[time-nuts] Practical considerations making a lab standard with an LTE lite

SAIDJACK at aol.com SAIDJACK at aol.com
Tue Nov 25 14:48:05 EST 2014

That's why I said its up to the user to decide what they want their  
trade-off to be.
For permanent installations I personally would not run the unbuffered 10MHz 
 output through more than about a foot of coax cable to the buffer.
The rise/fall time of the TCXO output is slow enough (typical spec is  4ns) 
to make that a lumped system rather than a reflected system. You won't  see 
any reflections on a foot or less of cable.
For short-term phase noise measurements I have run that signal through 6  
feet of coax no problem, but there are quite significant reflections at that  
point so I would strongly advise against that. If I break the TCXO here on 
my  bench due to my own stupidity its a different situation than if the 
customer has  that happen in their setting..
In a message dated 11/25/2014 09:34:11 Pacific Standard Time,  
csteinmetz at yandex.com writes:

Said  wrote:

>The increased current for the driver will cause heating near  the 
>crystal in both the CMOS driver and the 3.0V LDO as the LDO has to  
>convert the excess voltage into heat. This may or may not affect the  

There would be next to no additional heating in the CMOS  driver, 
because there is very little voltage across it in either logic  
state.  And the additional power supply current is so small that the  
increase in LDO dissipation will also be very low.  At the extreme  
worst, any such effects would be somewhere between imperceptible and  
negligible.  But on the other hand, if there is a possibility that a  
passive filter can create a clean, 50 ohm sine wave output for free,  
the potential up side is huge.

>Adding an external buffer is so  simple that I just did not think it 
>would be worth it..

An  external buffer is a fine way to go, but it would need to be very 
close to  the driver chip -- which is why I suggested on Sunday 
building it onto a  breakout card that plugs directly onto the LTE 
Lite's MMCX output  connector.  You really don't want to run a naked 
CMOS output at 10MHz  much farther than that, both for the corruption 
it may suffer and also for  the mischief that radiation and capacitive 
coupling can cause to other  nearby circuitry (the LTE Lite) as the 
loop gets larger than  that.

I'm not sure I see why a small additional source of heat is such  a 
dramatic concern with the 10MHz TCXO, but apparently not for the  
20MHz TCXO, which by accounts has an actual buffer amp that must  
create comparatively massive heating.  A temperature difference isn't  
a problem in and of itself -- only a changing temperature creates a  
problem.  Whatever the dissipation situation is, it should settle  
into stasis if one takes the slightest care with the thermal  design.

Best  regards,


time-nuts  mailing list -- time-nuts at febo.com
To unsubscribe, go to  
and follow the  instructions there.

More information about the time-nuts mailing list