hi, I've now designed an active RC filter with working frequency from 20K to 300K and I also designed the amplifier my self. Now I want to test the performance of the filter. How can I test the THD of the filter? For I use a real amplifier that I designed instead of a model, can I test the THD of the filter if a add a sin wave with frequency in the passband? If yes, how can I test the THD, will anyone give me a instruction? Thanks !
What is a THD ? The harmonics squared and summed or similar. There are now two ways, either measure the output and do the calculations, or measure the input and the filter and do the calculations this way.
Rene
-- Ing.Buero R.Tschaggelar - http://www.ibrtses.com & commercial newsgroups - http://www.talkto.net
I designed the opamp myself, so the active component in the filter isn't perfectly linear. On the other hand, I also don't know how to test the THD of the opamp. Looking forward to reply! thanks Rene Tschaggelar =E5=86=99=E9=81=93=EF=BC=9A
I'm assuming that you do not have access to a spectrum analyzer. One way to measure distortion is to null out the output of your amplifier with an inverted version of the input. You need to be able to shift the phase of the inverted input to compensate for phase shift in the amplifier. The difference between the amplifier output and the inverted, phase shifted input signal is THD + noise. Regards, Jon
thanks, Jon I'll try to do this tomorrow morning!
I' ve tried to do dft to the output of my filter,and I' ve found that if I input a 0.1v sinewave of 5KHz, I can get a spectrum of 99.8mv at
5KHz in the dft plot and the other harmonic waves are no more than 100nv. But when I calculated the THD of the filter, I found that it is only -63dB. I wonder how many harmonic waves has been calculated into THD? Why the THD calculation result is so small? Looking forward to reply. Thanks!By the way, I use Cadence Spectre to facilitate my design. When I calculate THD, I need to input the "range" and "sample number". If my input is 5KHz, how do I determine these two values? Jon =E5=86=99=E9=81=93=EF=BC=9A
-- Notch out the fundamental in the output and whatever\'s left will be distortion products.
Seeing as Google twats lost this lunchtime's treatise, this one is shorter. You need to be looking down beyond the 0.01% area. Good spectrum analysers only get you down to 0.1%. Audio band THD is straightforward. Sig gens can be pure and it is easy to knock up a precise notch filter, which is the usual method.
20k and up is bad news. Both for sig gen and active notch POV. Go for spot measurements. Use passive LC bandpass to clean up generator output and passive LC notch on filter output. Suggest 3rd order Butterworths with say 470ohm source/load resistance. 3rd order allows casual coil winds hence lower coil Qs, yet gives THD down to better than -100dB (.001% 10ppm) Bad news again ... 1] 3 spot measurements would need total of 18 inductors. 2] Need 3mH inductors at the bottom end. Can't use cores as they add THD. 3mH of air core needs hundreds of turns.Monumental task. Suggest just telling people your filter is "low distortion" but noting your second post I'm now wondering if you grasp the scale of the problem. john
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.