Is there good notching filter that can effectively remove the powerline harmonics? This is a real tough project for at least two reasons: 1. Frequency of power line signals flucturates. It is not a constant 50/60Hz 2. Harmonics can go 10th or beyond. Thanks for the advice in advance. James www.go-ci.com
notching filter that removes the power line harmonics?
Started by ●October 23, 2008
Reply by ●October 23, 20082008-10-23
James,
In your approach a comb filter would be more effective then
just a notching filter. I think essentially you mean that.
Search and understand comb filter design and i think your problem
will be solved. if you are not keen on phase/group delay distortions
then IIR comb might be more effective then FIR comb.
Regards
Bharat Pathak
Arithos Designs
www.Arithos.com
Dsp Design Consultancy and Training Company.
>Is there good notching filter that can effectively remove the
>powerline harmonics? This is a real tough project for at least two
>reasons:
>
>1. Frequency of power line signals flucturates. It is not a constant
>50/60Hz
>2. Harmonics can go 10th or beyond.
>
>Thanks for the advice in advance.
>
>James
>www.go-ci.com
>
Reply by ●October 23, 20082008-10-23
On Thu, 23 Oct 2008 00:20:35 -0500, bharat pathak wrote: (top posting fixed)> > >>Is there good notching filter that can effectively remove the powerline >>harmonics? This is a real tough project for at least two reasons: >> >>1. Frequency of power line signals flucturates. It is not a constant >>50/60Hz >>2. Harmonics can go 10th or beyond. >> >>Thanks for the advice in advance. >> >>James >>www.go-ci.com >> > James, > > In your approach a comb filter would be more effective then just a > notching filter. I think essentially you mean that. > > Search and understand comb filter design and i think your problem > will be solved. if you are not keen on phase/group delay > distortions then IIR comb might be more effective then FIR comb. > > Regards > Bharat Pathak > > Arithos Designs > www.Arithos.com > > Dsp Design Consultancy and Training Company. >Of course if the deviation from the nominal frequency is severe, the comb filter won't work well at all on the higher harmonics. _If_ there's a need to preserve the signal much above the line rate and _if_ you have a sample of the line frequency available, an cheap adaptive filter that acquires the line frequency and uses that to set the spacing on a comb filter would work well. If you _don't_ have the line frequency available, locking onto whatever is close to the expected line frequency and is stable may work, as may a full-blown LMS-error adaptive filter. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com Do you need to implement control loops in software? "Applied Control Theory for Embedded Systems" gives you just what it says. See details at http://www.wescottdesign.com/actfes/actfes.html
Reply by ●October 23, 20082008-10-23
Tim Wescott wrote:> On Thu, 23 Oct 2008 00:20:35 -0500, bharat pathak wrote: > (top posting fixed) >> >>> Is there good notching filter that can effectively remove the powerline >>> harmonics? This is a real tough project for at least two reasons: >>> >>> 1. Frequency of power line signals flucturates. It is not a constant >>> 50/60Hz >>> 2. Harmonics can go 10th or beyond. >>> >>> Thanks for the advice in advance. >>> >>> James >>> www.go-ci.com >>> >> James, >> >> In your approach a comb filter would be more effective then just a >> notching filter. I think essentially you mean that. >> >> Search and understand comb filter design and i think your problem >> will be solved. if you are not keen on phase/group delay >> distortions then IIR comb might be more effective then FIR comb. >> >> Regards >> Bharat Pathak >> >> Arithos Designs >> www.Arithos.com >> >> Dsp Design Consultancy and Training Company. >> > > Of course if the deviation from the nominal frequency is severe, the comb > filter won't work well at all on the higher harmonics. > > _If_ there's a need to preserve the signal much above the line rate and > _if_ you have a sample of the line frequency available, an cheap adaptive > filter that acquires the line frequency and uses that to set the spacing > on a comb filter would work well. > > If you _don't_ have the line frequency available, locking onto whatever > is close to the expected line frequency and is stable may work, as may a > full-blown LMS-error adaptive filter.It's almost always more satisfactory to exclude line noise from the signal than to try to remove it later. Using differential-input buffers with high CMRR between the signal source and the sampler, the buffer's output being referenced to the sampler's analog ground* often yields a 20 dB improvement right away, and can usually be tweaked to 30. A bullet-proof vest is a better first choice than the best-equipped emergency room. Jerry _____________________________________ * In other words, a properly connected instrumentation amplifier. Engineering is the art of making what you want from things you can get. ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
Reply by ●October 23, 20082008-10-23
DigitalSignal wrote:> Is there good notching filter that can effectively remove the > powerline harmonics?There is no problem to build a good notch/comb filter. However it helps only to the certain extent. The removal of the powerline interferrence is more complex then just notching out the harmonics.> This is a real tough project for at least two > reasons:> 1. Frequency of power line signals flucturates. It is not a constant > 50/60HzUnless you are running from some small individual alternator, the power frequency is very stable. The waveform is what fluctuating.> 2. Harmonics can go 10th or beyond.No. Harmonics are clearly seen to 500th and beyond.> Thanks for the advice in advance.What are you trying to accomplish? Vladimir Vassilevsky DSP and Mixed Signal Design Consultant http://www.abvolt.com
Reply by ●October 26, 20082008-10-26
Jerry, I agree. However this is a case that the powerline signal has been mixed in the measurement. In fact it is part of power line signal measurement. The interested signal content is somewhere other than the 50/60Hz harmonics. To follow up with this question, I wonder if there is a way to adaptively notch out the power line harmonics in the analog domain before it enters the A/D. Have you seen such a solution? James www.go-ci.com
Reply by ●October 26, 20082008-10-26
DigitalSignal wrote:> Jerry, I agree. However this is a case that the powerline signal has > been mixed in the measurement. In fact it is part of power line signal > measurement. The interested signal content is somewhere other than the > 50/60Hz harmonics. > > To follow up with this question, I wonder if there is a way to > adaptively notch out the power line harmonics in the analog domain > before it enters the A/D. Have you seen such a solution? > > James > www.go-ci.com1. What are you measuring? 2. Why are you measuring it? 3. What is the source of noise? 4. Is the noise stationary?
Reply by ●October 26, 20082008-10-26
DigitalSignal wrote:> Jerry, I agree. However this is a case that the powerline signal has > been mixed in the measurement. In fact it is part of power line signal > measurement. The interested signal content is somewhere other than the > 50/60Hz harmonics. > > To follow up with this question, I wonder if there is a way to > adaptively notch out the power line harmonics in the analog domain > before it enters the A/D. Have you seen such a solution?I think digital is the way to go for anything adaptive. I can imagine a system with servo-driven variable capacitors and inductors, but I wouldn't want to build (or maintain) one. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●October 26, 20082008-10-26
DigitalSignal wrote:> Jerry, I agree. However this is a case that the powerline signal has > been mixed in the measurement. In fact it is part of power line signal > measurement. The interested signal content is somewhere other than the > 50/60Hz harmonics. > > To follow up with this question, I wonder if there is a way to > adaptively notch out the power line harmonics in the analog domain > before it enters the A/D. Have you seen such a solution? >Like Jerry, I think digital is the way to go here. When the necessary operations get complex, it is much easier to maintain accuracy in a small form factor by doing things digitally. You _could_ make an adaptive comb filter using a delay line, but these are getting harder and harder to find. I think you're much better off just beefing up your ADC to capture more of the signal + harmonics. If part of your problem is huge, narrow spikes, consider making your analog front-end so that it doesn't get upset by them for any long duration, then sense the spikes and blank them (by not paying attention to those ADC readings, and interpolating over them) digitally as part of your noise rejection. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com Do you need to implement control loops in software? "Applied Control Theory for Embedded Systems" gives you just what it says. See details at http://www.wescottdesign.com/actfes/actfes.html
Reply by ●October 26, 20082008-10-26
DigitalSignal wrote:> To follow up with this question, I wonder if there is a way to > adaptively notch out the power line harmonics in the analog domain > before it enters the A/D. Have you seen such a solution?IIRC Maxim has ADCs with the built in rejection of 50 and/or 60 Hz. Don't know whether they do the filtering before or after conversion. BTW, making a comb filter in the analog domain is not a very big deal either. Vladimir Vassilevsky DSP and Mixed Signal Design Consultant http://www.abvolt.com
