[time-nuts] HP53132A vs SR625
ed_palmer at sasktel.net
Sun Mar 17 13:56:50 EDT 2013
That equation looks similar in form to the specs for any counter. What
are the comparable equations for the 53132A or the 5370(A or B)?
On 3/17/2013 10:41 AM, Said Jackson wrote:
> The error I have seen was in the high xE-011's to the low xE-010's. the only way around it was to turn on relative measurements, which then subtracted out this error. That error makes the unit almost useless to me.
> The factory told me as long as it is within specs they will ship it after cal and not bother trying to improve this.
> The "acceptable" specs are pretty crappy in tim-nuts terms: +/-350pico * frequency with a 1s gate time. Thats straight from the user manual and assuming no reference error. From the manual:
> Frequency Accuracy:
> < ± ((100ps typ [350 ps max])/Gate + Timebase Error ) x Frequency
> That equates to up to +/-3.5E-010 if my math is right!
> Not so impressive. I can confirm Ricks comment about HP 53132A units not showing that type of error.
> Not knowing how big the error is without another non-SRS counter to compare to, and if it may actually be bigger than spec is a problem. I don't remember if it shows up when feeding the counter its own reference.
> Sent From iPhone
> On Mar 16, 2013, at 16:06, Volker Esper <ailer2 at t-online.de> wrote:
>> What "small error" are we speaking about? The statement that SRS users have to tolerates a small error while HP users don't seems a little to general to me. IMHO we might be a bit more precise. Anyone who's already done an error analysis for - say - a 10MHz count and a comparison of the counters?
>> In real life every type of equipment has it's domain, where it has it's specific advantage. Could it be, that's the case for these counters, too?
>> Am 16.03.2013 19:57, schrieb Rick Karlquist:
>>>>> 1) I paid quite a bit of money and I had it "calibrated" and fixed by
>>>>> and it still exhibits a significant frequency offset with a "perfect"
>>>>> reference and "perfect" DUT!!!
>>>>> SRS says a small frequency error is "normal", well that prevents me from
>>>>> using the unit as a frequency counter, for me it's only useful as a
>>>>> display frequency counter. HP doesn't have such a frequency error, so no
>>>>> worries there.
>>> I worked with the guy who designed the HP53132A. He would
>>> never tolerate as "normal" a so-called small error. The term
>>> "frequency counter" brings to mind something that digitally counts
>>> zero crossings and should never have an error. First of all, even
>>> if that is all you do, it is still possible to screw it up.
>>> Secondly, "counters" have relied on analog interpolation even going
>>> back to the HP524 circa 1950. There is no theoretical basis of having zero
>>> error in this case, but the idea is that you display the number of
>>> digits that are commensurate with the worst case accuracy of your
>>> interpolator. Again, my colleague who designed the interpolator
>>> did very high quality work. I am pleased to learn that our stuff
>>> is better than the stuff from the company up the road.
>>> Rick Karlquist N6RK
>>> HP Santa Clara Division 1979-1998
>>> (still working for Agilent!)
More information about the time-nuts