[time-nuts] Re: pps pulse timestamp device

[Hal Murray]

As others have pointed out, the TICC is a wonderful tool for your problem.  
It's a lot more expensive than a Teensy but a lot cheaper than typical lab 
gear with equivalent resolution.  It needs an external 10 MHz clock.  It has 
much higher resolution than your ns.

I think your ns is a nasty case.  It falls in a hole between easy for a CPU 
and what you get with expensive lab gear.

If you need better than a ns, the TICC is wonderful.

If you need less than a a ns, you might make do with something like a Teensy.  
Are you happy writing that sort of software?  Check the data sheet and see how 
fast the counters actually run.  Often the counters connected to IO pins run 
slower than the CPU speed.

The TI ARM chip on the BeagleBone series of boards has a pair of 
fast-but-not-smart CPUs designed for this sort of thing.  I've never used 
them.  I'm pretty sure there is Linux software available for timestamping, but 
a quick search didn't find any.


What are you trying to measure?  Do you want the pulse to pulse jitter or the 
offset of the PPS from a GPS relative to some wonderful truth?

With a TICC, you can measure the pulse-to-pulse jitter with a stable reference 
clock (crystal)  that isn't exactly 10 MHz.  (If you stand on your head, you 
can turn things inside out and use a PPS from a GPS to measure the frequency 
of a TICC's external clock.)

If you want to measure relative to some truth, feed a PPS from that truth into 
the second channel on the TICC.



-- 
These are my opinions.  I hate spam.