LMX2594PSEVM: SYNC and MASH_SEED function

Hi Sandra,

in sync mode, N = 10 is less than the min. N divider restriction.

You need to reduce fpd to increase N divider. For example,

Now you are in Cat.2 sync, you can click the Toggle Sync Pin button to simulate a L-H-L operation at the SYNC pin, both devices on the board should be phase aligned. 

MASH_SEED is use to adjust the phase between input and output, not for sync.

Thanks for your answer. I will try this configuration and let you know. 

After the sync, the two devices are phase aligned between them and with respect the input, right?

Regarding the MASH_SEED function: we need to ensure certain delay between two devices. So, as we understood, if after sync both devices are phase aligned along with the input, if we apply certain phase misaligment (using MASH_SEED function) in one device we will obtain the required delay between devices. Is this correct?

Hi Sandra,

After synchronization, the phase between input and output become deterministic. if multiple devices share the same input clock, after sync, the phase between devices are also deterministic. We can use MASH_SEED to adjust their phases so that their edges are aligned. 

Hi Noel, 

Thanks for your answer once more. I tried this configuration, however I was unable to achieve full synchronization among input and outputs. It was not repeatable, the phase difference changed over time without any reason.

Since my application is Category 2 I understood that SYNC input can be generated from the reference pro board. That is why I connected the OUTP from reference pro board to OSCin port. Shall I divide the OUTP to both input ports OSCin and SYNC? Or shall I connect OUTP from the reference pro board to Oscin and OUTN to the SYNC port in the evaluation board? Or shall I introduce an external SYNC signal, a part from the OSCin signal coming from the Reference Pro board? 

Thanks in advance for your answer.

Dear Noel,

I synchronized both devices sucesfully. However I repeated the power-up sequence several times and I got different delay values between both synchronized outputs ranging from almost 0ps to 20ps.

Could you tell me what can be the reason for this? I need to ensure this value to be repeatable to later on adjust the mash_seed to stablish the desired phase delay.

Thanks in advance for your answer.

Hi Sandra,

Unfortunately, your configuration falls into the non-predicable condition.

I tried to reduce the N divider to > 45 with 2nd order modulator, 

we can change the phase correctly for a couple of steps but then it becomes not stable for another couple of steps.

keeps programming MASH_SEED = 20 will eventually make it unstable.

continue to program MASH_SEED, it resumes to normal again.

Is 1GHz your actual application frequency?

Yes, our application needs 1GHz signal.

Could you tell me which would be the configuration to obtain 1GHz output signal and be able to control in a deterministic way the delay with the mash_seed function? It seems to be not possible

Could you tell me a TICS Pro configuration that you have tested in the lab that proves that mash_seed function works as expected according to Equation 5 (no matters if it is a different frequency than our application)? I would like to reproduce it to verify that our evaluation board works properly.

Thank you very much!!

Hi Noel,

Kindly following up the thread. Could you let our customer know the appropriate TICS Pro configuration to test the mash_seed function?

Thank you

Hi There,

with LMX2594 device, whenever sync mode is enabled, IncludedDivide is greater than 1, so in order to prove the phase adjustment feature, we must not enable the phase sync bit.

If you program MASH_SEED = 50, you should get 2.25deg of phase shift. Program MASH_SEED = 50 one more time, total phase shift is 4.5deg.

Dear Noel, 

Thanks once more for your kind answer.

According to your last configuration on TICS Pro, output signal would be 4GHz, if you program MASH_SEED=50, the phase shift should be 9º. I guess it is a typo. I tried this example in the lab and it worked. Thanks. I also did it for 1GHz with MASH_SEED=50 and I got 2,25º phase shift.

However, we are still wondering how to use the VCO_PHASE_SYNC=1 and MASH_SEED function together at 1GHz, since this is our application frequency. Previously, you told me that our configuration falls into the non-predicable condition. How can we deal with this?

We can set a call first week of september to discuss possible solutions.

Looking forward to hearing from you soon.

Kind regards,

Sandra

HI Sandra,

right, 9deg at 4GHz; 2.25deg at 1GHz, I captured the wrong setting to you.

There is no workaround for LMX2594, you can use LMX2820, we have eliminated the IncludedDivide in this device, we can get smooth phase adjustment even in sync mode. 

Hi again Noel,

We plan to use LMX2615-SP, and one of your colleagues told us LMX2594 had similar behaviour, this is why we bought LMX2594 to test a COTS version in our lab. However, you tell us we can not use sync mode and mash_seed function together at 1GHz. Could you confirm if LMX2615-SP has the same problem or if is it solved, please?

Thank you very much in advance.

Hi Sandra,

Let me check tomorrow. 

Hi Sandra,

LMX2615 also has the IncludedDivide, the behavior of MASH_SEED in Sync mode is similar to LMX2594. I have a video showing how the phase shifted with MASH_SEED. The phase is not always shifted forward, it will shift backward and then forward again. Let me know if you want to watch the video, I will give you a link to access it. 

Hi Noel,

I will apreciate if you could send me the link to access to that video, thank you!

Regarding the phase you mentioned, that it shifted forward and backward, could you say that it is deterministic, I mean, does it shift the same amount every time or is it random?

Would you suggest any other space version that has not this problem with sync and mash_seed?

Thanks in advance for your answer.