[time-nuts] Thunderbolt settings

Richard Moore richiem at hughes.net
Mon Dec 1 20:59:43 UTC 2008


On Dec 1, 2008, at 7:08 AM, Chuck and Ulrich  wrote:

> ------------------------------
>
> Message: 2
> Date: Mon, 01 Dec 2008 09:02:33 -0500
> From: Chuck Harris <cfharris at erols.com>
> Subject: Re: [time-nuts] Thunderbolt settings -- one more time
>
> The T-Bolt is designed to be a turn key system, and is already
> set up for the best operation the factory knew how to achieve.
> I doubt that anyone has spent much time diddling the time constants.
>
> -Chuck Harris

Thx, Chuck for your reply.
>
> ------------------------------
>
> Message: 3
> Date: Mon, 1 Dec 2008 16:08:44 +0100
> From: "Ulrich Bangert" <df6jb at ulrich-bangert.de>
> Subject: Re: [time-nuts] Thunderbolt settings -- one more time
>
> Chuck and Dick,
>
>> The T-Bolt is designed to be a turn key system, and is
>> already set up for the best operation the factory knew how to
>> achieve.
>
> Yes and no! Clearly the factory settings have to be so that a lock of
> the pll is possible within a reasonable time, say minutes to hours or
> so. With a freshly powered OCXO the initial aging effects can be that
> large that really short time constants as the default 100 s of the
> Thunderbolt are necessary.
>
> I attach a measurement that I made on a freshly powered up HP10811  
> some
> years ago. Note: This was by no means a new device but one that had
> already thousands of hours lifetime. Clearly any regulation loop will
> need short time constants to follow this initial aging. Note that the
> aging effects are that big that frequency changes due to diurnal
> temperature changes in my flat are not visible! Temperature has
> otherwise the biggest impact of all environmental parameters on
> oscillator stability.
>
> After some weeks of continuous (!) operation the time constant of the
> Thunderbolt loop may be set to higher values. Mine currently runs at
> 500.0 s. Some tests indicate that the time constant may even be set a
> bit higher. I have made experiments on time constants with a number of
> different gps receivers and different OCXOs (including RBs). As an
> overall result it turned out that TCs > 1500 s are impraticcal in a
> normal living environment. With some degrees Centigrade temperature
> changes in a typical center European environment along the day that is
> the maximum time that allows the loop to follow the temperature  
> changes.
> This includes RBs which have an measurable coefficient of temperature
> too.
>
> Best regards
> Ulrich
>
>>
>> Richard Moore wrote:
>>> Dear nuts --
>>>
>>> Let me see if I can ask this in a way that will prompt you to supply
>>> some input:
>>>
>>> What is the maximum Loop time constant the TBolt (or perhaps the
>>> TBolt monitor sw) will accept as input?
>>> What TCs are you TBolt owners using?
>>>
>>> Thx,
>>> Dick Moore
>>>
>>> -------------- next part --------------
> A non-text attachment was scrubbed...
> Name: HP10811.pdf
> Type: application/pdf
> Size: 17055 bytes
> Desc: not available
> Url : http://www.febo.com/pipermail/time-nuts/attachments/20081201/ 
> f7077c5e/attachment.pdf
>
> ------------------------------

Thanks Ulrich. The default TC on my used TBolt was 100.0s. I ran it  
for a couple of days and it was stable with a very low level of total  
DAC change -- about +10mV. After a few more days, and after looking a  
various ADEV plots, particularly yours, Ulrich, I set the TC to  
1000.0 sec. It refuses to take 10,000 sec, but I haven't explored the  
limits of what it will accept, and was hoping, since Trimble won't  
tell me in the docs the acceptable ranges of anything you can adjust,  
that one of you would know.

Dick Moore



More information about the time-nuts mailing list