[time-nuts] pps vs. 10 MHz timing

[Magnus Danielson]

Matt Ettus skrev:
> On my Oscilloquartz GPSDO, the 10 MHz output goes low at very close to
> the same time as the 1 PPS output goes high.
> 
> On my Fury, the 10 MHz sine wave is just dropping off of its max high
> voltage as the 1 PPS goes high.  The 10 MHz CMOS output goes high just
> shortly before the 1 PPS.
> 
> What do other GPSDOs do?

You can expect anything. Some buffers the PPS from the GPS engine and 
then make some more or less elaborate 10 MHz training algorithm. Don't 
expect rising edges to coinside for all of them.

> Clearly, any device trying to latch the 1 PPS signal using the 10 MHz
> clock will need to choose which clock edge to use, depending on which
> type of GPSDO is used.  How is this normally handled?

Depends, several approaches is possible. A simple approach is just a 
dual flip flop approach which is simple and straightforward. There is a 
risk for it to switch between 10 MHz cycle, but it may be acceptable for 
some applications. You could use any of many forms interpolating 
techniques to reduce or remove this risk. Exactly what you can achieve 
depends not only on the receiver hardware but also the transmitter side.

The trouble is that the 10 MHz + PPS is a very weekly defined interface 
which is not properly standardised. The only standard I know off is an 
old US MIL-STD document.

> Also, I noticed that the TADD-1 has some weird distortion of a 10 MHz
> sine wave.  It turns it into a sort of a 3 valued square wave.
> Essentially, it stops and dwells at the midpoint for a while instead
> of making a sharp transition.  I think this would be bad if that dwell
> point is close to the threshold for the receiving device.  Has anyone
> else noticed this?

Yes, on a signal not properly terminated. It occurs if you T a signal to 
your scope while the signal goes further on... to an open end. Then the 
reflex signal will create this type of dwelling.

Cheers,
Magnus