Presonus 1818VSL mod/design issue

[kingkorg]

Hello, and marry Christmas!
I need to say my knowledge on dsp, dac, adc is very limited. While repairing my unit for some bad caps, and inputs i stumbled upon this potential issue in the design. C159 and C149 (adcdac.jpg) are 10uF each. In the CS4272(Cirrus.jpg) datasheet, they do not recommend values higher than 1uF, and say **Only one must be used**. There is also 0.1uF cap missing in the Presonus. Now the question is what might this cause? The use of oversized caps? I understand why 0.1uF should be used.

+15v, -15v (psu.jpg) for the opamps and pres doesn't seem to be regulated, just a bunch of additional caps for filtering for each pre besides what you can see in the psu image. Is this an issue? I was thinking about making external linear psu for this? Would this get me anywhere?

 

[Andy Peters]

Why the design deviated from data sheet recommendations for those pins is a question for the engineer.

Re: the ±15 V rails -- they are regulated. Look more closelier at the schematic.

 

[kingkorg]

Well, duh! It's regulated from the primary side, thanks

 

[Monte McGuire]

Some capacitors, especially the modern high value X5R and X5S parts "squish" when DC is applied, so their faceplate capacitance is not the capacitance that they present to the circuit. Some of these caps can lose 50 to 80% of their faceplate capacitance when significant DC voltage is applied. Still, too much capacitance is usually only a problem for the regulator - I cannot see how any load would care. It is also possible that the regulator requires more than 1µF of load capacitance for stability. For many modern LDOs, a minimum of 2µF is typical, with more output capacitance required by many of the new high performance LDOs.

 

[kingkorg]

The voltage across these caps is 2vdc.

Vcom stands for common mode voltage, which goes to the last op amp pair before the adc input. See the image.

CS4272 specs say that Vcom out has 25k impedance, so my best guess is they maxed the capacitance to filter out any noise before it goes to the opamps. However, with recommended 1uF cap cutoff frequency of the filter is 3hz, 20uF seems like overkill.

What i dont understand is what negative effect that extra capacitance( and inductance ) might have. Those 2V are internally regulated by the chip i guess. I'd lower the capacitor value, but i would like to know what to look for. Positive or negative effects.

 

[Khron]

My guess is that the "Only one should be used" refers to the "digital supply" feed of the codec - either a series resistor from the 5V that also feeds the analog side, OR a separate 3.3-to-5V straight to VD.

The extra Vcom capacitance may well be more as a current reserve rather than for frequency reasons, i'd reckon.

PS. Where the hell did you manage to dig up schematics for this thing? Presonus support wouldn't even supply (ready-compiled) firmware

 

[kingkorg]

Awesome, thanks