[time-nuts] Re: Why do have OCXO a Vref output? (was: help reviving Trimble UCCM-LPS GPSDO)

Bob kb8tq kb8tq at n1k.org
Fri Dec 24 14:15:12 UTC 2021


Hi

> On Dec 24, 2021, at 4:26 AM, Hal Murray <halmurray at sonic.net> wrote:
> 
> 
> kb8tq at n1k.org said:
>> The gotcha is the common ground found on most OCXO???s.  Putting the oven
>> current through that ground makes keeping the Vref (and the EFC) well behaved
>> a bit of a chore. For various reasons OEM???s have a hard time dealing with
>> multiple grounds that have multiple purposes ???. 
> 
> That sounds like something a good app-note could solve.
> 
> Is the problem more complicated than that?
> Has the app-note not been written?
> Do people not read it?

You would think so, but that’s not the case. Even going out and doing
a full up presentation about “why” turns out to be inadequate. 

> ...
> 
> Assuming the pinout on the osc is right...
> How much trouble does I*R drop on a ground plane contribute?

First up: do you even *have* a ground plane? 

How many ground layers do you have and how thick are they?

How many square feet of board do you have? 

Is it a double oven with the heaters running on 3.3V? (yikes …)

Is the current (even in that 3.3V double oven) completely insignificant 
compared to each of a dozen other devices on the board?


>  Do I need to run 
> a separate trace for the heater return back to the power supply?

If you are running traces, yes. If it’s a ground plane, that sort of 
eliminates the value of the plane. In typical low power designs a top
and bottom ground plane should do the job well enough to not 
need to go super crazy.

If you are going to run a trace for something, the negative side of 
your external Vref (and / or  DAC) is what you normally do. Noise issues
may not allow you that luxury ….

Bob

> 
> -- 
> These are my opinions.  I hate spam.
> 
> 
> _______________________________________________
> time-nuts mailing list -- time-nuts at lists.febo.com -- To unsubscribe send an email to time-nuts-leave at lists.febo.com
> To unsubscribe, go to and follow the instructions there.




More information about the Time-nuts_lists.febo.com mailing list