Better sounding anti-imaging filter
Still in the vein of AIF designs which utilize off-the-shelf inductors, I'm currently listening to this one and its thoroughly enjoyable. I built it with some very cheap bobbin-wound (unscreened) 10mH chokes I found on Taobao.
I don't recommend using unscreened inductors for this but if that's all you have the results are still aural candy. They do pick up a little low level mains hum, not noticeable except when close to the speaker. Also they interact - when I was checking the prototype filter for continuity I hooked up the LCR meter to the 4 inductors in series and found different answers - none measured 40mH. The result was from 32mH to 45mH (from memory - my filters are balanced so I have four sets of series Ls). I have them spaced apart by almost one coil diameter (about 6mm) so if you're going to build this I suggest greater spacing.
You'll note that this one's industrial strength in terms of its stop-band rejection - around 60dB. Whether this is obtained in practice depends on factors like the component tolerances and the Q of your inductors. Mine are fairly lossy, showing about 25ohms @ 20kHz although only 14ohms at DC. You might also note I've not made the stop band flat - this is an experiment because the DAC's output already has a sinc response meaning a zero at 44k1 so I adjusted the stopband to suit that.
While on the subject of tolerances - the chokes I bought were surprisingly good in terms of tolerance - most were better than 1% right out of the bag. Be aware though that when measuring inductors the stimulus current needs to be fairly low - my LCR meter allows me to choose the stimulus voltage and this I set to 100mV when checking highish value inductors.
I don't recommend using unscreened inductors for this but if that's all you have the results are still aural candy. They do pick up a little low level mains hum, not noticeable except when close to the speaker. Also they interact - when I was checking the prototype filter for continuity I hooked up the LCR meter to the 4 inductors in series and found different answers - none measured 40mH. The result was from 32mH to 45mH (from memory - my filters are balanced so I have four sets of series Ls). I have them spaced apart by almost one coil diameter (about 6mm) so if you're going to build this I suggest greater spacing.
You'll note that this one's industrial strength in terms of its stop-band rejection - around 60dB. Whether this is obtained in practice depends on factors like the component tolerances and the Q of your inductors. Mine are fairly lossy, showing about 25ohms @ 20kHz although only 14ohms at DC. You might also note I've not made the stop band flat - this is an experiment because the DAC's output already has a sinc response meaning a zero at 44k1 so I adjusted the stopband to suit that.
While on the subject of tolerances - the chokes I bought were surprisingly good in terms of tolerance - most were better than 1% right out of the bag. Be aware though that when measuring inductors the stimulus current needs to be fairly low - my LCR meter allows me to choose the stimulus voltage and this I set to 100mV when checking highish value inductors.
Total Comments 0
