[time-nuts] time-nuts, frequency counters

[Magnus Danielson]

From: "CHazlitt" <chazlitt at mtaonline.net>
Subject: Re: [time-nuts] time-nuts, frequency counters
Date: Wed, 26 Sep 2007 13:41:30 -0800
Message-ID: <F58E254685A84F8194E49A4EDF6FC56B at Vista>

Hi!

> I just joined the group a couple of weeks ago and have one remark about an 
> HP counter that I bought on ebay and also one question. The HP 5386A I own 
> incorporates a 12-character display which works out wonderfully when using 
> the 10 MHz output of a Rubidium source to clock the counter. The one I 
> bought goes up to 3 GHz but there is a less expensive version of the same 
> thing which although doesn't go to 3 Gig, has the same resolution. I found 
> this unit to be a great way to compare the Rubidium's frequency with an HP 
> GPS clock source.

The cheaper ones is the HP 5384A (50 Hz - 225 MHz) and HP 5385A (50 Hz - 1 GHz).

> I bought a Frequency Electronics FE 5680A Rubidium card on ebay too and 
> found that when comparing it's output at 10 MHz to the output of the HP GPS 
> standard that they did not agree when comparing the two using the above 
> mentioned HP counter (using the HP GPS as the clock for the counter). The 
> Rubidium I purchased is a few years old but locks up every time in just 
> three minutes so I know it is in good shape but yet it didn't agree to 
> the -9 on the frequency counter.
> 
> So, here is my question, do Rubidium standards drift that much over a period 
> of years to where they have to be brought back on frequency?

Yes.

> If so, what is tuned on the Rubidium to do so, C-field?

You can either tune the C-field or you can tune the synthesis-chain.

For my old R&S Rubidium the synthesis-chain is fixed and the C-field is
adjusted instead. Big vernier knob on the front and a CV-input on the back.

> I didn't touch anything on the physics package of the FE 5680A but the unit
> has a DDS divider built into it so I was able to adjust the Rubidium to agree
> with the HP GPS standard through an RS 232 serial connection and hexidecimal
> inputs to change the frequency of the dividers output. The divider they use
> can adjust the frequency far finer than the best resolution of any Rubidium
> standard so I thought that would be a good way to match the output of the GPS
> standard. 

I think it is a good strategy.

> I'm new to time standards, GPS or Rubidium units and purchased it for some 
> telecom work so I'd appreciate any suggestions or comments anyone here might 
> have.

We tend to use time interval counters and the HP 5370A or HP 5370B is popular.
Get once of those.

> One last question in regard to counters. Is it possible, or does anyone make 
> a frequency counter which has 12 digits but only displays 1 Hz and below 
> regardless of the frequency which is fed into it? i.e. 1 Ghz reads out as 
> 1.000 000 000 00 (if on frequency) only displaying the last cycle of the 
> count down to -11 ?

I have counters which display more digits than 12 if needed.

Also, for stable sources we use spaced time-interval measures and plot those
over time. We time-nuts seems eager to calculate our Allan Deviation charts,
mumble obscuria about not optimal here and there... and for tau this and that
you would have to consider... :)

Welcome to the group!

> Thanks for the bandwidth

You didn't waste any bandwidth, but you did use up some bits. Not the same! :)
(time-nuts are notorious nitpickers too!)

Cheers,
Magnus