[time-nuts] Observations and opinions needed on graph by new nut
John Ackermann N8UR
jra at febo.com
Wed Aug 3 21:50:58 EDT 2011
Just a note on the 8165: the primary output is from a decent quality
OCXO that's in an FLL (not PLL) circuit. The basic idea is that there
is a counter with a 1000 second gate time generated from (I think) a 10
MHz crystal that is tightly locked to WWVB with a very short time
constant. Because of propagation impacts, the CPU looks for several (I
think 3) consecutive gate times and if they are within a small tolerance
range makes a frequency adjustment.
So, two things: (a) the 10 MHz "Frequency Standard" output won't
(shouldn't) show short term propagation or modulation; and (b) you may
see a sawtooth or other step effect as the OCXO isn't continuously
locked -- since the EFC changes at most only once every several minutes,
you may see some drift before the FLL tweaks the frequency.
The overall accuracy of the unit is spec'd as 1x10e-9, so it's a few
orders of magnitude worse than the Z3801A.
Now, if you're tapping into the 10 MHz that's directly locked to WWVB,
that should closely follow the received phase (and noise), but when
unlocked you're looking at a free-running crystal that could be several
Hz off at 10 MHz. You could be seeing significant rollover effects from
A few years ago I replaced the failed OCXO in my 8164 (same basic design
as the 8165 but configured for lab use) and did quite a bit of digging
into the design. I have some stuff about it here:
paul swed said the following on 08/03/2011 09:11 PM:
> I think you misunderstand my suggestion. Forget the efc it should replicate
> the phase errors your seeing. Granted anyplace in that chain you could have
> a failed or failing component. I don't think you do.
> My suggestion is this. If your 3801 is working and Bobs right about the 3801
> software, I use it. Then for this discussion you have a accurate reference
> beyond what you will see from wwvb. (Lordy please don't rip me up for that
> comment.) So feeding that into the 8165s 10 mc vco instead of the vco lets
> the system work and whats left is the phase errors in propagation.
> Essentially you disconnect the 10 mc vco and use the 3801s reference.
> You can then use the 3801 ref in to a scope on 1 channel and the adjusted
> dancing 8165 into another channel and have some fun watching it.
> Should you see the phase error drop significantly then indeed the VCO or
> loop is causing a problem. However if the phase is still jumping at the
> phase detector and especially if semi in sync with the modulation depth.
> Guess what? Its propagation and my concern the added modulation depth.
> On Wed, Aug 3, 2011 at 8:55 PM, Paul A. Cianciolo<paulc at snet.net> wrote:
>> See comments below
>> -----Original Message-----
>> From: time-nuts-bounces at febo.com [mailto:time-nuts-bounces at febo.com] On
>> Behalf Of paul swed
>> Sent: Wednesday, August 03, 2011 6:31 PM
>> To: Discussion of precise time and frequency measurement
>> Subject: Re: [time-nuts] Observations and opinions needed on graph by new
>> Have not thought this through very well.
>> But it is what time-nuts is about.Assuming your 3801 is locked and checked
>> as Bob suggests. It puts out a 10 mc signal. At least mine does. Well if
>> stuck that into the oscillators output instead of the VCO. Guess what? You
>> sort of just have propagation and modulation errors left from your
>> That's what I thought that I would see by doing this experiment, the
>> propagation and modulation effects. I figured the oscillator might drift a
>> little but the main artifact would be prop and mod.
>> Maybe it is? Right now there is a graph building for 24 hours of just the
>> output of the Spectracom 8165, no antenna connected.
>> The result for the past 4 hours has been a smooth consistent saw tooth as
>> shown in the last part of my earlier graph.
>> Is the EFC the only way that a disciplined oscillator get controlled? With
>> your idea would not allow the WWVB signal to "modulate" the GPS oscillator.
>> I wonder if monitoring the correction voltage going to the 8165's OCXO
>> be worth looking at.
>> That might show something.
>> There are provisions for an RS232 connection on the back of the 8165, that
>> needs to be looked into.
>> Thank you
>> Or at least mostly that. Its reasonably easy for you to try and would be
>> great to see your graphs after you did that.
>> Also run longer then a day I seem to have seen interesting stability on a 1
>> week time scale. Sections of the chart that matched other days offsets.
>> Fun stuff.
>> On Wed, Aug 3, 2011 at 5:54 PM, Bob Camp<lists at rtty.us> wrote:
>>> What is the Z3801 saying it's doing if you hook it up to a computer?
>>> There are several programs out there that will give you a pretty good
>>> idea of what's going on inside the beast.
>>> On Aug 3, 2011, at 1:05 PM, Paul A. Cianciolo wrote:
>>>> Hello Folks,
>>>> New nut here.
>>>> I have posted a graph here.
>>>> First time doing this sort of thing.
>>>> There is a comparison made between a z3801 GPS 10 MHz and a WWVB
>>>> Could someone please look at it and validate my assumptions?
>>>> Being new to this I would like to make sure I am not going off on a
>>>> false path.
>>>> Thank you
>>>> Paul A. Cianciolo
>>>> Our business computer network is powered exclusively by solar and
>>>> wind power.
>>>> Converting Photons to Electrons for over 20 years
>>>> time-nuts mailing list -- time-nuts at febo.com To unsubscribe, go to
>>>> and follow the instructions there.
>>> time-nuts mailing list -- time-nuts at febo.com To unsubscribe, go to
>>> and follow the instructions there.
>> time-nuts mailing list -- time-nuts at febo.com To unsubscribe, go to
>> and follow the instructions there.
>> time-nuts mailing list -- time-nuts at febo.com
>> To unsubscribe, go to
>> and follow the instructions there.
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
More information about the time-nuts