EMRFD Message Archive 116
Message Date From Subject 116 2006-08-30 17:40:34 Hans Summers Crystal motional parameter measurement
I am temporarily estranged from my EMRFD book, in so far as it is in
one location and I am in another, with no means to access it for
perhaps some weeks. So I wonder if I might ask some questions about
motional crystal parameters?
Today I constructed a Butler oscillator and made measurements of
motional inductance and motional capacitance on a set of 50 10.14MHz
crystals. I used the G3UUR method which I believe is referenced in
EMRFD. However, whilst G3UUR uses a Colpitts oscillator, I used a
Butler.
I documented my construction and measurements on my web page here:
http://www.hanssummers.com/radio/xtalchar .
My first question is, is it valid to use a Butler rather than Colpitts?
And the second question, assuming that my motional inductance and
capacitance measurements are to some extent useful: I want to build a
ladder filter using these crystals, what criteria should I use to
select a set of 4 (say) well-matched crystals from the 50? What
constitutes "well matched"? I wish to build a signal frequency crystal
filter to cover the 30m QRSS band 10,140,000 to 10,140,100 . (for more
info on 30m QRSS see http://www.hanssummers.com/radio/qrssjb and also
scroll down to the bottom to see links to other pages).
Thanks & 73 de Hans G0UPL
http://www.hanssummers.com117 2006-09-01 11:12:51 Jim Kortge Re: Crystal motional parameter measurement
> Hello,Greetings Hans,
>Nice job on all of the documentation! :-)
> I am temporarily estranged from my EMRFD book, in so far as it is in
> one location and I am in another, with no means to access it for
> perhaps some weeks. So I wonder if I might ask some questions about
> motional crystal parameters?
>
> Today I constructed a Butler oscillator and made measurements of
> motional inductance and motional capacitance on a set of 50 10.14MHz
> crystals. I used the G3UUR method which I believe is referenced in
> EMRFD. However, whilst G3UUR uses a Colpitts oscillator, I used a
> Butler.
>
> I documented my construction and measurements on my web page here:
> http://www.hanssummers.com/radio/xtalchar .
>Probably, but I don't consider myself an expert on the various oscillator
> My first question is, is it valid to use a Butler rather than Colpitts?
topologies. That said, the basic premise of the G3UUR method is to cause the
crystal to oscillate at a different frequency related to adding series
capacitance. Since that happens with your Butler oscillator, I think you are fine.
>Use the 4 that are the closest to each other. The departures in Hz should be
> And the second question, assuming that my motional inductance and
> capacitance measurements are to some extent useful: I want to build a
> ladder filter using these crystals, what criteria should I use to
> select a set of 4 (say) well-matched crystals from the 50?
less than 1/5 to 1/10th the filter bandwidth. In your case, I'd say a filter
with a 200 Hz bandwidth ought to be about right to cover the 100 Hz spread of
the QRSS signals. So if the crystals are matched withing 20 to 40 Hz, you ought
to be ok. The tighter the better.....
What
> constitutes "well matched"?See above. When I do crystal filter sets, I can always get them within 10 Hz of
each other, and often within 10 Hz across the whole filter set. Usually I'm
working with 100-300 crystals in the set, so lots of selection.
I wish to build a signal frequency crystal
> filter to cover the 30m QRSS band 10,140,000 to 10,140,100 . (for moreI've attached a schematic of a 200 Hz BW Cohn filter and its response that were
> info on 30m QRSS see http://www.hanssummers.com/radio/qrssjb and also
> scroll down to the bottom to see links to other pages).
generated with the AADE Filter program. I assumed the crystal Q to be 100K,
since we don't have an Rs or Q value, and you need one or the other in Neil's
filter program. The attachments will probably get stripped off of the reply
that goes to the EMRFD group, but you will get them. Note the input and output Z
of this 4-pole filter approximately 10 Ohms, so you'll probably need to use some
impedance matching into and out of the filter in the circuitry into which it
will go.
One of the drawbacks that I see with the G3UUR method is that you don't have any
measure of each crystal's internal loss, or Rs parameter (Q too) I've found that
crystal Rs values vary all over the place. Within a set of 100 crystals, the
mean Rs could be 10 Ohms, but maybe 10% of the units will have Rs values 3 times
(or more) that. Those work fine for Local Oscillator service, but would add
significant loss in a filter using one of those elements, or worse yet, a whole
set of them. My preferred method of measuring and matching crystal elements is
to drive them with a generator and measure the output of each unit with an RMS
meter. Those with excessively high Rs values really stand out using that method.
72 and thanks for an interesting (at least to me) post,
Jim, K8IQY118 2006-09-01 18:25:02 Hans Summers Re: Crystal motional parameter measurement
I did a fair amount of reading, after somebody else suggested the AADE
program to me and I downloaded/installed it. The help file is very
informative, the section on crystal filters includes several magasine
articles. The whole subject is still only marginally clearer than mud
to me right now, but I'm sure a couple more readthroughs and it will
probably start making a little more sense.
> > And the second question, assuming that my motional inductance andThat's great, thanks. The one remaining question, well I suppose TWO
> > capacitance measurements are to some extent useful: I want to build a
> > ladder filter using these crystals, what criteria should I use to
> > select a set of 4 (say) well-matched crystals from the 50?
>
> Use the 4 that are the closest to each other. The departures in Hz should be
> less than 1/5 to 1/10th the filter bandwidth. In your case, I'd say a filter
> with a 200 Hz bandwidth ought to be about right to cover the 100 Hz spread of
> the QRSS signals. So if the crystals are matched withing 20 to 40 Hz, you ought
> to be ok. The tighter the better.....
remaining questions... firstly, when you say the closest to each
other, in frequency terms: do you mean the frequency of the crystal
just by itself, or with the series capacitance present? My table of
values has two frequency columns and I'm wondering which of the two
frequency columns should be matched.
The second question is that if one was building an IF filter, then the
exact centre frequency of the crystal filter is arguably not too
crucial. You'll just adjust the BFO later to be at the right place on
the filter's skirt. However in my case, I am thinking of using the
10.14MHz ladder filter as the receiver's input filter, specifically at
10,140,000 to 10,140,100 Hz. None of the 50 crystals I tested are
quite high enough, so presumably I will have to do something like pull
them up to the right frequency band before putting them in the cohn
filter?
Excuse my ignorance ;-) The beauty of this hobby is that there's
ALWAYS something new to learn!
73 Hans G0UPL
http://www.hanssummers.com