CDCE913 noise problem

user459077038

Other Parts Discussed in Thread: CDCE913, PCM4204

Hi,

I'am using CDCE913 with my design to generate 48KHz(Y1) and 12.288MHz(Y2 or Y3) clock . Crystal frequency=24.576MHz

With the Angilent E3631A to power the CDCE913.

I also place a 0.1uf(ceramic capacitor) and a 10uf(aluminum electrolytic capacitor) bypass capacitor on both power supplys(VDD and VDDOUT).

But when I generate the 48kHz clock,I found a about 500mV noise on the oscilloscope, see the attach picture. According to the oscilloscope,f(noise)=25Mhz which is approximately equals to crystal frequency.

What should i do to eliminate this noise ?

Does it ok for devises like PCM4204 to tolerance this kind of noise????

over 9 years ago

0 Julian Hagedorn over 9 years ago

TI__Mastermind 20900 points

Hello,

do you use the CDCE913 EVM or is it a new design? Could you share the power supply decoupling scheme (and full schematic around CDCE913)?
Could you also share the register programming?

regards,
julian

0 user459077038 over 9 years ago in reply to Julian Hagedorn

Prodigy 210 points

Hi Julian,

It's a new design,

Here is the full schematic.

I didn't use the PLL module (powered down).

So What's wrong with my design?

Thank you

Regards

Fantasotne

0 user459077038 over 9 years ago in reply to Julian Hagedorn

Prodigy 210 points

Hi Julian,
Sorry I forgot to explain my reply.
The FB means ferrite bead.
D01H means "Data of register offset 01h".
Registers that didn't show here remain their default value.

0 user459077038 over 9 years ago in reply to Julian Hagedorn

Prodigy 210 points

Hi Julian,

I think I have figured out the problem. When I measure the Y1 and Y2 using two channels on single oscilloscope. It will show a lot of noise.

But when I only measure one output(Y1 OR Y2), It will be normal.The dc wire and the oscilloscope was ground together.

So where is the problem?

0 Julian Hagedorn over 9 years ago in reply to user459077038

TI__Mastermind 20900 points

Hello,
did you have a ground connection for the scope probe?
also it might be that the scope grounds are not well isolated between both channels.

schematic and register settings are ok and do not explain your previous measurements. (in case of 50 ohm transmission lines R1 to R3 should be changed to 33 ohms).

Best regards,
Julian

0 user459077038 over 9 years ago in reply to Julian Hagedorn

Prodigy 210 points

Hi Julian,
Yes , I connect the two scope probe grounds with DC supply wire ground together.
Thank you for your advice!
Regards,
Fantastone

Clock & timing

Clock & timing forum

CDCE913 noise problem