[time-nuts] General questions about making measurements with time interval counter.

Dr. David Kirkby (Kirkby Microwave Ltd) drkirkby at kirkbymicrowave.co.uk
Thu Jan 12 04:36:10 EST 2017

On 12 January 2017 at 02:31, Bob Camp <kb8tq at n1k.org> wrote:

> Hi

Hi Bob

> The most basic issue you are going to run into is that your counter is not
> high enough
> resolution / accuracy to give you meaningful data for time intervals under
> a few hundred
> seconds.

Is that true if I'm not testing very high quality sources? There are
significant differences observed between these two setups

1) START and STOP from distribution amplifier.
2) START from distribution amplifier. STOP from Stanford Research SR345 30
MHz function generator set to produce 10 MHz.

1) ADEV plot at http://www.kirkbymicrowave.co.uk/time-stuff/ADEV.PNG
where ADEV is about 5x poorer on the function generator at 100 seconds

2) MDEV plot at http://www.kirkbymicrowave.co.uk/time-stuff/MDEV.PNG
where MDEV is about 5 x poorer on the function generator at 10 seconds.

The raw data is in the same directory
http://www.kirkbymicrowave.co.uk/time-stuff/ . I should have zipped that,
as it is quite large

I'm pretty sure the HP 83623A 20 GHz sweep generator will be a lot worst
than either of those, as a simple check of the standard deviation on the
display of the HP 5370B (no data collected from GPIB), showed much higher
SD on the microwave sweeper than the function generator.

I can see if I were testing masers, Cs source, and decent GPS receivers,
what you say would be true. But is it true for lesser quality sources?
Maybe testing lesser quality sources is not such a stupid idea, as the
instrumentation is less of a limiting factor.

I admit I did say in my original post SHORT/MEDIUM time scales, and I guess
10/100s is not short.

> I would focus on improving on that part of things before I went off on a
> major
> “test everything” adventure.

But is there any way without spending lots of cash? The following
instruments are out of the question due to price

* Keysight 53230A 350 MHz Universal Frequency Counter/Timer, 12 digits/s,
20 ps
* John's Timepod

A Stanford Research SR620 is not out of the question. I did have one
before, but swapped it, along with a 4.2 GHz signal generator for an
HP4391B impedance/material analyzer. The 5370B came along fairly cheap
($300), but I don't mind spending more on  a SR620. But will that gain me
much? I know the single shot resolution is a bit better than the 5370B, but
it does not appear to be a massive improvement, given they are 3~4 x the
cost.on the used market.

I also have many other contraiints, which limit what I can realistically

* Small lab in my garmage- opening door on lab will change temperature.
* Single glazed window
* Air con that is a standard unit designed for offices - not metrology

Simply collecting a lot of data that is resolution limited is
> not a lot of fun …..

Agreed. But am I doing that at > 10 seconds? Is the counter my limit on
those two sets of data? I will collect some from the HP 83623A 20 GHz sweep
generator later today, but that generates a lot of heat, and I'd rather let
that warm up for a couple of hours before doing anything with that.

> Bob

You clearly have a lot of knowledge Bob. Can I learn anything useful with
what I have, or am I wasting my time?


More information about the time-nuts mailing list