CDCE6214: accuracy

Allan Fan

Part Number: CDCE6214

Hi expert,

I want to follow up the e2e thread CDCE6214: Design Support that customer pcb is back. The measured output frequency is 53.696MHz but the target is 54Mhz.

Is there any way to improve the accuracy.

By the way, here is customer PN/jitter spec requirements:

Thanks,

Allan

over 1 year ago

0 Kadeem Samuel over 1 year ago

TI__Mastermind 23330 points

Allan,

What is the frequency accuracy of the 10MHz clock? The CDCE6214 will track the internal clock, so if there is any accuracy issue from the input clock it will be reflected on the CDCE6214 output.

Thanks,

Kadeem

0 ching-ting prng over 1 year ago in reply to Kadeem Samuel

Prodigy 60 points

Hi Kadeem

this is the frequency accuracy of the 10MHz clock, about +/-2ppb

Regards,

Amy

0 Cris Kobierowski over 1 year ago in reply to ching-ting prng

TI__Expert 6435 points

Hello,

Kadeem is currently OOO but will return on Tuesday 3/12. Thank you for your patience!

Best,

Cris

0 Allan Fan over 1 year ago in reply to Cris Kobierowski

TI__Expert 7460 points

Hi Chris,

Is any other team member can give us some advice e.g. what things can check or test that we can do before Kaedeem is back?

Thanks,

Allan

0 Cris Kobierowski over 1 year ago in reply to Allan Fan

TI__Expert 6435 points

Allan,

From looking at the previous thread, it seems like you have a CMOS output correct? But I notice in your schematic you only have 0ohm resistor on the output. Can you share the rest of your schematic for the output? Without the proper termination, reflections can occur, causing the frequency to not appear accurate.

Best,

Cris

0 ching-ting prng over 1 year ago in reply to Cris Kobierowski

Prodigy 60 points

Hi Cris

This is the CDCE6214 schematic

Regards,

Amy

0 Cris Kobierowski over 1 year ago in reply to ching-ting prng

TI__Expert 6435 points

Hello,

Apologies. This part does not require external termination for CMOS output, so the schematic is good.

Though, you're only using OUT1_P as an output correct? You're not using OUT1_N? If that's the case then here is an updated .tcs file. The prior one had both OUT1_P and OUT1_N outputting, which can also lead to ringback and frequency inaccuracy. Alternatively, you can rewrite R59[14] from a 1 to a 0, as everything else should be the same as the original configuration.

Best,

Cris

10 MHz in 54 MHz out (2).tcs

0 ching-ting prng over 1 year ago in reply to Cris Kobierowski

Prodigy 60 points

Hi Cris

Thank you for your suggesion. I found the reason for this. It's because the value of R27[1:0] is set to 0h (Integer mode). When I set it to 1h (1st Order Modulator enabled), the frequency output I see on the GUI matches what I measure. If there are any errors, please let me know.

But now I have another issue. When I write the registers into the EEPROM and power restart, I notice that the frequency becomes a different value. I have to recalibrate the PLL (R0 [4] = 1h) to return to 54MHz. Is there a way to resolve this? thanks

Regards,

Amy

0 Cris Kobierowski over 1 year ago in reply to ching-ting prng

TI__Expert 6435 points

Amy,

- Could you provide the config you are trying to write to the EEPROM?

- Are you writing to the device using the TICSPro GUI or your own I2C software?

Best,

Cris

0 ching-ting prng over 1 year ago in reply to Cris Kobierowski

Prodigy 60 points

Hi Cris

I using the TICSPro GUI.

This is the config file.

10 MHz in 54 MHz out 240328.tcs

Regards,

Amy

0 Kadeem Samuel over 1 year ago in reply to ching-ting prng

TI__Mastermind 23330 points

Amy,

We are OoO for a US holiday. We will check this configuration on an EVM when we return Monday.

Thanks,

Kadeem

0 Kadeem Samuel over 1 year ago in reply to Kadeem Samuel

TI__Mastermind 23330 points

Amy,

Ensure that you calibrate the PLL after writing the registers. When I write the registers without calibration I see ~52.1MHz. When I calibrate the PLL, I get the correct frequency. After writing to EEPROM and power cycling the frequency is still correct.

Thanks,

Kadeem

0 ching-ting prng over 1 year ago in reply to Kadeem Samuel

Prodigy 60 points

Hi Kadeem

I tested the frequency of power cycling and it is still different. Then I checked the PLL lock status and it is unlocked. Will this have an impact?

0 Kadeem Samuel over 1 year ago in reply to ching-ting prng

TI__Mastermind 23330 points

Amy,

Are you calibrating the PLL before you program the device registers? The device being unlocked absolutely will have an impact here.

Thanks,
Kadeem

0 ching-ting prng over 1 year ago in reply to Kadeem Samuel

Prodigy 60 points

Hi Kadeem

I will give it a try. Thank you for your suggestion.

Regards,

Amy

Clock & timing

Clock & timing forum

CDCE6214: accuracy