enrico.bellotti at gmail.com
Thu Feb 25 17:50:18 EST 2016
Bruce, Mark and Magnus
thank you for your comments. Incredibly enough, I have found some time to
follow up and address some of the issues you outlined.
I measured the rise time of the start and stop signals for different cable
lengths and types of generators.
I have also used a HP8901A/8903A that generates square waves with ~350ps
transitions and 1.2V amplitude. This is the fastest source i have.
I attach a pdf with the details and the Adev results.
It looks like that the length of the cable effects the Adev, probably
caused by the stop signal rise time degradation.
It is interesting to notice that for short cables (~3ft on HP5370B #1) the
Adev increases. I wonder if, unless of other measurement errors, the cable
is too short to "decorrelate" start and stop.
I have just had time to try one different references (TB 10MHz) instead of
the internal one but i do not see much of a difference. I assume a large
set of measurements is needed.
Finally, HP5370B #2 is off anyway. I am going to look into the multiplier
10->200 section as it was suggested.
On Wed, Feb 24, 2016 at 2:23 PM, Magnus Danielson <
magnus at rubidium.dyndns.org> wrote:
> Dear Enrico,
> On 02/24/2016 06:55 AM, Enrico Bellotti wrote:
>> Hello to all,
>> first of all, thank you for the great and useful work that the time-nuts
>> have done over the years.
> I have finally been able to gather all my counters (HP5335A, HP51131A,
>> HP5370A, HP5370B #1, HP5370B #2) and try to do some "simple" measurements.
>> I have started testing and comparing the instruments I have available
>> the approach that was discussed a while ago on the list and also outlined
>> by John Ackermann (http://www.febo.com/pages/hp5370b/). Specifically, the
>> Adev for a time interval measurement on a (90ft of) RG58A/U cable.
> It's a bit of a cable, so it can eat in on the rise-time somewhat. Did you
> check the risetime?
> I have attached a PDF file with the results and some additional details of
>> the test setup.
> Nice setup.
> From what I have understood the Adev at one second is related to the
>> counter resolution. The results I have obtained seem to be reasonable
>> except for HP5370B #2. This instrument seems to be marginal at best. Does
>> anybody know if the measured value for HP5370B #2 is a symptom of a
>> multifunction or simple need for calibration?
> Yes, it is somewhat related. The 1/tau slope you see is expected.
> You can usually expect the slope to be in the neighborhood of
> single-shot-resolution/tau, which is rule of thumb. It's more complex as it
> depends on the experienced trigger jitter, which depends on the noise and
> the slew-rate at the trigger point. You can thus optimize the jitter by
> adjusting the start and stop trigger voltage.
> The cable delay will act to decorrelate the triggers, but for most
> designs, you don't need to go to 135 ns but can keep them tighter, of the
> benefit of maintaining nice slope. However, when the stop trigger comes
> just handfull of ns after the start trigger, then the remains of the
> start-event can shift the stop trigger. The cable decorrelates this effects
> so it behaves more as separate signals, so that is good.
> The ripples you see for shorter taus for PPS signals would be interesting
> to see the reason for, the phase plot should help to illustrate the reason.
> However, it is curious how you provide measures from 0.1 s for a 1 Hz (1 s)
> PPS signal.
> However, it is nice to see the relative close correlation between the PPS
> and 1 kHz signals. It would be nice to see if a slew-rate measurement of
> the two sources could be related to the ADEV differences.
> Thank you for any comments/suggestions/corrections you may have.
> Hope you got some input from my ramblings.
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to
> and follow the instructions there.
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 218859 bytes
Desc: not available
More information about the time-nuts