[time-nuts] A couple of Tbolt questions

Magnus Danielson magnus at rubidium.dyndns.org
Sat Jul 25 22:17:06 UTC 2009


Tom Van Baak wrote:
>>> Typical TAPR Thunderbolt phase noise is about -100 dBc/Hz
>>> at 1 Hz, falling to about -150 at 100 Hz, with a floor just above
>>> -160 beyond 10 kHz. Variations on this, and spurs, seem to be
>>> a function mostly of power supply. For example, see:
>>> http://www.leapsecond.com/pages/tbolt/noise.htm
>>
>> Given the apparent sensitivity to the power supply that the Tbolt  
>> exhibits, has anyone put any effort into optimizing the power supply  
>> so as to minimize that set of effects?
>>
>> ie: what is the best power supply to use to ensure maximum accuracy  
>> for the Tbolt?  A couple of linear supplies?  Any particular switcher?
>>
>> Tom Frank, KA2CDK
> 
> This could be done but do keep it in perspective. The talk about
> power supplies is in the context of phase noise measurements
> not GPSDO accuracy or stability. The latter are more a function
> of TBolt design, or antenna placement, or environmental control,
> etc.
> 
> I have yet to see any difference in TBolt stability from one p.s.
> to another.
> 
> For most of us who use a TBolt as a precise time or 10 MHz
> frequency reference the choice of power supply is less of a
> concern. I suspect line/load regulation may be more important
> than the number of spurs or the linear vs. switcher question.
> 
> If someone has time it would be nice to know the df/dV for
> each of the TBolt +12, +5, and -12 volt inputs.

I was just wondering about that, which supplies breaks it?

The +12V supply surely goes to the heater of the OVEN. +5V for logic and
+12V and -12V for serial interface. I don't think that there is any 
analog EFC processing relating to the +/- 12V but then again, maybe 
there is?

I don't know, but what if the EFC DAC has no reference voltage? Just 
speculating. It should have it and it should be fairly obvious for 
holdover, but stranger things have happend before.

> I take it it's only the guys who need to multiply their 10 MHz up
> to GHz that phase noise becomes a significant issue.

Indeed.

A simple linear step-down supply could be built without too much hazzle 
and then feed it from +/- 15 VDC supply.

Cheers,
Magnus



More information about the time-nuts mailing list