LMK5B12212: Multiple LMK5B12212 synchronization

Hi Eddie, 

Can you provide the customer's .tcs configuration file? For this kind of use case, I would recommend operating the slave LMK5B devices in zero-delay mode to ensure that the output clocks have phase determinism with respect to the input clocks. 

Hi Connor, 

Confirm customer with below clock tree, 

Master： All output 156.25 MHz

Slave1 ：All output 312.5 MHz

Slave2 ：Only8_9_10_11_OUTPUT=156.25 MHz ,  Other_OUTPUT=312.5 MHz

tcs file attached for different LMK5B, 

master.tcsslave1.tcsslave2.tcs

Customer run the tcs files and slave output doesn't synchronized with IN1 ( from master output)

Would you help check check customer setting? we can go through e-mail if you need customer schematic for reference. 

Thanks & Regards

Eddie

Hi Eddie, 

Sorry for the delayed response here. Tomorrow I'll test out the .tcs file that you sent on my setup and see if I can replicate the same behavior on my end. 

Hi Eddie, 

I reviewed the .tcs configuration files and it seems like there was an issue with the configuration for DPLL1. I re-ran the script in step 10 of the start page and now the settings look correct to me. Can you try using the updated configs I've attached here and see if there's any improvement? 

slave1 - modified.tcs

slave2 - modified.tcs

Hi Connor,

Customer still not able to get clock synchronized with tcs file you provided?

Below is are waveform and status page snapshot, is there something with HW setting?

Thanks & Reagrds

Eddie

Hi Eddie, 

Since REF1 is not being validated, there could be some issue with the input clock or reference validation settings. Could you ask the customer to try following the debug steps outlined in this application note? https://www.ti.com/lit/an/snaa421/snaa421.pdf Assuming the output frequency is correct then the APLL should be locked, so they could try starting from section 2.3 and onward. 

Hi Connor,

I get EVM back to office and try input 156.25MHz clock to IN0_P,  (Output from CDCE913 EVM) and remove XO VDD jumper, then open the tcs file you provided. 

I can get REF0_VALID_STATUS 

but I cannot clock output, is it possible EVM damaged?

Thanks & Regards

Eddie

Hi Eddie, 

The LMK5B device always requires an XO input, it looks like the .tcs config sent earlier in this thread is configured for a 48MHz XO. If you place the XO VDD jumper back are you able to see the correct output and lock to the 156.25MHz signal on IN0? 

Hi Connor,

Yes, I place XO VDD jumper back but clock not output not synchronized with REF

Is there setting we need modify in tcs? 

Below is status after jumper install ed. 

Can we setup a webex call discuss next week?

Thanks & Regards

Eddie

Hi Eddie, 

Sure, we can set up a webex call to discuss in more detail. In the meantime since REF0 isn't being validated, you could try relaxing some of the validation thresholds on the "validation" page in TICS Pro to see if you can get REF0_VALID_STATUs to read back as '1'.