Undamped LC Filter - Electronic Circuits Projects Diagrams Free

Prof_Plum · 24th May 2006, 01:34 PM

On the output of a voltage regulator I have for example 75KHz noise. If I want to put an LC filter arrangement on the output of the regulator, do I specify the 3dB point to be much lower than the 75KHz noise. If so how much lower?

**Nigel Goodwin** · 24th May 2006, 01:48 PM

Quote:

Originally Posted by Prof_Plum

On the output of a voltage regulator I have for example 75KHz noise. If I want to put an LC filter arrangement on the output of the regulator, do I specify the 3dB point to be much lower than the 75KHz noise. If so how much lower?

What sort of regulator is it?, if it's a linear one you need to build it properly as it's unstable. For that matter, if it's switch-mode it needs designing properly as well, what level is the 75KHz, and what's the output voltage?

Prof_Plum · 24th May 2006, 01:55 PM

Its a switching regulator 17.5V to 3.3V, its stable but the switching noise of the regulator is inherent on the output. Also I use linear reg's elsewhere and depending on the circuit load these can have switching noises depending on the switching frequency of the load.

So in general where is it better to design the filter 3db frequency response for an LC filter.

The output noise is approx. 40mV peak-peak, but I'm using it in a sensitive RF circuit.

Hero999 · 24th May 2006, 03:59 PM

Connect a 100nF, 10nF and 1nF capacitors on the output and add a small inductor 10uH in seris with the load.

Roff · 24th May 2006, 04:45 PM

Quote:

Originally Posted by Hero999

Connect a 100nF, 10nF and 1nF capacitors on the output and add a small inductor 10uH in seris with the load.

If you mean like in the schematic below, this illustrates the perils of using an undamped LC filter with a switching or high impedance load.

Hero999 · 24th May 2006, 05:01 PM

No, I'm not talking about switching high impedance loads.

I'm talking about adding an LC filter to the output of the switching regulator to remove the 75kHz ripple.

If you want to be exact a 470uH choke in series with the 10nF capacitor connected in paralell with the load might help (Fo would be 73.kKhz) but I don't think you need a tuned filter.

Roff · 24th May 2006, 05:14 PM

Quote:

Originally Posted by Hero999

No, I'm not talking about switching high impedance loads.

I'm talking about adding an LC filter to the output of the switching regulator to remove the 75kHz ripple.

If you want to be exact a 470uH choke in series with the 10nF capacitor connected in paralell with the load might help (Fo would be 73.kKhz) but I don't think you need a tuned filter.

My point is, an LC filter on a power supply can wreak havoc if your load excites it at the resonant frequency.

Hero999 · 24th May 2006, 05:44 PM

I see your point you need to watch out for that.

However it wouldn't be in issue for an FM transmitter, or a digital circuit so long as it doesn't switch at 151kHz or an odd harmonicfor exmple a longwave transmitter running at 151KHz or another switching PSU would cause a disaster.

**Nigel Goodwin** · 24th May 2006, 05:57 PM

Quote:

Originally Posted by Prof_Plum

Its a switching regulator 17.5V to 3.3V, its stable but the switching noise of the regulator is inherent on the output. Also I use linear reg's elsewhere and depending on the circuit load these can have switching noises depending on the switching frequency of the load.

So in general where is it better to design the filter 3db frequency response for an LC filter.

The output noise is approx. 40mV peak-peak, but I'm using it in a sensitive RF circuit.

40mV is a very low ripple level, assuming it's feeding a sensitive RF circuit then that circuit itself should have supply decoupling to remove any noise from the supplies.

Are you sure it's been picked up down the supply rails?, if the circuit is sensitive I would tend to suspect it's more likely to be picking up radiation from the switching supply - probably a lot more intrusive than 40mV on the supply?.

Roff · 24th May 2006, 06:37 PM

Quote:

Originally Posted by Hero999

I see your point you need to watch out for that.

However it wouldn't be in issue for an FM transmitter, or a digital circuit so long as it doesn't switch at 151kHz or an odd harmonicfor exmple a longwave transmitter running at 151KHz or another switching PSU would cause a disaster.

But frequency of the load transients is relatively unimportant. Fast rising, high current edges of a low frequency switching load will excite the filter and cause it to ring. If you have a simulator, try some different examples.

Hero999 · 24th May 2006, 10:11 PM

That only happens at frequencies lower than or near the resonant frequency - you won't have this problem with an FM transmitter.

Anyway, I was wrong when I mentioned an inductor, you should use a ferrite bead which is a kind of inductor but it has a very low Q so it'll be damped, or you could go for a tuned filter.

Roff · 25th May 2006, 01:00 AM

Quote:

Originally Posted by Hero999

That only happens at frequencies lower than or near the resonant frequency - you won't have this problem with an FM transmitter.

Anyway, I was wrong when I mentioned an inductor, you should use a ferrite bead which is a kind of inductor but it has a very low Q so it'll be damped, or you could go for a tuned filter.

I agree that you won't have a problem with an FM transmitter. I was addressing the general problems that can arise if you use an LC power supply filter without considering the load characteristics.

dknguyen · 25th May 2006, 01:26 AM

Will an LC filter wreak havoc if it is on the input side of the switching voltage regulator?

Thread Tools
Show Printable Version Email this Page
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

24th May 2006, 01:34 PM	(permalink)
Prof_Plum	Undamped LC Filter On the output of a voltage regulator I have for example 75KHz noise. If I want to put an LC filter arrangement on the output of the regulator, do I specify the 3dB point to be much lower than the 75KHz noise. If so how much lower?

24th May 2006, 01:55 PM	(permalink)
Prof_Plum	Its a switching regulator 17.5V to 3.3V, its stable but the switching noise of the regulator is inherent on the output. Also I use linear reg's elsewhere and depending on the circuit load these can have switching noises depending on the switching frequency of the load. So in general where is it better to design the filter 3db frequency response for an LC filter. The output noise is approx. 40mV peak-peak, but I'm using it in a sensitive RF circuit.

24th May 2006, 03:59 PM	(permalink)
Hero999	Connect a 100nF, 10nF and 1nF capacitors on the output and add a small inductor 10uH in seris with the load. __________________ I also post at the following sites: http://www.stop-microsoft.org http://www.heated-debates.com Screen name: Aloone_Jonez

24th May 2006, 05:01 PM	(permalink)
Hero999	No, I'm not talking about switching high impedance loads. I'm talking about adding an LC filter to the output of the switching regulator to remove the 75kHz ripple. If you want to be exact a 470uH choke in series with the 10nF capacitor connected in paralell with the load might help (Fo would be 73.kKhz) but I don't think you need a tuned filter. __________________ I also post at the following sites: http://www.stop-microsoft.org http://www.heated-debates.com Screen name: Aloone_Jonez

24th May 2006, 05:44 PM	(permalink)
Hero999	I see your point you need to watch out for that. However it wouldn't be in issue for an FM transmitter, or a digital circuit so long as it doesn't switch at 151kHz or an odd harmonicfor exmple a longwave transmitter running at 151KHz or another switching PSU would cause a disaster. __________________ I also post at the following sites: http://www.stop-microsoft.org http://www.heated-debates.com Screen name: Aloone_Jonez

25th May 2006, 01:26 AM	(permalink)
dknguyen	LC Will an LC filter wreak havoc if it is on the input side of the switching voltage regulator?