3
\$\begingroup\$

This circuit is from the datasheet for a demo board for the CS4334 i2s audio DAC. schematic

I'd like some help analyzing the circuit. Not shown is the audio source, which is immediately before ALA and ARA. There is a blocking cap from the output of the aforementioned CS4334.

From what I can gather, this is a (near) unity gain inverting amp with high-pass and low-pass filtering added. The virtual ground is set to roughly half VCC (VA+5, +5 volts) because the amp is using a single supply polarity. That virtual ground will add DC bias to the input (which has a blocking cap off-screen) allowing "positive" and "negative" excursions of the audio waveform. The output has a DC blocking cap and a minimum load (100 kΩ). From what I can tell, the MC33202 is only special for the purpose of this circuit in that its output is rail-to-rail.

Am I missing anything in that description?

Is there a way to analyze the high- and low-pass bandwidths here? Also, what is the purpose of R2 being terminated on the input side between R24 and R16? Put another way, what is the purpose of the extra 1.2 kΩ past the point of feedback on the input side? Is the entire network of R3, R24, R16, C34 and C36 to be taken as some sort of multi-pole filtering network?

\$\endgroup\$
7
  • \$\begingroup\$ Each channel has a simple R-C low-pass filter followed by a sallen-key active filter. Together, these provide the reconstruction function required for the DAC. \$\endgroup\$
    – Dave Tweed
    Aug 2, 2021 at 17:24
  • \$\begingroup\$ These are most often designed assuming a voltage source having low resistance (Thevenin), as Voltage-spike has in his SPICE simulation. In your circuit, source resistance should be significantly lower than R4. \$\endgroup\$
    – glen_geek
    Aug 2, 2021 at 17:49
  • \$\begingroup\$ +1 for bsd devil dude \$\endgroup\$
    – Pete W
    Aug 2, 2021 at 18:03
  • \$\begingroup\$ nsayer, are you looking for analysis? Or just the results you already have below? \$\endgroup\$
    – jonk
    Aug 2, 2021 at 18:23
  • 1
    \$\begingroup\$ @jonk Really just help in understanding what was going on. Which I think I got. :) \$\endgroup\$
    – nsayer
    Aug 2, 2021 at 19:43

2 Answers 2

2
\$\begingroup\$

Is there a way to analyze the high- and low-pass bandwidths here?

Yes, use spice with AC analisys which is much easier\faster than trying to calculate all the total circuit impedance by hand. You can think of the R2/R3 as a voltage divider, and it makes C3 more effective at getting a higher rolloff slope on the low pass filter component of the circuit (as shown below, green is the AC response of the 'original' circuit and red is a modified circuit which shows a more 'rounded off' rolloff when C3 is connected on the other terminal of R16.

enter image description here

\$\endgroup\$
1
\$\begingroup\$

It's basically a 3rd order Butterworth using only 1 amp. with a -3dB at >50kHz

Without the feedback cap it is 2 stage LPF's with 2nd stage around 65kHz and 1st stage around 100 kHz and unity gain.

With the 270 pF at 10% of 2700 pF and negative feedback, it converts the summing junction below R2 as a 1st order HPF+ 2nd order LPF with the 3rd cap 270pf as the 3rd order LPF. This is a modification of the Sallen & Key 2nd order filter to create a 3rd pole. The ratios are specific to the Butterworth functions for staggered f-3db and Q.

This attenuates DAC noise with -60dB @ 1MHz.

enter image description here

Simulation

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.