[time-nuts] [qs1r] Looking for good, cheap, external reference

Ed Palmer ed_palmer at sasktel.net
Wed Sep 22 23:39:19 UTC 2010


And if you buy a newer unit with the better oscillator, new firmware, 
and newer (but brain-dead) temperature sensor, it's trivial to replace 
the sensor.  It took me longer to take the unit apart than to change the 
chip and I don't have a lot of experience with surface mount.

Ed

Bob Camp wrote:
> Hi
>
> That's my feeling. The firmware isn't the issue. The OCXO and the temp sensor are the important points. The firmware is just an easy way to get the right OCXO.
>
> Bob
>
>
>
> On Sep 22, 2010, at 7:00 PM, Mark Spencer <mspencer12345 at yahoo.ca> wrote:
>
>   
>> Would a newer unit with the 3.00 firmware, the trimble branded OCXO and the 
>> older temperature sensor be one of the better (or perhaps best ?) TBOLT 
>> combinations ?
>>
>>
>>
>> ----- Original Message ----
>> From: John Miles <jmiles at pop.net>
>> To: lester at veenstras.com; Discussion of precise time and frequency measurement 
>> <time-nuts at febo.com>; David <n7aig at yahoo.com>
>> Sent: Wed, September 22, 2010 12:42:06 PM
>> Subject: Re: [time-nuts] [qs1r] Looking for good, cheap, external reference
>>
>>
>>     
>>>       John, unless I am going mad, the ADEV plots for both auctions you
>>> mentioned look the same - maybe the bad one has been replaced since
>>> you posted your message?
>>>       
>> The black-and-white plot that everybody uses is from the original
>> Thunderbolt data sheet, which was way conservative even for the older units.
>> Towards the bottom of the 290308733659 auction, though, the seller has
>> appended a half-dozen or so photographs of a TSC 5120A screen showing what
>> are (for a Thunderbolt) some weak results.  The ADEV plot shows poor
>> long-term disciplining and the PN plot doesn't look good either.  Possibly
>> he's using a noisy reference, or the unit was tested during a 2-day-long
>> earthquake.
>>
>> I've never seen one do ~-88 dBc/Hz at 10 Hz.  Even my original one with the
>> lower-quality OCXO was good for -105 dBc/Hz, while the ones with the
>> Trimble-labelled OCXO can do -130 or better.  Likewise the 'good' ones will
>> do several dB better than the -150 dBc/Hz broadband floor he shows.
>>
>> I've also never seen one in a black anodized(?) housing like that (has
>> anyone else?)  Mark is right in that the 3.00 firmware doesn't necessarily
>> do much for you, but the newer units also had the good OCXOs, which IMO is
>> more important than the temp sensor.
>>
>> -- john, KE5FX



More information about the time-nuts mailing list