[time-nuts] Re: Custom Quartz crystals

[Bob kb8tq]

Hi

In the era that the fixture was used, doing this with an network 
analyzer wasn’t even a pipe dream. Doing up an adequate set of
data with the gear of that era was a long and drawn out process. 
Getting a measurement with a simple fixture was a second or two. 

Indeed if 1% on resistance and a tenth of a ppm frequency wasn’t
adequate, you might head off to other approaches. Given the make
tolerances on the crystals, neither one was a limit to production testing. 

The biggest limit in reality was the accuracy of the load capacitor. 
That’s why you never / ever swapped it out. The value that was in there
was confirmed by some non-trivially exciting testing with a number of 
“golden” crystals. 

Bob

> On Aug 19, 2022, at 10:30 AM, Brooke Clarke via time-nuts <time-nuts at lists.febo.com> wrote:
> 
> Hi:
> 
> IEC 444 is a PI-Netowrk measurement method for crystals.  It is supposed to me more accurate than the method that uses a physical CL.
> The HP 41900A has the look and feel of the linked PI-Network fixture so may just be a re-branded version.
> https://prc68.com/I/4395A.shtml#E5100
> 
> -- 
> Have Fun,
> 
> Brooke Clarke
> https://www.PRC68.com
> axioms:
> 1. The extent to which you can fix or improve something will be limited by how well you understand how it works.
> 2. Everybody, with no exceptions, holds false beliefs.
> 
> -------- Original Message --------
>> Hi
>> 
>> Thanks to a bit of off list coaching from those who’s memory (obviously)
>> better than mine :) :
>> 
>> This is the magic test set that Motorola (and many others) got pretty
>> heavily into:
>> 
>> https://isolalab.com/pinetwork.html <https://isolalab.com/pinetwork.html>
>> 
>> The gizmo is right at the top of the page. There’s a nice paragraph next
>> to it and a diagram of what’s inside.
>> 
>> The original outfit that made it ( Cathodeon / Pye ) has been out of business
>> for quite a while. They apparently made the fixture to support their
>> crystal manufacturing business. Just how many fixtures they sold to
>> their competitors …. no idea.
>> 
>> Bob
>> 
>>> On Aug 17, 2022, at 8:17 AM, Bob kb8tq <kb8tq at n1k.org> wrote:
>>> 
>>> Hi
>>> 
>>> One “interesting” way to look at this is a “delta F” test.
>>> It’s based on a system that (essentially) always measures
>>> things at series. You put the (maybe) 20 pf  load cap in series
>>> with the crystal and zip off it goes to measure the net result
>>> of the crystal at a 20 pf load.
>>> 
>>> Typically you did this by pushing a paddle on a fixture. Yes
>>> that’s very specific to one rig up. Turns out that Motorola
>>> decided it was a fun thing and we tested a *lot* of crystals
>>> that way. It’s been “a while” … but .. fundamental crystals
>>> came in at > 400 ppm. Third overtones might be around
>>> 30 ppm. Fifth overtones might be around 3 to 5 ppm.
>>> 
>>> No this isn’t for the same crystal or even the same crystal
>>> package. It also isn’t for SC cuts, these all where AT’s.
>>> 
>>> … and yes, if I could remember the name if the UK outfit
>>> that made the test heads, I’d give them credit for their cute
>>> little gizmo.
>>> 
>>> One interesting outcome of testing this way was to spec
>>> crystals as some number of PPM off series rather than
>>> with a load. Changing the load cap in the fixture was “not
>>> done”.  AFIK that was very specific to internal Motorola
>>> documents. The only time I saw it used elsewhere was
>>> on crystals that operated below series.
>>> 
>>> Bob
>>> 
>>>> On Aug 17, 2022, at 5:28 AM, Richard (Rick) Karlquist via time-nuts <time-nuts at lists.febo.com> wrote:
>>>> 
>>>> It is hard to answer your question.  The tuning range
>>>> depends on the varactor(s) used in the respective
>>>> oscillators, and the range of the tuning voltages.
>>>> I think I could roughly say from memory that putting
>>>> in a crystal meant for the other oscillator would
>>>> probably result in not being able to reach 10.000000
>>>> MHz.  Or just barely reaching it and having no
>>>> aging margin.  I had to apply a lot of arm twisting
>>>> to Jack Kusters to get him to make the series
>>>> resonant (at 10 MHz) crystals.  I made sure that
>>>> this was really necessary before going to war on it.
>>>> 
>>>> Rick N6RK
>>>> 
>>>> On 8/14/2022 3:35 PM, Hal Murray via time-nuts wrote:
>>>>> Richard (Rick) Karlquist said:
>>>>>> We actually used the upper turnover because the E1938 had to operate at
>>>>>> ambient up 85  degrees.  BTW, the E1938 crystals had to be calibrated for
>>>>>> 10.000000 MHz at series resonance, unlike the 10811 crystals which were
>>>>>> calibrated for 20 pF (IIRC).
>>>>> How big is the difference between series and 20pF compared to the tuning range?
>>>> _______________________________________________
>>>> time-nuts mailing list -- time-nuts at lists.febo.com
>>>> To unsubscribe send an email to time-nuts-leave at lists.febo.com
>> _______________________________________________
>> time-nuts mailing list -- time-nuts at lists.febo.com
>> To unsubscribe send an email to time-nuts-leave at lists.febo.com
> 
> _______________________________________________
> time-nuts mailing list -- time-nuts at lists.febo.com
> To unsubscribe send an email to time-nuts-leave at lists.febo.com