Coaxial choke balun??
KG4GGX
I'm looking for recommendations on a coaxial choke balun for a 6 meter
dipole.I plan to make the choke from RG-8X. At this time, I've made one from
RG-8X that is 4 feet of coax, 5 turns at about 2.5 inches diameter. How far off
am I? Do I need a longer length of coax, more turns, smaller or larger
diameter? I'd really like to use an effective choke to keep the feed line from
radiating since 6 m seems prone to RFI.
Thanks for all replys.
73 Rob KG4GGX
'Doc
Rob,
I'd have used a coil diameter of around 6", and 8 turns.
It isn't very critical. Small diameter coils with RG-8x is
asking for problems from conductor migration (center wire
gradually moves to the outer side of the coil, changes
impedances, and can eventually short out).
'Doc
Mark Keith
That might not be too far off. In the arrl ant book they state appx 8
ft-8 turns for 20m. 4 ft-6-8 turns for 10m. "For rg-58, which is nearly
the same size as rg-8x". So I would think appx 3 ft-4to 6 turns might be
fairly close to optimum just by guessing. You are not too far off from
that, and it may be working ok. If you have no problems using it, I
would consider it good enough. MK
--
http://web.wt.net/~nm5k
Reg Edwards
wound with a conductor as thick as the outside diameter of the coax you
intend to use.
A solenoid, 4 inches long, 2 inches former diameter, close-wound with 16
turns of 1/4 inch outside diameter coaxial cable, has a self-resonant
frequency of about 50 MHz.
For other dimensions of coil and coax download in a few seconds program
SOLNOID2. In addition to the essentials of coil design this program
calculates the self-resonant frequency due to parasitic capacitance, and the
impedance at the resonant and lower frequencies.
The PVC jacket on coax will cause some uncertainty in the exact resonant
frequency but in choke-balun applications this is a relatively unimportant
matter.
( Having said all that, I think a 2" diameter ferrite ring filled with
loosely-wound coax does a better job. It has a smaller loss, it has a much
wider operating bandwidth, it weighs less and is mechanically more
convenient to install.)
---
***********************************
Regards, Reg, G4FGQ
Free radio design & modelling software
http://www.btinternet.com/~g4fgq.regp
***********************************
KG4GGX <kg4...@aol.com> wrote in message
news:20010212174348...@ng-fs1.aol.com...
KK7HO
KG4GGX
have and go from there with the info you all have given me. Thanks again.
73 Rob
Reg Edwards
the solenoid.
(2) Connect it to an Autek or MFJ antenna analyser via a loosely
twisted-pair line about 10" long.
(3) Switch the antenna analyser to the Z range.
(4) Suspend the solenoid in mid-air.
(5) Z will be in the range 0 to a few ohms. Vary frequency slowly searching
for a very sharp, resonant increase in Z.
Maximum Z will very likely be in the range 50 to 500 ohms at the
self-resonant frequency. Q will be extremely high, too high to be determined
from the 3dB points on the resonance curve.
The method works for solenoids 2 or 3 inches long, and for helically-loaded
dipoles, 1" diameter, 6 feet long.
Its an old wives' tale that Q is poor when most of the capacitance is of the
parasitic variety. Unless frequency is varied very slowly the resonant point
will be missed.
Program SOLNOID2 predicts the self-resonant frequency before you've wound
it. Program MIDLOAD deals with continuously-loaded dipoles which are just
very long solenoids damped by radiation resistance.
--
***********************************
Regards, Reg, G4FGQ
Free radio design & modelling software
http://www.btinternet.com/~g4fgq.regp
***********************************
Reg Edwards <g4fgq...@btinternet.com> wrote in message
news:96bene$m90$1...@uranium.btinternet.com...
Richard Harrison
Reg`s statement: "It`s an old wives` tale that Q is poor when most of
the capacitance is of the parasitic variety." was inserted to draw a
response, I suspect. Q usually depends upon high capacitance in a
resonant circuit.
For any resonant circuit, its frequency is a constant when the product
of L and C is a constant. The same resonant frequency can be obtained
with high C and low L, or high L and low C.
High L and low C is the case of the self-resonant coil. A high-L
resonant circuit maximizes the lossy element in a resonant circuit.
Q is the inductive reactance of a coil divided by its resistance. In
most resonant circuits Q is the same as the coil Q because a properly
made capacitor has very low losses.
A self-resonant coil is a coil without added capacitance in which most
of the capacitance is of the "parasitic" variety. As such, the lossy
coil is maximized, the low-loss capacitance is minimized, and the Q of
the resonant circuit is relatively low.
Best regards, Richard Harrison, KB5WZI
Reg Edwards
> A self-resonant coil is a coil without added capacitance in which
most
> of the capacitance is of the "parasitic" variety. As such, the
lossy
> coil is maximized, the low-loss capacitance is minimized, and the
Q of
> the resonant circuit is relatively low.
Rich, I'm afraid you are in danger of becoming an old wife.
The actual value of coil Q is, as you say, its
reactance/resistance, X/R. Q depends only on frequency because X
and R depend on frequency. Actual Q remains unchanged regardless of
the capacitance across it, self C or otherwise. ( The
between-turns capacitances are negligible. Why ? Because they are
all in series across the coil.)
So for a given frequency, actual Q remains at its normal high value
even when the only capacitance in the circuit is the coil's
self-capacitance.
But what the circuit designer is interested in is EFFECTIVE Q and
this depends on what sort of circuit the coil is used in.
If the coil is used in a PARALLEL-TUNED circuit, with a capacitor
in parallel as in an antenna trap, then Q' is the normal value of
X/R. If the self capacitance and additional capacitance are in
parallel with each other there is no reason to suppose circuit Q is
any different from actual Q. For the same voltage across the coil
its loss remains the same regardless of the extra shunt
capacitance.
However, if the coil is used in a SERIES-TUNED circuit the
situation may dramatically change. There is a circulating current
around the coil and its self-capacitance. The current flowing in
the coil is greater than that flowing through the series resonating
capacitor. So coil loss is greater than that due just to the
current flowing in the main series circuit. Effective circuit Q
therefore is then less than actual coil Q.
Let Fself = Self-resonant frequency.
At a frequency F it is not difficult to prove that the effective Q'
of a series tuned circuit is given very simply by -
Q' = Q * ( 1 - k )
where k = Square( F / Fself ) and
Q is actual Q = X/R of the coil at the frequency F.
The effective series loss resistance, R', is given by -
R' = R / Square( 1- k )
Where R is the actual coil loss resistance.
The effective series inductance L' of the coil is -
L' = L / ( 1 - k ).
It will be seen that when the series capacitance is equal to the
stray capacitance, effective circuit Q is half of the coil's actual
Q and this occurs at a frequency of 0.707 of the self-resonant
frequency.
At the coil's self resonant frequency, circuit Q is obviously zero
because the series tuning capacitor reduces to zero pF and there is
an open circuit. Also the effective series loss resistance becomes
infinite.
When the circuit is series resonant at a frequency of 1/3rd of the
self resonant frequency, effective Q is 89% of actual coil Q which
is acceptable for nearly all applications.
A series tuned circuit of interest to the antenna newsgroup, where
the self-capacitance of a coil can seriously affect performance, is
an antenna loading coil. A loading coil is in-series resonance
with the capacitance of the short length of antenna wire or rod or
top hat beyond the coil. When attempting to increase radiation
resistance, as the coil is raised nearer to the end of the antenna
its inductance necessarily increases very rapidly. And so does its
resistance even for the same Q. Simultaneously the coil's
self-resonant frequency decreases. Effective Q falls like a brick.
Loss resistance increases much faster than radiation esistance.
The coil becomes hot. The optimum coil height on a very short
antenna is usually less than 2/3rds of total height. Often on
moderate length antennas any additional loss caused by not
externally loading the antenna at all is hardly worth bothering
about.
Program LOADCOIL examines the effect on radiating efficiency of
sliding a loading coil along an antenna. Self-capacitance of the
loading coil is taken into account.
Tom W8JI
<g4fgq...@btinternet.com> wrote:
Reg,
>Rich, I'm afraid you are in danger of becoming an old wife.
>
>The actual value of coil Q is, as you say, its
>reactance/resistance, X/R. Q depends only on frequency because X
>and R depend on frequency. Actual Q remains unchanged regardless of
>the capacitance across it, self C or otherwise. ( The
>between-turns capacitances are negligible. Why ? Because they are
>all in series across the coil.)
>
I've noticed your programs greatly overestimate Q in some cases Reg.
Many software inductor programs do so.
That's probably because the software writers think the turn-to-turn
capacitance is negligable, and that shunting capacitance can not lower
Q at a given frequency with all else constant.
73 Tom
Peter O. Brackett
Q is a somewhat nebulous concept. Quality is always subjective I suppose.
Q has several definitions in the literature and one has to make sure what
definition is being described and discussed before entering upon a
measurement or calculation of its value.
The "Q" found in the electrical and electronics literature refers usually to
the "quality" of energy storage devices.
If energy storage devices were perfect we wouldn't need Q. If companies
were honest we wouldn't need a Better Business Bureau.
Q then arises only because of secondary or undesirable effects.
In other words, in a perfect world, Q would be infinite and we would not be
discussing it.
There is a reciprocal concept often called "dissipation factor" sometimes
labeled with the character "d", usually defined such that dissipation factor
d = 1/Q.
In a perfect world d is always zero and again, we would not be discussing
it.
Any devices capable of storing energy can have a Q. A capacitor [condenser]
can have a Q. An inductor can have a Q. A tuned circuit, which is a
combination of an inductor and a capacitor can have a Q.
Some think that Q or d is a constant. Because it arises from secondary or
undesirable side effects Q or d is not a constant and varies with other
parameters, such as frequency, current levels, etc...
I believe that the "best" definition of Q for any energy storage system and
the one that most often corresponds to the formulae found in most references
is:
Q = Energy Stored Per Cycle/Energy Dissipated Per Cycle
Simple calculations for an inductor L having a parasitic resistance
equivalent to a series resistance R then result in:
Q = (2 pi f L)/R, with similar formula for a capacitor with an equivalent
parallel conductance G, etc...
The problems in finding simple formulae and or software algorithms to
calculate this Q arise in calculating the values of L and R from first
physical principles. For instance the parasitic resistance varies with
frequency due to skin effect, and the equivalent L varies with frequency
because of the effects of parasitic interwinding capacitances, all of which
are functions of the physical geometry of the devices, etc...
How are we to believe in any computer program that purports to calculate Q?
We could ask the programmer to share his/her formulae and code with us so
that we could understand what factors enter into the calculations.
Reg? Care to share?
But clearly, if we are from that great State of Missouri in the USA, "the
show me State", we should compare the computer program output to laboratory
measurements of Q.
How are we to do that without a Q meter at our disposal?
And... even if we have a Q meter, whose design is it? And how does it
measure Q?
Was it made by Boonton, made by HP [Sorry Agilent], or Marconi, or Wandel
and Golterman, or Siemens' or what...
Reg, care to share with us which instruments you have used to develop the
experimental measurements that you have used to validate the formulae and
algorthims in your programs?
Thoughts, comments?
Peter K1PO
"Tom W8JI" <2w...@contesting.com> wrote in message
news:3a8e744d...@news.accessunited.com...
Reg Edwards
> I've noticed your programs greatly overestimate Q in some cases
Reg.
> Many software inductor programs do so.
>
> That's probably because the software writers think the
turn-to-turn
> capacitance is negligable, and that shunting capacitance can not
lower
> Q at a given frequency with all else constant.
Tom, as a non-software writer, please tell us what YOU think.
In your reply could you be quantitative and include some accurate
numerical information? Otherwise words like "overestimate" and
"negligible" have no meaning.
By the way have you ever previously come across the formula
relating Q', L' and R' for series resonance to the corresponding
parallel resonance values, which I gave on my previous posting to
this thread.
----
Reg
Tom W8JI
<ab...@ix.netcom.com> wrote:
>But clearly, if we are from that great State of Missouri in the USA, "the
>show me State", we should compare the computer program output to laboratory
>measurements of Q.
>
>How are we to do that without a Q meter at our disposal?
When I measure the Q of an inductor with nearly zero turn-to-turn
spacing, the Q can easily be half or less what it is when the turns
are spaced one wire diameter.
When I look at "programs", they almost always show the opposite.
I can design RF inductors using software that have Q's in the
thousands, but the highest I have ever measured is in the 800
range!!!! Those are generally large tank coils used in high power
transmitters.
I tried to do a notch filter for 160 meters, and the inductor Q is
what made the project not "do-able".
Of course if I used software inductors, it would have worked like a
charm. IMO, the BBB ought to put out a consumer watch for inductor Q
software programs.
73 Tom
Tom W8JI
On Sun, 18 Feb 2001 03:05:04 -0000, "Reg Edwards"
<g4fgq...@btinternet.com> wrote:
>In your reply could you be quantitative and include some accurate
>numerical information? Otherwise words like "overestimate" and
>"negligible" have no meaning.
You seem to think a definite number that is wrong is better than
saying "the result varies". I can't give any exact figures, because
the errors are all over the place. The software simply does not track
the real world, it isn't even remotely close as the frequency is moved
up towards the self-resonant frequency of the inductor.
Have you measured inductors, and compared them to you software? What
is the highest Q you have measured in a reasonably sized inductor at
radio frequencies, and how did it compare to your program's "estimate"
of Q?
73 Tom
Reg Edwards
> When I measure the Q of an inductor with nearly zero turn-to-turn
> spacing, the Q can easily be half or less what it is when the
turns
> are spaced one wire diameter.
>
Tom, if you take the trouble of using program SOLNOID2 before
criticising it you will find that at HF the computed Q is at a
maximum when the ratio of wire diameter/winding pitch is usually
between
0.5 and 0.7
Also, it is not the increase in capacitance between turns which
reduces Q when the spacing between turns decreases.
----
Reg, G4FGQ
Reg Edwards
Sorry Tom, if you are unable, as I requested, to provide any
quantitative evidence or solid facts about the uncertainty in your
own calculations or in your own estimates or measurements of Q,
then your criticism is clearly unfounded and there's nothing more
to be said in the matter. Have a good day !
----
Reg, G4FGQ
Tom W8JI
<g4fgq...@btinternet.com> wrote:
>Tom, if you take the trouble of using program SOLNOID2 before
>criticising it you will find that at HF the computed Q is at a
>maximum when the ratio of wire diameter/winding pitch is usually
>between
>0.5 and 0.7
I'll look at solenoid2. The program I downloaded did not behave that
way, so I assume 2 corrects somewhat for winding capacitance?
>Also, it is not the increase in capacitance between turns which
>reduces Q when the spacing between turns decreases
It isn't?
73 Tom
Roy Lewallen
Q prediction program I've seen (admittedly not including Reg's) predicts
higher Q -- often much higher -- than I can measure. I've felt this was
most likely due to failure to account for proximity effect, but I've
never known for sure. We did have some discussion on either this or the
homebrew newsgroup a while back about the real meaning of inductor Q and
whether it should or shouldn't include the effect of shunt capacitance.
Some methods of measuring Q include the effect of shunt C (that is, the
Xl becomes modified by the effect of shunt C) and some don't. The
consensus was that you really need to specify whether shunt C is
included or excluded when you specify inductor Q, and you need to be
aware of what your measurement includes.
Roy Lewallen, W7EL
Tom W8JI
wrote:
>More important might be current crowding due to proximity effect. Every
>Q prediction program I've seen (admittedly not including Reg's) predicts
>higher Q -- often much higher -- than I can measure. I've felt this was
>most likely due to failure to account for proximity effect, but I've
>never known for sure. We did have some discussion on either this or the
>homebrew newsgroup a while back about the real meaning of inductor Q and
>whether it should or shouldn't include the effect of shunt capacitance.
>Some methods of measuring Q include the effect of shunt C (that is, the
>Xl becomes modified by the effect of shunt C) and some don't. The
>consensus was that you really need to specify whether shunt C is
>included or excluded when you specify inductor Q, and you need to be
>aware of what your measurement includes.
>
I'm not sure what the problem is Roy, but the Q values are
unrealistic. This is especially true above about 1/2 of the
self-resonant frequency of the inductor.
While Reg's program (even solenoid2) shows Q steadily increasing, real
world measurements show the parallel resonant Q peaks at about half of
the self-resonant frequency and then smoothly and rapidly decreases as
the self-resonant frequency is approached. Of course series resonant
systems peak much lower.
It is also very easy to "design" an inductor with a Q of well over
1000, yet I have never seen one over the high 800's (although there
might be some out there).
I assumed the error was in how it handles the capacitance, but it
might be anything. Anyway, as long as you stay away from anywhere
near the self-resonance it is a useful program. Except the damn
numbers are all in some odd meteric form, instead of the US standard.
73 Tom
Tom Twist
<snip>
>I assumed the error was in how it handles the capacitance, but it
>might be anything. Anyway, as long as you stay away from anywhere
>near the self-resonance it is a useful program. Except the damn
>numbers are all in some odd meteric form, instead of the US standard.
Correction.
The numbers are according to international standards ;)
73, Tom LB8X
---
Email: t...@twist.no (ttw...@online.no, lb...@qsl.net)
Phone: +47 35972928/90662366, Ham call sign: LB8X Locator: JO49UA
Snail: Tom Twist, Kirkeveien 8A, N3970 Langesund, Norway
Cecil
>
> On Tue, 20 Feb 2001 04:25:22 GMT, 2w...@akorn.net (Tom W8JI) wrote:
> >Except the damn
> >numbers are all in some odd meteric form, instead of the US standard.
>
> Correction. The numbers are according to international standards ;)
One wonders why an Englishman doesn't use English units. :-)
--
http://www.mindspring.com/~w6rca
-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 80,000 Newsgroups - 16 Different Servers! =-----
Peter O. Brackett
I can only assume that your programs do what you say they do... however you
are testing my assumptions about your good work now, because... from your
observations below, I see that...
Your pique is showing...
That is a sure sign of insecurity. The retreat into denial and
communications breakdown. What?
How about sharing with us some of your algorithms/formula and comparisons to
actual measurements.
I't unfair of you to ask Tom to make measurements to verify your programs
for you.
Thoughts, comments,
-Peter K1PO
"Reg Edwards" <g4fgq...@btinternet.com> wrote in message
news:96pr15$rp7$1...@plutonium.btinternet.com...
Reg Edwards
Peter O. Brackett wrote -
> Reg:
>
> I can only assume that your programs do what you say they do
Pete, there's a flaw in your first line. Regarding predicting
accuracy I don't state say what my programs do. I leave it to users
to discover for themselves they are accurate enough for the
intended purposes.
==========================
>... however you are testing my assumptions about your good work
now, because... from your observations below, I see that - Your
pique is showing...
==========================
Not pique - just tiredness caused by evasive replies.
==========================
>
> How about sharing with us some of your algorithms/formula and
comparisons to
> actual measurements.
>
==========================
I can assure you, you would be unable to understand the source code
and and follow the track of reasoning! After a few months absence I
have difficulty myself. There would be more quibbling and haggling
and sidetracking between unqualified participants than has occurred
with Conjugate Matches and Fractals put together. This would wreck
the confidence which has accumulated during the last couple of
years. In any case this is a radio amateur newsgroup - not a
mathematics and computer programming training course.
==========================
>
> It's unfair of you to ask Tom to make measurements to verify
> your programs for you.
>
==========================
There are no rules to this game. ;o) I did not ask Tom to veryify
my programs for me. I asked him to numerically quantify his
criticisms rather than voice an opnion. In other words - what was
the source of HIS data with which he had compared programs.
Everyone already knows the source of MY data - it's the programs
themselves. And his criticism was of only one of the computed
parameters in one of 35 or so programs.
To clarify your thoughts -
Tom originally said that all computer programs over-estimate Q and
max Q occurs when coil turns are almost touching. He said this was
due to programmers ignoring the effect of capacitance between
turns. He specifically included my programs with the others. ( I
have never seen any others myself.) But he did not state what HIS
measurement standards were.
I asked him to look at SOLNOID2 for the wire spacing at which
maximum Q occurs. In another branch of this topic he then withdrew
his criticism about my calculation of Q and replaced it with a
criticism of the International Standards of Measurements. He also
withdrew his view that Q decreases with close spacing because of
increased capacitance between turns.
Tom did not replied to my question as to whether he had heard of
the formula I gave which relates Q, L and R to stray capacitance.
In reply to your request for the secret "algorithms" it is used in
SOLNOID2. It implies that the effective Q of a coil depends on
whether the coil is used in a series LC or a parallel LC circuit.
Tom did not mention that important (in the present context) fact.
For example, the effective Q of a coil is lower in a Pi-match
output circuit than the same coil when used in a tuned tank
circuit.
A comment on Q measurements -
If, as Tom says, Q values at HF cannot be greater than 1000 or
thereabouts, then to obtain any degree of accuracy the Q of the
conventional Q-meter itself would have to be several times greater
than that and accurately known. And to correct measurements
calculations are unavoidable.
Working values of Q when inside a screening enclosure or adjacent
to poor materials may be appreciably less than values measured on
an isolated coil. Calculated values obviously apply to isolated
coils. It is important to note it is not the coil which gets warm
under enclosed conditions but the enclosure. When a magloop or
dipole is near the ground it is not the magloop or dipole which
warms up but the ground.
But accurate values of Q are not required for engineering purposes.
It is sufficient to know Q is greater than some minimum value and
calculated values are good enough. Indeed very high calculated
values may be more accurate than the measured variety.
How is the intrinsic Q of a Q-meter determined ?
The only other use for Q values is to provide an enduring topic for
argument between perfectionists ( bless them ) on this newsgroup.
;o)
=========================
>
> Thoughts, comments,
>
=========================
See above.
Reg
Richard Harrison
My response concerned a self-resonant coil as compared with coils using
external capacitors in parallel with the coil to resonate at at the same
frequency.
I believe that Reg noted that in a series resonant circuit, the
parasitic capacitance increases current in the coil by the amount of
circulation between itself and the coil. Or, maybe his point was that
the parasitic capacitance vitiates some of the inductance. I don`t
remember now exactly what his response was, and I`m not clear how Q
might be improved by parasitic capacitance.
I rediscovered that Terman has done all the work already and published
it on page 67 of the second edition (1937) of his "Radio Engineering".
As Reg might say: Maybe someone else did the work and Terman just copied
it. If so, Terman checked the work and found it believable.
Terman doesn`t say Q is poor when most of the capacitance is of the
parasitic variety. Terman does say: "It is to be noted that the apparent
Q of a coil is decreased by self-capacitance."
Reg Edwards
Reg Edwards asked
> How is the intrinsic Q of a Q-meter determined ?
By connecting it to an inductor of even higher, accurately known Q
?
And how is - - - - ?
Peter O. Brackett
Thanks for the detailed explanation of the progress of the thread and Tom's
and your positions. I had not been following as closely as I should have.
You did a good job of summarizing the thread.
After due consideration, heere's my comments on your controversy over Q with
Tom...
When it comes to measuring high values of Q, the most applicable "old wives
wisdom" applicable is the old cliche, "in theory, theory and practice are
the same, in practice they are different."
I do understand that the "accuracy" of a Q measurement or calculation
depends critically upon stray or secondary parameters which are often
difficult to measure and that the higher the value of Q, the more inaccurate
the calculation or measurement must be.
The "perfect" Q is infinite, or at least very very large compared to our
experience, but the closer one approaches the infinite the more difficult it
becomes to measure or calculate, and the higher the inaccuracies in its'
determination, both relative (%) and absolute!
Q suffers from that affliction common to many "difficult to determine
quantities", which is that it can only really be determined as the
difference between two large numbers which are very close in value.
Generally determining such a difference is a thing you don't want to be
doing!
Most laboratory measurements of Q are made by simple instruments by making
use of the "approximate" formula Q = f0/(f2 - f1) where fo is the resonant
frequency and f1 and f2 are the two "3dB down" frequencies on either side of
a very sharp resonant peak.
Now if Q is very large, say up to 10,000 or more and f0 is in the MHz, then
f1 and f2 are two very large numbers quite close together, and their
numerical difference will dimish in signifigance so as to frustrate the most
persistent of metrologists! Heh, heh.
In cases of very high Q's even 10 digit calculators won't be able to
determine the difference between f1 and f2, even if f1 and f2 could be
individually measured to the required accuracy.
Because of the nature of Q, i.e. perfection is immeasurable [infinite], and
as Q approaches its' perfection, the accuracy with which one can know the
precise value of Q decreases. Frustating! This is because it is
essentially a strong function [reciprocal] of the tiny difference between
two very large numbers of approximately equal value. Thus all discussions
of Q are destined for difficulties and controversy.
And so... for satisfactory resolution of these controversies, the
protagonists, [Tom and Reg?] require the most candid and detailed sharing of
both laboratory results and mathematical models.
To make progress, one must push forward into the details of both the lab
measurements and their error analysis and with the mathematical models used
for calculation.
I do understand that, because of these facts, low values of Q are easy to
calculate and measure, in terms of their relative accuracy.
And that [extremely] high values of Q are extremely difficult to measure
accurately.
For instance the Q of a quartz crystal [Which is the equivalent of a simple
LC series tank circuit.] can often exceed a value of 25,000 - 100,000 or
more! I challenge anyone here to produce for me a standard "off the shelf"
Q meter made by any manufacturer that will even give a satisfactory
"indication" of, let alone an accurate value for the Q of any quarz
crystal.
Does anyone reading this thread know who makes Q meters for measuring the Q
of quartz crystals? A very narrow field of instrument sales indeed.
And so, I agree that measuring, and indeed calculating, the values of Q,
especially for high values, is fraught with difficulties and not for the
faint of heart.
Getting agreement between measurement and calculation for high Q values is
extremely difficult and depends critically upon a full understanding of all
the things, both practical laboratory measureable things as well as
mathematical model things, involved in the measurements and calculations.
And as we know, most of those things are "secondary" or "stray", i.e. they
are things like interwinding capacitances for example, that do not appear on
anybody's schematic diagram. They are things which often no one knows are
there until there is a disagreement between theory and practice. And so
getting agreement on them is difficult.
Only when we get disagreement between theory [mathematical model] and
practice [measurements] can we refine and perfect both!
And with Q, as with life, we are ultimately destined to "fail" for as we
push forward to ever higher values of Q trying to reconcile theory and
practice we will continue to find ever larger areas of disagreement. In
life as we push forward we face death. That's exactly the end result for
the final successful determination of very high Q's. We will all be dead by
the time we can successfully measure a Q of 100 million! Ultimately the
difference between theory and practice will itself become the difference
between two very nearly equal large numbers... and once we reach that lofty
level of disagreement, well...
Reg says the Q is infinite [Oops his computer registers just overflowed!]
and Tom agrees that he also measures an infinite value on his lab Q meter,
who among us will dare...
to punch those two infinities into their lowly desk calculator and press the
difference key (that's the "-") to find the difference between those two
infinities and determine the winner... Tom or Reg....
In theory, theory and practice are the same, in practice they are
different...
We are all presented with an opportunitiy to learn new things here... if
only theory (Reg) and practice (Tom) would collaborate on this thread, if
only they would share, and be more candid and open about their mathematical
models and their laboratory practice.
If you both hold your cards to the vest then no one will win! We all have
something to learn here. Cooperate, "just do it!"
Thoughts, comments?
-Peter K1PO
"Reg Edwards" <g4fgq...@btinternet.com> wrote in message
news:970mg6$cma$1...@uranium.btinternet.com...
Richard Clark
<ab...@ix.netcom.com> wrote:
>
>For instance the Q of a quartz crystal [Which is the equivalent of a simple
>LC series tank circuit.] can often exceed a value of 25,000 - 100,000 or
>more! I challenge anyone here to produce for me a standard "off the shelf"
>Q meter made by any manufacturer that will even give a satisfactory
>"indication" of, let alone an accurate value for the Q of any quarz
>crystal.
>
Hi Peter,
You have succumbed to the old wive's tale of the Q meter posed in this
thread (off the shelf instrumentation indeed!). The Q Meters that
have passed over my calibration bench had incredible Q's by virtue of
being cavity resonators. I doubt if this specie qualifies as what
this board would consider as "off the shelf." For one it dismisses
the artificial premises here as absurd.
Your question above can resolve itself in simply exciting that
crystal and measuring its frequency spectrum. This is already
responded to by Gary in the classic determination of Q. You pine away
about 10 place measurements, and yet frequency can be measured out
many more places without too much effort (but certainly at a lot of
expense). Given that the frequency standard of the Cesium Beam is
absolute, any measurement against it will result in a beat frequency
that is easily measured. I am not going to hedge my bet with the
native frequency of the Beam against a crystal of the same frequency,
that would be a slam dunk (if you could in fact find such a crystal).
However this focuses on the obvious solution at the expense of a
repertoire of other methods that can achieve the same end. I assure
you that metrologists are not so easily discouraged at technical
challenges. This, however, hardly rises to the merit of "technical
challenge."
Industry and the marketplace care not a whit about super Q devices.
Hence the quality of "off the shelf" instrumentation is limited to
their less demanding needs. Science and Engineering is not limited to
the marketplace for solutions. The tenor of the question of how to
calibrate a Q Meter reveals a basic ignorance of measurement, or is
offered as a troll.
73's
Richard Clark, KB7QHC
Wes Stewart
<ab...@ix.netcom.com> wrote:
[snip]
>
>For instance the Q of a quartz crystal [Which is the equivalent of a simple
>LC series tank circuit.] can often exceed a value of 25,000 - 100,000 or
>more! I challenge anyone here to produce for me a standard "off the shelf"
>Q meter made by any manufacturer that will even give a satisfactory
>"indication" of, let alone an accurate value for the Q of any quarz
>crystal.
>
>Does anyone reading this thread know who makes Q meters for measuring the Q
>of quartz crystals? A very narrow field of instrument sales indeed.
Well, it didn't have a label on it reading "Q Meter", but I used an
H-P 3577 network analyzer and software I wrote to measure 1000s of
quartz crystals and calculate their motional parameters. From this, Q
is easily determined.
Saunders & Associates makes crystal measuring equipment to do this on
the fly during crystal processing. http://www.saunders-assoc.com/
Wes Stewart N7WS
>
[snip]
Roy Lewallen
>. . .
> Now if Q is very large, say up to 10,000 or more and f0 is in the MHz, then
> f1 and f2 are two very large numbers quite close together, and their
> numerical difference will dimish in signifigance so as to frustrate the most
> persistent of metrologists! Heh, heh.
>
> In cases of very high Q's even 10 digit calculators won't be able to
> determine the difference between f1 and f2, even if f1 and f2 could be
> individually measured to the required accuracy.
It's not at all difficult to resolve a 1.4 kHz difference between two
frequencies at 14 MHz. Your receiver does it without strain, as does
even a very inexpensive frequency counter. In any case, high accuracy
isn't required, just moderate resolution.
>. . .
> For instance the Q of a quartz crystal [Which is the equivalent of a simple
> LC series tank circuit.] can often exceed a value of 25,000 - 100,000 or
> more! I challenge anyone here to produce for me a standard "off the shelf"
> Q meter made by any manufacturer that will even give a satisfactory
> "indication" of, let alone an accurate value for the Q of any quarz
> crystal.
Somewhere around 8-10 years ago Wes Hayward, W7ZOI, wrote some excellent
articles on designing crystal ladder filters. I believe one was in QST
and one in QEX. Some filters require selecting crystals, and in the
article was a simple circuit for doing that, which anyone could build at
home. I'm quite sure it could be used to get the crystal Q with
reasonable accuracy. Perhaps someone with easy access to the article can
confirm that.
> . . .
> In theory, theory and practice are the same, in practice they are
> different...
If theory and practice are different, the theory is faulty and should be
revised or replaced.
>. . .
Roy Lewallen, W7EL
Cecil
> If theory and practice are different, the theory is faulty and should be
> revised or replaced.
Depends upon whether one subscribes to a primacy of existence concept
or a primacy of consciousness concept - Aristotle VS Plato. Some people
here on r.r.a.a believe that the math model *IS* the reality.
Reg Edwards
come up with any solid, numerical, quantitative, support in favour
of his assertions. Which he has dropped anyway. So let's forget
about it !
----
Tom W8JI
<g4fgq...@btinternet.com> wrote:
>There's no controversy between me and Tom. Its just that he hasn't
>come up with any solid, numerical, quantitative, support in favour
>of his assertions. Which he has dropped anyway. So let's forget
>about it !
I haven't modified or dropped my assertions, it just isn't possible to
discuss anything with you.
Your programs show Q increasing all the way up to resonance, my test
equipment does not.
73 Tom
Peter O. Brackett
Peter O. Brackett
Aristotle vs Plato... What?
Wow!
It has taken me this long to discover, from your pithy postings, that
indeed, they are civilized at TX A & M. Long live the "Aggies".
Now if only we could get Tom and Reg to be so philosophical.
Tom, what is the name rank and serial number of YOUR Q-Meter?
Just askin?
-Peter K1PO
"Cecil" <Cecil....@IEEE.org> wrote in message
news:3A95659F...@IEEE.org...
Peter O. Brackett
[SNIP]
So let's forget
> about it !
> ----
> ***********************************
> Regards, Reg, G4FGQ
> Free radio design & modelling software
> http://www.btinternet.com/~g4fgq.regp
>
> ***********************************
[SNIP]
Fugetaboutid! What kind of "cop out" is that?
Reg, you can't get off so easy, and neither can Tom. Come back here and
fight it out!
Theory versus Practice, Mathematics versus Laboratory, Software versus
Hardware, Aristotle versus Plato, Ali versus Liston...
We want it now!
Reg what is the mathematical model that you are using for Q, and Tom what is
the manufacturer, model number, serial number and last calibration date of
your Q-Meter?
Best,
-Peter K1PO
Peter O. Brackett
been impressed by HP Instrumentation, but now, thanks to Carly, I don't know
what to expect... We now have about 50 years to establish the veracity of
Aigilent Instruments I guess. -Peter K1PO
"Wes Stewart " <nospa...@arrl.net> wrote in message
news:3a955a0a...@news.aepnet.org...
Peter O. Brackett
Talk about "Technology Jocks", hey... give us a break.
Wes Haywood is a God!
What about the rest of us mere mortals. How shall we measure Q?
What is the Q-Meter that Tom W8JI uses in his daily Q measurements?
Boonton, model x, serial no. y, last calibrated by ??? on 12/21/41? What?
-Peter K1PO
"Roy Lewallen" <w7...@eznec.com> wrote in message
news:3A9561DC...@eznec.com...
Roy Lewallen
included a simple circuit that any mere mortal, presumably even you and
I, can build and use to measure crystal Q.
Roy Lewallen, W7EL
"Peter O. Brackett" wrote:
>
> Roy:
>
> Talk about "Technology Jocks", hey... give us a break.
>
> Wes Haywood is a God!
>
> What about the rest of us mere mortals. How shall we measure Q?
> . . .
Cecil
> It has taken me this long to discover, from your pithy postings, that
> indeed, they are civilized at TX A & M. Long live the "Aggies".
My ex-wife says the only class I'll ever have is the one I took
at Texas A&M.
Reg Edwards
practical use. It is just a ratio. It may be used to calculate or
estimate some other circuit parameter. In the case of quartz
crystals the other circuit parameter is usually the series loss
resistance which, in any case, can be more easily measured
directly.
An accurate way of determining coil Q, if anyone should insist on
knowing the value of Q, is to measure the coil's loss resistance at
resonance using an impedance bridge - then CALCULATE it !
Why am I reminded of the method of determining ground conductivity
by setting up a high power radio transmitter and measuring
ground-wave field strength at various distances up to several
hundred miles ? Having tediously determined ground conductivity it
is then possible to predict field strength at various distances
from a transmitter of known power output.
But it does keep the philosophical conversation going ;o)
Reg Edwards
Joke, joke.
I do not have a Q-meter. Like SWR and S-meters I don't trust them.
I have my own, more reliable methods of determining Q.
But in any case I (and very few others) seldom need to know the
value of Q. It's not a particularly useful quantity. As I have said
elsewhere, Q is just a ratio, an intermediate value in the
calculation of some other parameter of more practical interest. It
also provides a topic of conversation.
My relatively recent interest in Q arose only because I wanted to
check the accuracy of some software. Before throwing my SOLNOID2
program to the wolves I wound 20 or 30 coils of various dimensions
to check what you would refer to as 'algorithms'. The algorithms
were developed from basic engineering and mathematical
rinciples - the tools of any electrical/mechanical engineer worthy
of the name. On these coils, 1 inch to 5 feet long, 1 inch to 6
feet diameter, I determined, where feasible, inductance,
self-resonant frequency and that nuisance parameter Q. Having
served their purpose I tossed the scribbled measurement results
into the garbage bin.
So I do not have a calibration certificate for a non-existent Q
meter. But having during my career set up from an empty floor an
electricaI standards and calibration laboratory on an echelon
imediately beneath the British National Physical Laboratory
(equivalent to NBS), I am familiar with assessment of statistical
measurements uncertainties and the need for traceabilty to national
and international standards.
----
***********************************
Regards, Reg, G4FGQ
Free radio design & modelling software
http://www.btinternet.com/~g4fgq.regp
***********************************
Peter O. Brackett <ab...@ix.netcom.com> wrote in message
news:974f2r$kn$1...@slb5.atl.mindspring.net...
Richard Clark
<g4fgq...@btinternet.com> wrote:
>So I do not have a calibration certificate for a non-existent Q
>meter. But having during my career set up from an empty floor an
>electricaI standards and calibration laboratory on an echelon
>imediately beneath the British National Physical Laboratory
>(equivalent to NBS), I am familiar with assessment of statistical
>measurements uncertainties and the need for traceabilty to national
>and international standards.
Ah, Yes!
Then that explains it, not necessarily a lack of knowledge so much as
a troll.
...as if there were any doubt about Punchinello :->
Wes Stewart
A few weeks ago, while taking a short "reading break" from a DX
pileup, I was flipping through one of the electronics trade magazines
and browsed an article.
The premise, if I remember correctly was that turn-to-turn capacitance
in an inductor was more-or-less a myth because, and this is fuzzy,
there wasn't a significant difference in signal phase and amplitude
between adjacent turns and self capacitance was really just
capacitance between the inductor and the measuring instrument.
I said to myself, "Self, that is interesting you should study this
more later." Unfortunately, the library was rearranged before I
remembered to look this up and now I can't remember where I saw it.
Any comments?
Wes Stewart, N7WS
Roy Lewallen
that inductors couldn't be self resonant. But they are.
Roy Lewallen, W7EL
Richard Clark
wrote:
Hi Wes,
Comments follow.
Strange observation indeed, and probably argued from a perspective
where this issue is marginal at best (in other words they don't worry
about it because it isn't a problem at their scale of operation). I
say this because of their provisos offered insofar as signal phase and
such (rather an exotic consideration that does not exist at all
scales).
There are clear and simple solutions working in a billion dollars
worth of gear. In my own 6M rig (a Motorola mobile converted from
public service) are a string of at least 6 shielded inductors that
each show considerably more Q than if they were simply sitting on an
open PC board (or in free space for that matter). There is a better,
technical name for them, but I am not close to my resource to
resurrect that term right now. The point of the matter is that all
capacitance is overwhelmed by the surrounding shield, but that shield
is made part of the resonant circuit and would certainly swamp the
inter-turn capacitance.
In the Nav, I taught a course on what was then called the TED and the
RED (UHF transmitter and Receiver). One of the amusing tuned circuits
was a simple metal post in the middle of a cavity (no, the cavity was
not resonant in the conventional expectation of modes excited by the
physical dimensions - more probably another example of my 6M rig's
design). No turns at all, nor any capacitors. Knowing the specs
demanded by the Nav, I would say that the Q was very good.
John S. Dyson
"Richard Clark" <rwc...@seanet.com> wrote in message news:3a97c8c6....@news.seanet.com...
> scales).
>
> There are clear and simple solutions working in a billion dollars
> worth of gear. In my own 6M rig (a Motorola mobile converted from
> public service) are a string of at least 6 shielded inductors that
> each show considerably more Q than if they were simply sitting on an
> open PC board (or in free space for that matter). There is a better,
> technical name for them, but I am not close to my resource to
> resurrect that term right now. The point of the matter is that all
> capacitance is overwhelmed by the surrounding shield, but that shield
> is made part of the resonant circuit and would certainly swamp the
> inter-turn capacitance.
>
John
Gary Peach
"Peter O. Brackett" wrote ..
> Roy:
> What about the rest of us mere mortals. How shall we measure Q?
By whether you get into the cinema or not before the big picture starts.
--
Gary7SLL
51° 23' 50" N 1° 19' 58"W
http://www.gary.peach.dial.pipex.com/
Richard Clark
wrote:
Hi John,
Thank you, yes, helical resonators. My reference is "Reference Data
for Radio Engineers," Chapter. 22 Transmission lines.
If you wanted to optimize the Q of the coil (this is largely the only
structure if you don't count the walls) there are nomographs for any
consideration.
I will simply use the already illustrated exercise found in the book.
We have a 30 turn inductor for resonance at 10 MHz. The unloaded Q is
the desired engineering goal (or marketing goal, what have you...) and
for this example is idling at 1000.
To fulfill the expectations of the preceding paragraph we come up with
that we need a 30 turn 3 inch diameter coil inside a 6 inch diameter
can. To obtain the unloaded Q of 1000 the turns pitch needs to be
0.16 inch.
The geometry of this clearly reveals that the turn to turn capacitance
is much more dominant (by nearly 10:1). However:
"It operates as a distributed-parameter system equivalent to a
quarter wave co axial transmission line resonator."
This quote becomes obvious when the coil unwinds to maximize Q with a
straight wire in a coaxial, quarter wave length cavity.
Let's examine that defining tenet recently expressed:
>Its an old wives' tale that Q is poor when most of the capacitance is of the
>parasitic variety.
A straight wire is nature's way of minimizing parasitic capacitance.
Peter O. Brackett
What is the Q of a queue? Or indeed of a cue? MI5 had a "Q", but got no
respect from Bond. Does that make him a low Q?
-Peter K1PO
"Gary Peach" <xc...@dial.pipex.com> wrote in message
news:978nm2$n5e$4...@lure.pipex.net...
Gary Peach
"Peter O. Brackett" wrote ...
> > > What about the rest of us mere mortals. How shall we measure Q?
> > By whether you get into the cinema or not before the big picture starts.
> What is the Q of a queue? Or indeed of a cue?
Oh! balls can't get out from the 8
> MI5 had a "Q", but got no respect from Bond.
> Does that make him a low Q?
No, a fur cue.
For the answer go down to Kew at Lilac Time, it isn't far from London.
While you're there pop into the PRO and look up Bond.
Is an educated agent a Unibond?
--
Gary7SLL
51° 23' 50" N 1° 19' 58"W
http://www.gary.peach.dial.pipex.com/
For our American Cousins Unibond is a product line of building materials
mostly paints and plasters and fillers and things; hope this helps you get a
grip on this subject.
Oops! America doesn't have Subjects. :)
Reg Edwards
relative contributions depend on the number of turns.
The strays can be lumped together to form an equivalent capacitance
in parallel with the coil's inductance at the natural self-resonant
frequency. Depending on an isolated coil's length/diameter ratio,
the strays can also be considered to be uniformly distributed along
the length of the coil which then behaves as 1/2-wave transmission
line at its self resonant frequency.
The principal component can crudely be decribed as the capacitance
between the two half-cylinders when the coil is cut at its center.
The electrostic field spreads out into the surrounding space in a
manner similar to the magnetic lines of force around a bar magnet.
The other component is due to capacitance between turns. With a
single turn, as in a magloop, the between-turn capacitance is
obviously zero. With two turns the capacitance in shunt with the
the coil is a maximum. The two components are one and the same.
But with N turns the equivalent parallel capacitance is only 1/Nth
of the capacitance between adjacent turns because the individual
capcitances are all in series with each other. So as coil length
and N increase, the first component rapidly predominates.
In any case, if the insulating material between turns is of high
quality, even with only two turns, turn-to-turn capacitance will
not degrade Q any more than any other sort of circuit component.
When a helical resonator is enclosed in a screen it increases
capacitance somewhat and distributes it more uniformly along its
length. But the screen does NOT increase coil Q - it merely
maintains the natural high Q of the coil it by preventing the
external field from interacting with poor uncontrolled dielectric
materials and lossy conductors in the immediate vicinity.
It is the interaction of a coil with its environment which makes
the working Q of a tuned circuit less than that which may be
calculated on the assumption that the coil is isolated in space. Q
meters are part of the environment !
Richard Harrison
"Q meters are part of the environment."
My Heathkit HD-1250 dip meter will indicate the self-resonance of the
1.6 to 3.4 MHz coil by using the 3.2 to 6.6 MHz coil in the dip meter.
The dip occurs at the 6.6 MHz end of the dial.
The inductance of the self-resonant coil is given by Heath as 171
microhenrys.
To resonate at 6.6 MHz requires 3.4 picofarads by my ARRL Lightning Coil
Calculator. That is the apparent distributed capacitance of my 171
microhenry coil. Looks reasonable to me.
Heath recommends using the plug-in coils this way to determine unknown
capacitor values. One can start with the coil`s parasitic capacitance
value. It seems to work.
Best regards, Richard Harrison, KB5WZI
Peter O. Brackett
We colonialists were once British Subjects, but our ancestors decided the
King didn't really know what was best for us, and so we embarked upon an
effort to increase our own Q.
Unibond...heh, heh... we don't have that trade name here in the colonies,
but since you have "splained" it for us, now we "get it".
Kew? Is that where Kew Gardens is located? Is this the Q that Reg
disparages?
Best,
-Peter K1PO, a former Subject.
"Gary Peach" <xc...@dial.pipex.com> wrote in message
news:97akm0$ju3$3...@lure.pipex.net...
Gary Peach
"Peter O. Brackett" wrote ...
> We colonialists were once British Subjects, but our ancestors decided the
> King didn't really know what was best for us, and so we embarked upon an
> effort to increase our own Q.
O. "Stanley that's another fine mess you got me into."
fx(screws up face pretending to cry and scratches top of head)
Yeah! your best home grown was the Marx Brothers:)
And of course Doris Day
> Unibond...heh, heh... we don't have that trade name here in the colonies,
> but since you have "splained" it for us, now we "get it".á
It'll cover up the cracks very nicely.
Wouldn't want you to think that it was come sort of college fraternity,
phi, beta, kappa.
On the square, more likely Toc H.
I hope that your builders are better than the Cowboys that we get here.
Just how much DIY do you guys do on your own houses?
> Kew? Is that where Kew Gardens is located?
Yes, certainly is:
Go down to Kew at lilac time,
It isn't far from London.
> Is this the Q that Reg disparages?
We take what Reg disparages with a pinch of salt.
--
Gary7SLL
51° 23' 50" N 1° 19' 58"W
www.gary.peach.dial.pipex.com/
Reg Edwards
>
> We take what Reg disparages with a pinch of salt.
It is not I who disparages Q.
It is others who exalt it, extol it, worship it, even adore and pay
homage to it, and place it on a pedestal with SWR and computer
programs alongside the Angels.
But everything has a use. Is a widget the unfledged offspring of a
widgeon ?
----
Reg, G4FGQ
Reg Edwards
unknown
> capacitor values. One can start with the coil`s parasitic
capacitance
> value. It seems to work.
Indeed it does. Although when connections are made to a coil its
self-capacitance re-distributes itself and changes somewhat, the
value of an unknown capacitor placed across the coil can be
determined from the inductance and the new resonant frequency
followed by subtracting the value of the stray capacitance. It's
accurate enough.
But there's no need to become involved with the stray capacitance
at all. All that's needed are the two resonant frequencies and the
LF inductance.
In the 21st Century, instead of an 1/2-hour +/- 15% job, the
value of a capacitor can be determined by your hand-held analyser
within 15 seconds with an accuracy of a few percent. But I cannot
say I'm any happier for all these mod-cons.
----
Reg
Peter O. Brackett
Heh, heh. I know a widget when I see one, but....
I've never known a widgeon... what?
Q is so ephemeral... just when you get it right, it disappears right over
the horizon into infinity!
Reg, have you ever used commercial pot cores and wound precision inductors
from the manufacturer's "iso-Q" contour charts?
I have often wondered if there is a similar iso-Q chart that one can create
for the solenoid inductor.
Iso-Q charts correspond to a given pot core design, which has a fixed
physical dimension and a fixed magnetic air gap resulting in a fixed value
of Al, the charts have inductance on the ordinate and frequency on the
abscissa and consists of closed "loops" tracing the "curves" of constant Q.
It is actually a graphical representation of a surface of Q values over the
co-ordinates of L and f. To be a valid inductor the pot core has to be
"filled" with wire. i.e. once the number of turns is chosen using the Al
value to fix a given L, then the size of the wire must be chosen to fill the
pot core cavity, and then the resulting Q can be read from the iso-Q
contours.
Reg, if one were to fix a given set of solenoid dimensions, i.e. say a
length l and a diameter d could one arrive at a set of iso-Q contours over a
plane of L versus f for a solenoid? Does this make sense? Can your program
plot such a set of curves?
-Peter K1PO
"Reg Edwards" <g4fgq...@btinternet.com> wrote in message
news:97o3pl$pki$1...@neptunium.btinternet.com...
Richard Harrison
"I`ve never known a widgeon...what?
In the U.S. it is a small civilian flying boat very popular with oil
companies in the bayou country of Louisiana.
Peter O. Brackett
Thanks for that "widgeon" information. I've never heard that term before,
of course I've encountered lots of "widgets", both low Q and high Q widgets.
The high Q widgets are the best!
Incidently, I've been to Lafayette, LA a few times. Went to a Doug Kershaw,
Jimmy C. Newman concert, and danced with those Cajun girls... and I do have
a few Cajun friends from there, but never saw those flying boats to which
you refer.
In bayou country, I just saw and rode in the airboats, like the ones they
use in the Florida everglades. Been on those swamp airboats in both places,
LA and FL. Fun but noisey.
On small flying boats, I've never been on one myself, but I have seen em in
action. They are really neat...
Chalk Airlines used to fly a wing of small flying boats out of Miami, FL
harbor over to the Bahamas on a regular schedule. Don't know if they still
do that. I never flew on them, but used to watch them taking off and
landing alongside the ocean going vessels in the harbor back when I lived in
Miami.
Would those Chalk flying boats have been widgeons?
ThanQ
-Peter K1PO
"Richard Harrison" <richard...@webtv.net> wrote in message
news:21763-3A...@storefull-116.iap.bryant.webtv.net...
Richard Harrison
"Thanks for the "widgeon" information."
A widgeon is a small fresh-water duck between the mallard and teal in
size. Grumman named several civilian aircraft after water-fowl.
Chalk Airlines would probably use larger aircraft unless they were doing
charter service. The widgeon only can carry a handful of people.
When I lived in Portugal, I watched a start-up airline unpack and
assemble several U.S. surplus PBY flying boats (Catalinas). These are
larger than widgeons. They were to be outfitted with a few rows of seats
and other passenger amenities. Unfortunately, their fledgling pilots
crashed them all before they ever went into regular service. This is no
reflection on TAP the government airline. They have one of the best
safety records in the business. No reflection on the PBY`s either. They
are superb aircraft.
When Peter visited Lafayette, had he poked arund the hangars there, he
probably would have seen several widgeons. At least Texaco and Union Oil
stationed widgeons there.
Take-off from a runway is unremarkable. They are amphibians. Take-off
from the water is something else. At one point the props pick up floods
of water and inundate the cockpit. The pilot can`t see a thing. You just
hope there isn`t a log or big gator lurking in the take-off path.
Reg Edwards
Being a practical engineer I've never found the iso-Q contour
charts for ferrite pot cores of any use. The operating instructions
for bobbin winding machine have always been the most essential
documents.
I suppose it would be possible to use a solenoid computing program
to assist a team of 4 people to produce iso-Q charts for coils. It
would take about 5 years and a lot of guesswork by which time any
need for them would have completely evaporated.
The present need must be quite small. After all, we've managed very
well without solenoid iso-Q charts for the last 100 years.
Regards, Reg, G4FGQ
***********************************
Peter O. Brackett <ab...@ix.netcom.com> wrote in message
news:97pi2a$fp1$1...@slb1.atl.mindspring.net...
Peter O. Brackett
Iso-Q curves...
Well... back in the "old days" in one of my former electronic company
employers, when we were designing and selling a lot of "precision" LC
filters and LC group delay equalizers, we made extensive use of pot cores
and pot core iso-Q curves.
Most of those filters and all-pass group delay equalizers that we designed
were for use in the frequency range from audio up to a couple of MHz and
were never in the HF band.
In many cases they were "narrow band", i.e. say spanned a 3.4kHz voice
channel say in the "group" band of 60kHz to 104kHz, or VFCT, voice frequency
carrier telegraph channels, etc...
These were "precision" filters and equalizers, often of very high order.
The highest order that I ever personally designed and put into production
was of order 22. i.e. a precise design comprising 11 inductors and 11
capacitors, not counting Norton transformations and tapped inductors, etc...
Now in order to get the actual production filters to "hit" the design spec
within the desired tolerances [Which were often of the order of +-0.1dB in
the passbands], we had to actually synthesize "pre-distorted" designs.
We had to pre-distort the inductors for two effects, both the Q and the
stray capacitance of the coil, or self-resonant frequency of the coil. In
practice most of our capacitors could be considered to be "ideal" and
consequently we seldom pre-distorted a design for the C's. We did this only
for the L's. This is/was called "semi-uniform" pre-distortion.
To do this predistortion accurately we had to know the inductor "Q"s in the
filter passbands fairly precisely, we used the iso-Q contours to help us
with this. We also had to pre-distort the design for the "effective or
apparent" inductance. i.e. the actual apparent inductance in the filter
passband taking into account the stray intra and inter winding capacitance
and the self resonant effect of the inductor. In other words we actually
had to offset the design inductance by an amount to take into account the
capacitive reactance of the inductor stray capacitance, at least across the
band of interest. Our precise value of design inductance was the apparent
inductance at a given frequency of the structure of the inductance and stray
capacitance acting together. Below resonance, this apparent value of L,
call it Lapp, is always slightly higher than the actual inductance. The
deviation of Lapp from L rising as the self-resonant frequency is
approached.
The pot core manufacturer's iso-Q contours and their "effective inductance"
correction factors allowed us to design and manufacture inductors which were
then used in the "pre-distorted" precision designs with great accuracy. The
iso-Q contours were actually plotted on a plane of actual or real frequency
f versus apparent inductance Lapp.
We used computer analysis and synthesis to generate the "pre-distorted"
filter and/or equalizer designs. i.e. we would begin with a specific ideal
"target" design, say for instance a Cauer Parameter [Elliptic Function
Filter] filter of order12 or more... then we would first design the ideal
filter to get the first pass non-predistorted values for L's and C's... then
we would design, on paper, the "ideal" inductors required. Then we would
find the actual practical Q for those inductors from the iso-Q contours
provided by the pot core manufacturer, within our passband of interest.
Then we would average the value of those inductor Q's to get an average coil
Q which we called Qavg for the whole filter. Then we would take the
mathematical "ideal poles" of the ideal filter and pre-distort or shift them
to the right [towards the jw axis] by an amount equal to 1/2*Qavg. Then we
would use our computer algorithms to re-synthesize the whole filter from
those "pre-distorted poles". We generally did not pre-distort the stopband
zeros, since the "error" in the notches in the stopbands was tolerable, it
was the passband errors that really caused the difficulties. And we would
then design and manufacture the actual filter using the new, nearby, values
of inductors obtained from the synthesis of the pre-distorted design.
And... presto... when the filter was built and put in production using the
actual coils [rather than the mathematical ideal coils] the actual filter
would be "right on the money". Note the capacitor values would also now be
slightly different as well.
Thus we could design filters that used practical finite Q coils, including
the effect of self-resonance, to manufacture very nearly exactly perfect
precison filters and equalizers. Our competitors often tried to achieve the
same thing by adding additional correction amplitude and delay equalizers to
a poor non-predistorted design. i.e. since they did not know how or could
not pre-distort the design, they just took the poor design and added a bunch
more of inductors and capacitors to correct the errors in the design.
Consequently our precision designs cost usually 20 - 30% less than theirs
and were much more precise, with fewer parts and smaller altogether.
Pre-distortion for all-pass group delay equalizers was a little more
complicated, since the group delay eqalizers have both poles and zeros that
need to be pre-distorted. No one seems to care about this passive LC
all-pass equalizer stuff anymore today folks do delay equalization digitally
today, but we did some really neat stuff in that passive LC all-pass network
area! I believe that we were the only firm that did pre-distorted delay
equalizer designs.
Over a period of about 10 years we created perhaps 50 different passive LC
filter and group delay equalizers, and produced them in many thousands of
units for use in all kinds of commercial FDM equipment.
And so... I don't know about others, but we did find iso-Q contours to be
both useful, and profitable.
Now I agree that the average ham or person doing low order, low precision,
designs in the HF band would have no use for iso-Q contours, however...
those curves do illustrate some interesting and practical physical
phenomena, and help with the intuitive understanding of the underlying
physical phenomena, for instance...
The fact that the iso-Q contours for all pot cores seem to be closed curves
in the (Lapp, f ) plane is interesting.
Why are the curves all closed? What can we learn from that?
And.. the fact that there seems to be only one particular (Lapp, f)
co-ordinate where a particular physical dimension of pot core inductor has
maximum Q... why is that? i.e.
the Q surface over the (Lapp, f) plane is mono-modal. Why? Why does the
maximum occur where it does? etc...
and so... getting back to the thread topic, I have often wondered for a
solenoid single layer inductor... of length l and diameter d...
what is the maximum value of Q?
....and for what value of Lapp and f does it occur?
etc...
is there in fact a unique single maximum value of Q for a solenoid on the
(Lapp,f) plane?
For instance, can your program be used to plot the maximum values of Q for a
solenoid of dimensions l and d over the range of 1 MHz to 30 MHz, and the
range of 1uHy to 100uHy... I presume that the wire diameter at each point
would be set by the program to maximize the Q at any point etc....
Then once the first set of iso-Qs for a given (l, d) are plotted, one could
go on to plot curves for various (l, d) values to perhaps eventually
discover the question which we all desparately want answered!
What is the solenoid coil which has the highest Q value in the HF band?
i.e. does there exists a solenoid of given l and d, wound with copper wire
and of apparent inductance Lapp, at frequency f which has the maximum value
of Q of all possible inductors? If so, what is it? Can we not all use it?
Let us designate this optimum solenoid HF inductor the "closest inductor to
the perfect inductor" and we will name it the "Henry" or something like
that!
Then, for it's discovery, I will personally offer the "Brackett" prize! A
free dinner at KFC!
How about it? Your program "solenoid" is a good start!
Best,
-Peter K1PO
"Reg Edwards" <g4fgq...@btinternet.com> wrote in message
news:97u5t0$dvf$1...@neptunium.btinternet.com...
Richard Harrison
"In many cases they were "narrow band", i.e. say spanned a 3.4kHz voice
channel say in the "group" band of 60kHz to 104kHz, or VFCT, voice
frequency carrier telegraph channels, etc."
OK, except the usual frequency division multiplex allows 4kHz per voice
channel to permit signaling and supervision at the edge of the voice
channel. The bandwidth retained for the voice band might vary slightly
from 3.4kHz.
The "base group" is normally 12 of these 4-kHz channels between 60 and
108 kHz. These same 12 channels (the base group) are used over and over
again throughout the baseband by translation. They are simply
heterodyned either directly or via some super-group to any group-spot in
the baseband.
The 12-channel group occupying 108 to 156 kHz is exactly the same
12-channel group as that occupying the 60 to 108 kHz spot. It has just
been translated.
Modern analog practice is to use a universal channel which can be
programmed for any carrier frequency multiple of 4 kHz within the
baseband. These are SSB channels. Only a single modem is needed as a
replacement for any channel modem in the system.
As Peter implied, it is desirable to have the same delay for every
frequency in the baseband.
Filter flanking is a big deal to fit all the channels together.
Peter also wrote:
"Below resonance, this apparent value of L, call it Lapp, is always
higher than the actual inductance."
In my experience, Lapp was a manufacturer of high-power vacuum tubes,
vacuum capacitors, and ceramic insulating hardware.
Peter O. Brackett
I note that you too messed around with FDM systems. I used to do a lot of
passive LC filter and equalizer designs but those skills and knowledge are
now un-marketable.
In recent years I have set my objectives toward a similar task, but using
more modern technology (DSP), and am now committed to putting more
bps/Hz/Dollar down long runs of twisted pairs of copper.
Next we are looking at doing the same for single strands of Silica. Sort of
like Kestrel Solutions and Centerpoint are doing.
Electrical waves, light waves....
Some things never change!
Best,
-Peter K1PO
"Richard Harrison" <richard...@webtv.net> wrote in message
news:14981-3A...@storefull-116.iap.bryant.webtv.net...