PRS10 PPS and 10MHz

[tchiwam]

Ok, this is bugging the hell out of me....

I tried to configure the PRS10 PPS and 10MHz so that the phase is kept in sync between the 2... but it is still jumping around, kind of 3 steps in a loop on my scope, PPS real steady then 10MHz moving a bit in 3 positions. If I do a Rise PPS+high 10MHz, the 10MHz stays in place and so is my 1.290GHz carrier, dead on, so I know my TDS694C is NOT bad ...

I am now thinking to add a PPS output with a little of logic and keep the PPS start at the edge of a 10MHz period and let it go down 100µs later on another rising edge....

Am I missing, some settings in the PRS10

[DaJMasta]

If you take a look at the datasheet for the PRS10 you can see the architecture - the PPS output is computed by the processor to give output stability in the long term, it is not a divided down 10MHz output signal, so the OCXO driving the 10MHz output is effectively disciplined according to the PPS output signal, but it will never perfectly be in phase with it in the short term - it is designed to be perfectly in phase with the output for time constants approaching infinity.


If you want a perfectly in phase PPS output from a 10MHz clock, you should just divide down the clock with some or a configurable divider chip, the architecture of the PRS10 will never guarantee you that in the short term.

[tchiwam]

Are they any good Rb clock with displined PPS and 10MHz that are in phase ? available at some reasonable cost that is ... 

PS:  PRS10 new is still in the reasonable price bracket...

[DaJMasta]

Do you actually need 1PPS for your application or some other frequency?  If your needed frequency is high enough, you just need to lock a PLL for your output frequency to the 10MHz output, if you actually want something as slow as PPS, you just need a divide by 10,000,000 circuit or frequency synthesizer that takes the 10MHz as a reference.

I don't really know of one with an output where the PPS output is just a divided down output of the main OCXO, but I don't know the architecture of that many Rb oscillators - it's not hard to do, but I don't think most implementations bother because the short term phase drift on the PPS is going to be less than the GPS PPS doing the initial disciplining, so it's sort of clean enough.

[tkamiya]

You could perhaps use this and generate pps from 10MHz output?

https://tapr.org/product/tadd-2-mini-pulse-per-second-divider/

It's open source, so you can even make your own.