Hi,
I designed a board for a CS43122 DAC, wich has voltage differential outputs. I used a simple output stage consisting of two non-inverting op-amps performing a small gain to get 4Vrms on an XLR output. A first order low pass filter at 120kHz, symetrical, between the CS43122 and the amps. Two 22uF MKT caps and another first order low pass on the amp's output at 380kHz. It is almost the same as represented in the demo board.
My question is, wich op-amp for the best results? In my opinion, there is no need of a very high slew rate because of the voltage output already filtered at 120kHz... however some fast amp like an AD826 could maybe do a great job?
I choosed a non-inverting amp, in this case many fet op-amps presents distortion at high frequencies probably due to a bad CMRR, like OP2134.
A old NE5532, OP249, LM6172, do you guys have any advice for me?
Thank you!
Ychnaval
I designed a board for a CS43122 DAC, wich has voltage differential outputs. I used a simple output stage consisting of two non-inverting op-amps performing a small gain to get 4Vrms on an XLR output. A first order low pass filter at 120kHz, symetrical, between the CS43122 and the amps. Two 22uF MKT caps and another first order low pass on the amp's output at 380kHz. It is almost the same as represented in the demo board.
My question is, wich op-amp for the best results? In my opinion, there is no need of a very high slew rate because of the voltage output already filtered at 120kHz... however some fast amp like an AD826 could maybe do a great job?
I choosed a non-inverting amp, in this case many fet op-amps presents distortion at high frequencies probably due to a bad CMRR, like OP2134.
A old NE5532, OP249, LM6172, do you guys have any advice for me?
Thank you!
Ychnaval
a couple of things they don't explain very well; the dac output op amp's are directly feed by a sw cap I-V circuit guaranteeing high frequency clock feedthru, and one reason for the differential output is the digital noise in the dac chip's substrate gnd - differential output puts the onus of removing the dac gnd noise on the rf common mode rejection of the users' buffer/filter circuit
given the 10s of mV of MHz hair you really want passive RC pole(s) in front of your op amp's input stage, fet inputs are more linear when hit by rfi beyond the op amp's feedback control range, really fast op amps can have the bandwidth to linearly treat more of the interference leaking through the passive input filtering
distortion from common mode V in noninverting fet op amps is largely due to nonlinear input fet capacitance's charging current flowing in the source impedance, low source impedance and balancing + and - op amp input impedance largely eliminate it (to maintain low noise your filter source impedance should be low enough to minimize this effect already)
fast fet input op amps often have relatively high Vnoise and high 1/f noise corner - most would degrade the 122's noise but still would deliver way more dynamic range than could be appreciated outside of an anechoic chamber
to preserve the full noise performance of the '122 dac you may have to go to bipolar input op amps, only experimentaion and measurement can tell if there is meaningful differences due to input rfi rectification
the next step in distortion reduction is to go to composite op amp circuits like Walt Jung's, esp for line driving from the dac buffer's output
given the 10s of mV of MHz hair you really want passive RC pole(s) in front of your op amp's input stage, fet inputs are more linear when hit by rfi beyond the op amp's feedback control range, really fast op amps can have the bandwidth to linearly treat more of the interference leaking through the passive input filtering
distortion from common mode V in noninverting fet op amps is largely due to nonlinear input fet capacitance's charging current flowing in the source impedance, low source impedance and balancing + and - op amp input impedance largely eliminate it (to maintain low noise your filter source impedance should be low enough to minimize this effect already)
fast fet input op amps often have relatively high Vnoise and high 1/f noise corner - most would degrade the 122's noise but still would deliver way more dynamic range than could be appreciated outside of an anechoic chamber
to preserve the full noise performance of the '122 dac you may have to go to bipolar input op amps, only experimentaion and measurement can tell if there is meaningful differences due to input rfi rectification
the next step in distortion reduction is to go to composite op amp circuits like Walt Jung's, esp for line driving from the dac buffer's output
Yes, AD826 is very good!
Depending on your taste, you may like OP275...
2134 is too warm and wobbly - good to hide harsh-sounding digital signature of so many 'HighEnd' stuff that's around.
Try feeding Vout straight from CS chip to XLR’s – you might like it… (if your amp can give enough gain). If you opt to do this experiment, use the coupling capacitors wire leads to connect the Vout pins and XLR’s . Do not use PCB or hook-up wiring - if possible.
Regards,
Extreme_Boky
Depending on your taste, you may like OP275...
2134 is too warm and wobbly - good to hide harsh-sounding digital signature of so many 'HighEnd' stuff that's around.
Try feeding Vout straight from CS chip to XLR’s – you might like it… (if your amp can give enough gain). If you opt to do this experiment, use the coupling capacitors wire leads to connect the Vout pins and XLR’s . Do not use PCB or hook-up wiring - if possible.
Regards,
Extreme_Boky
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.