CS4398 - Output Stage

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
PM2aCcl.png


What's wrong in my schematic ?
These measures are normal?
Should I keep the buffer in this way before the differencial , is that right?
Using symmetric source can rule out the use of capacitor at the output of opamp?
I'm out of scope at the moment , I feel that the sound is strange :confused:

I am using the OPA1611 in all parts , but is equal to the NE5532 , really strange.
 
Member
Joined 2009
Paid Member
The DC voltages seem to be correct.

There could be several issues here:

1. You have not shown the supply decoupling. With a relatively fast op-amp as OPA1611 this can be critical. Incorrect decoupling may can easily lead to instability.

2. The buffer op-amps will get a high level of high frequency signals directly on their inputs. That may create distortion.

3.You don't have much low pass filtering. This will send a lot of high frequency noise to the amplifier following you DAC buffer/filter.

If you want to use DC coupling, you could try using the circuit suggested by Cirrus Logic (Figure 11 in the data sheet) , modifying it by shorting the 100uF and the 22uF capacitors. It will give you DC coupling, but with more filtering.

You may also want to include a mute circuit as described in Figure 12, to avoid clicks on the output.
 
1. You have not shown the supply decoupling. With a relatively fast op-amp as OPA1611 this can be critical. Incorrect decoupling may can easily lead to instability.

Just do not put in the schematic, but I did so, 1000uF + 100nf in each op amp (+V to GND, -V to GND).


2. The buffer op-amps will get a high level of high frequency signals directly on their inputs. That may create distortion.

Do you think unnecessary to place buffers before the differential?


3.You don't have much low pass filtering. This will send a lot of high frequency noise to the amplifier following you DAC buffer/filter.

Indeed, the noise (hiiissss) is constant regardless of the dac is on or off. That is, is generated at this stage, or the PS, right? But I do not believe that comes from the PS, it is linear with multiplier capacitance, 8000uF each side of the symmetry. I think it's enough.

You think this is more from the filter, or is it more from the buffers before the differential?

Thanks for listening JensH. :D
 
Last edited:
Member
Joined 2009
Paid Member
The op-amp buffers are probably unnecessary (best case) and worst case they could create problems. As far as I remember there is a very strong switching noise on the CS4398 output, coming from the sigma-delta conversion.

Is the noise you hear like white noise?

What is the "multiplier capacitance" ? Is it the decoupling capacitors in the power supply? If so, that should be more than sufficient if it is only to supply the op-amps of the DAC output stage.
I agree that the hiss is unlikely to come from the PS. Of course I don't know the details of your PS. Noise from the PS is more likely to be in the form of hum (100Hz and harmonics).

In terms of op-amp decoupling I think the the 100nF capacitors are the most important. 1000uF on the op-amp supplies should not be necessary if you have a regulated power supply. But the layout around the op-amps, for both signal lines and supplies, is important.

Do you turn the DAC ON and OFF, but still power the op-amps? Or how should that be understood?
 
The op-amp buffers are probably unnecessary (best case) and worst case they could create problems. As far as I remember there is a very strong switching noise on the CS4398 output, coming from the sigma-delta conversion.

Okay, I'll remove the buffers.


Is the noise you hear like white noise?

The noise is like a breath.
It is as if the noise floor was around -60dB when it should be below -100db, do not you think? I'm no scope at the time to confirm this. :(


In terms of op-amp decoupling I think the the 100nF capacitors are the most important. 1000uF on the op-amp supplies should not be necessary if you have a regulated power supply. But the layout around the op-amps, for both signal lines and supplies, is important.

Yes, regulated power supply with 100nf in each electrolyte. Discard that comes from the source, as always used the same topology.
I will modify the layout, you can really be this many signal components near the power lines.

Do you turn the DAC ON and OFF, but still power the op-amps? Or how should that be understood?

Exactly, even with this isolated stage DAC, the noise floor remains.

I will redo, and post results.
Thanks! :)
 
Last edited:
I removed the buffers, and changed the schematic also with lower values of resistors.
Major cause problem was caused by using 47k in the input buffers.

Now, all right, without any noise. :D

But I am intrigued by these OPA1612.
It sounds to me, very like to the NE5532, perhaps with a little more SNR.
Is it because I'm used to the OPA2132? Or I have caught fake op amps 1612? :confused:

Look at the image, I suspect that they are fakes, but ... what do you think?

jUA7cYa.jpg
 
Member
Joined 2009
Paid Member
It is strange that the 47k resistors should generate so much noise. If they are in parallel with the (low impedance) outputs of the DAC, it shouldn't really be an issue. Or were they in reality in series with the DAC outputs? That would definitely add some noise.

The OP1612 is a nice, low noise, op-amp. I have used it in some designs with good results. The marking on the ones you have look different from mine, but TI may just have changed the marking at some point in time. I don't really know.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.