2.2uF C0G/NP0

[fabiodl]

I am developing a board around the TI pcm5122.
The datasheet at page 66 reports that all caps should be 2.2uF C0G.
I cannot find any caps with that dielectric/temperature coefficient on Digikey with capacitance over 560nF.
Do bigger C0G caps exist? What/Where should I search for?

[langwadt]

use 4?

[DaJMasta]

I'd be sort of surprised if they do, I'm actually surprised you can find a 500nF C0G, the ones I'm used to seeing top out in the single digits.


Why not a film cap instead?

[Jay_Diddy_B]

Hi,

I think is a an error the correct value is 2.2nF

They use 2.2nF in other places in the documentation.

2.2nF and 470 is 153kHz

2.2uF and 470  is 153 Hz low pass filter.

Jay_Diddy_B

[duak]

Wow, I'm surprised there are even 560 nF parts.  I think it's a mistake.  The output filter should be 2.2 nF C0G as per the schematic, while the 2.2 uF for the charge pump can be an X7R.

Here's a link to someone who thinks the same: https://electronics.stackexchange.com/questions/386216/pcm512x-charge-capacitor-value

[floobydust]

Na, it's several typo's. There are no C0G/NPO's that big and you would not put something so huge at the DAC's outputs.
Those are properly 2.2nF C0G?, and 2.2uF X7R are used for the charge-pump. Look at PCM5122 Evaluation Module http://www.ti.com/tool/PCM5122EVM-U

[BravoV]

A 100nF C0G/NPO is about 24 dollars a pop  , example -> Kemet C0G Cap

I can't imagine what will be the price if at 2.2uF.   

[fabiodl]

Thank you very much everybody.
The evaluation board's schematic/bill of materials really helps a lot, it was a typo indeed!

[exmadscientist]

A 100nF C0G/NPO is about 24 dollars a pop  , example -> Kemet C0G Cap

I can't imagine what will be the price if at 2.2uF.   

Huh? No they're not. I've used a bunch of 100nF NP0 caps when I need stability and it's not worth expending more design effort to use something cheaper. They're usually about a quarter, like GRM31C5C1E104JA01L at US$0.23 in quantity 100.

It also looks like Kemet has recently introduced a new NP0 line that's about a factor of 5 cheaper, but they're not stocked at distributors yet. (Example part at Mouser.)

[BravoV]

A 100nF C0G/NPO is about 24 dollars a pop  , example -> Kemet C0G Cap

I can't imagine what will be the price if at 2.2uF.   

Huh? No they're not. I've used a bunch of 100nF NP0 caps when I need stability and it's not worth expending more design effort to use something cheaper. They're usually about a quarter, like GRM31C5C1E104JA01L at US$0.23 in quantity 100.

It also looks like Kemet has recently introduced a new NP0 line that's about a factor of 5 cheaper, but they're not stocked at distributors yet. (Example part at Mouser.)

Yeah, you're right, those I pointed out is for 100V rated. 

[MagicSmoker]

Since the OP's question has been answered, note that there are C0G caps rated for up to 330n and 100V that only cost $5.78 on Mouser:

https://www.mouser.com/ProductDetail/KEMET/C2220C334J1GACTU?qs=sGAEpiMZZMs7ZEmUmaUL0%252BHfn%252BkysyX5xyIpm1WOce8%3D

[ZaphodBeeblebrox]

Replace with polymer caps. Only better, nothing worse, at least for audio DC blocking.

There is this thing about leakage current you might want to consider...

[fabiodl]

It is "funny" that in their schematic/bom they use x7r and not c0g even for the low pass output filter

[floobydust]

It's better C0G as they have lower dielectric absoption, critical in say a reconstruction filter, but not as much here because the DAC's output impedance and 470R resistor work around it.
Signal distortion from high-K ceramic capacitors
Maxim AN4333 only looks at X7R, Y5V and the low freq. <300Hz gets badly distorted.

The eval boards are usually to play with developing S/W and few have test results showing it actually achieves 24-bit resolution with low noise. +3.3VA also powers the DAC IC's
VD digital portion which might not be ideal.