high frequency PWM signal is switching between 0V and 5V with 50% duty cycle, an ideal LPF should filter as a result of steady 2.5V instead of 0V.
In the context of your question there is no difference between 0V and 2.5V, both are DC and neither, if perfectly clean, has any frequencies associated with it. That's what matters, the filter removes that which is not DC.
mannok:
@MarkT, I see. Sorry for messing things up... because i was gathering information from different online resources...
So what should I call it when I am using it for DC smoothing intention? e.g. getting steady average voltage from switch PWM power supply.
PerryBebbington:
In the context of your question there is no difference between 0V and 2.5V, both are DC and neither, if perfectly clean, has any frequencies associated with it. That's what matters, the filter removes that which is not DC.
Isn't my context associate with frequency? As it is a PWM... 5V 0V 5V 0V 5V 0V
Isn't my context associate with frequency? As it is a PWM... 5V 0V 5V 0V 5V 0V
PWM is a mixture of AC and DC. The AC part comes from it switching between 0V and 5V and back rapidly. The DC part comes from it only ever being positive, not negative. It could instead for example switch between -2.5V and +2.5V. From an AC point of view that's exactly the same, but from a DC point of view it's not as half way between -2.5V and + 2.5V is 0V, whereas half way between 0V and 5V is 2.5V. The AC part doesn't care, if the frequency is the same and the amplitude is the same then as far as AC and a filter is concerned there is no difference. The fact that one will give 0V as the output and the other will give 2.5V is down to the average DC level going in*.
If that doesn't make sense then, sorry, I can't think of a better way to explain it.
*I have assumed PWM set at 50% for this, setting it at a different ratio will give a different DC output as you know, but that's adding complication.
PerryBebbington:
PWM is a mixture of AC and DC. The AC part comes from it switching between 0V and 5V and back rapidly. The DC part comes from it only ever being positive, not negative. It could instead for example switch between -2.5V and +2.5V. From an AC point of view that's exactly the same, but from a DC point of view it's not as half way between -2.5V and + 2.5V is 0V, whereas half way between 0V and 5V is 2.5V. The AC part doesn't care, if the frequency is the same and the amplitude is the same then as far as AC and a filter is concerned there is no difference. The fact that one will give 0V as the output and the other will give 2.5V is down to the average DC level going in*.If that doesn't make sense then, sorry, I can't think of a better way to explain it.
*I have assumed PWM set at 50% for this, setting it at a different ratio will give a different DC output as you know, but that's adding complication.
Thx @PerryBebbingtong, this is so clear.
May I know is there any name for this kind of filter when it is used in this purpose? I mean turning a switched PWM power supply into its average output.
May I know is there any name for this kind of filter when it is used in this purpose?
I can only think to call it a low pass filter. If there's another name someone will be along shortly to tell me what I once knew and have forgotten 
In case you don't know this, DC is zero Hz.
May I know is there any name for this kind of filter when it is used in this purpose?
This isn't technically correct, but I usually just say "DC blocking capacitor". (You need the resistor or DC load to make a complete filter to "pull down" the DC voltage to zero.)
DVDdoug:
In case you don't know this, DC is zero Hz.
This isn't technically correct, but I usually just say "DC blocking capacitor". (You need the resistor or DC load to make a complete filter to "pull down" the DC voltage to zero.)
No no no, its a low pass filter and nothing is pulling down anything to zero!! There is no DC-blocking, its AC-blocking. Please read the question carefully...
DVDdoug:
In case you don't know this, DC is zero Hz.
This isn't technically correct, but I usually just say "DC blocking capacitor". (You need the resistor or DC load to make a complete filter to "pull down" the DC voltage to zero.)
I would think "DC blocking capacitor" in general usage is a High Pass Filter (HPF), that is the desired output is the AC portion of the signal.