This is the output filter section of the TX-NR1010/3010/5010 DAC board:
This is the recommendation from TI:
So in addition to the filtering in the I/V converters and the diff amp, they used an extra four op-amps per channel for extra active low pass filtering. Any idea why they deemed this necessary? With the noise shaping in the DACs, there should be little HF junk to filter out, and a simple passive low pass is probably better at handling HF junk than these slowish op-amps.
So in addition to the filtering in the I/V converters and the diff amp, they used an extra four op-amps per channel for extra active low pass filtering. Any idea why they deemed this necessary? With the noise shaping in the DACs, there should be little HF junk to filter out, and a simple passive low pass is probably better at handling HF junk than these slowish op-amps.
In both cases, an NE553x has to handle most of the ultrasonic stuff, an NE5534 in the TI circuit and an NE5532 in the Onkyo circuit.
I don't recognize the last part of the Onkyo filter. I would guess it's either some sort of notch filter or some sort of group delay equalizer.
Noise shaping suppresses (re-)quantization noise in the band of interest and makes it worse outside the band of interest, so I don't agree with your statement about noise shaping and HF junk.
I don't recognize the last part of the Onkyo filter. I would guess it's either some sort of notch filter or some sort of group delay equalizer.
Noise shaping suppresses (re-)quantization noise in the band of interest and makes it worse outside the band of interest, so I don't agree with your statement about noise shaping and HF junk.
Don't really know other than its likely additional filtering as maybe they used a less sharp digital filter. I recognize the extra op amps from the design recommendation for the PCM63 many years ago so maybe you can find some answers reading up on that. Either way I would not worry to much and it should be easy enough testing before and after the extra op amps and decide for yourself what you like best.
It was mainly curiosity about their design decisions, but I am playing with the idea of scavenging the DAC board for a standalone I2S multichannel DAC. It is true that the digital filter in the PCM1796 goes down only to about - 100 dB stopband attenuation as opposed to -120 dB for the PCM1792. But surely, getting the better DACs and saving those op amps and passives would have been cheaper?
Marcel, it is true that you get HF junk with noise shaping, but usually a first order passive is enough to get rid of most of it.
I haven't yet analyzed what the 5th and 6th op amp do - but why would one want to have a notch filter there?
Marcel, it is true that you get HF junk with noise shaping, but usually a first order passive is enough to get rid of most of it.
I haven't yet analyzed what the 5th and 6th op amp do - but why would one want to have a notch filter there?
This is the output filter from the PCM63 data sheet with a generalize immitance converter (no idea what that means). It uses one op amp less than the Onkyo circuit (even after taking into account that no diff amp is needed).
The circuit around A2 in post #5 looks like an FDNR to me, a frequency-dependent negative resistor, a circuit of which the impedance drops with frequency squared. It's a trick to emulate LC filters without inductors: all inductors are replaced with resistors, all resistors with capacitors and all capacitors with FDNRs
Yes, I remember reading the app note. BB were pushing the idea heavily a long time ago. But the Onkyo circuit is not the same.
