EMRFD Message Archive 135
Message Date From Subject 135 2006-09-10 23:52:31 Steven S. Coles 7 MHz Oscillator Capacitors Anyone,
I'm testing the 40-meter transmitter oscillator (Fig 1.39) with
ceramic capacitors. Do I remember correctly that I should change the
390 pF capacitors to silver mica type (for more stabile
characteristics & lower dissipation factor) before putting the
transmitter on the air? Did I miss the discussion of capacitor types
in EMRFD?
Best regards,
Steven, KD7YTE136 2006-09-11 10:04:59 Wes Hayward Re: 7 MHz Oscillator Capacitors Hi Steven, et al,
Ceramic capacitors are just fine in this application. I probably
put down Mica on the list because that is what I used in the version
I built. I used what I had in the junk box. That was an old time
DeMaw and League rule, and is generally a good one. But either
ceramic or mica would be OK.
There may not be a discussion of "capacitor types" in EMRFD. The
purpose of the book was not to be an encyclopedia. The ARRL
Handbook is aimed at that. Rather, EMRFD was aimed at presenting
examples of some facets of the design process and to encourage folks
to experiment. You may have to (1) look in additional places for
some answers and (2) do some experiments along the way to answer
some of the questions that arise.
Enjoy the experimental process. It can be great fun.
73, Wes
w7zoi137 2006-09-11 10:37:54 Steven S. Coles Re: 7 MHz Oscillator Capacitors Wes,
Thank you very much for the reply.
I've been a little shy of ceramic capacitors since a few years ago
when some from a hobby store drifted in a 3-watt application. The
hobby ceramic capacitors had measured dissipation factors > 0.015.
Ceramic capacitors from American Technical Ceramics and silver mica
capacitors were solid in the same circuit. The hobby grade capacitors
produced a measurable temperature rise. Of course that was a very
different circumstance.
Best regards,
Steven138 2006-09-11 13:18:58 Stan Re: 7 MHz Oscillator Capacitors Steve, I prefer polystrene caps in VFO circuits. However, they
are becoming more difficult to locate. Like you I would like to see
a good discussion about the various types we can purchase from
Mouser or DigiKey. One has to watch the tolerance on even those
market NPO/COG. Most of us do not have the proper test equipment
to really measure the dissipati139 2006-09-11 23:26:55 Dave Re: 7 MHz Oscillator Capacitors If you do a Google search for "Type I" "Type II" "Type III" and
"ceramic capacitor" you will ultimately find a mil document that
covers the subject in very nice detail.
The Type I capacitors are generally the ones we want to use. They run
from negative temp coeff Nxxx to NP0 to the positive temp coefficieint
Pxxx series. NP0 is as close to zero as can be made, at +/-30 ppm per
degree C...about equal to a silver mica, and roughly the same Q as a
silver mica, but usually lower voltage rating. Somewhat lower loss
(higher Q) is found in the Pxxx series caps, but they do have a
positive temperature coefficient. American Technical Ceramics offers
a line of RF current-rated P90 and P100 capacitors for transmitting
service. As you go negative to N470, N750 and further, you find the
capacitors physically smaller for the same capacitance value, and
somewhat more lossy. Those of us designing analog LC oscillators
often find an N470 or N750 capacitor in parallel with about an equal
NP0 can compensate for the temperature drift of the typical inductor.
Note that C0G is the same as NP0. Mouser generally has NP0 in stock,
and can order N470/N750, but the Pxx series tends to be a specialty item.
Type I caps are otherwise well-behaved; capacitance does not change
with voltage, for instance.
X7Rs are sometimes classed as Type I and sometimes Type II. Whatever,
they are smaller for the same capacitance value, and higher loss, and
have a non-linear temperature coefficient.
Z5U and many other grades have capacitance values that vary with
applied voltage and can even be microphonic. I reserve them for use
as coupling caps - their high losses actually make them poor bypass caps.
The Mouser catalog will tell you if it's NP0, etc, and often indicates
Type I or Type II.
A good general rule is to stick with NP0 (aka C0G), unless you know
you need a specific temp coefficient, or if you're looking for a
large-value (e.g. 0.1uF or larger) cap and it's just not available as
an NP0.
I was at the bench tonight, picking out caps for some audio LC
networks. I have a cheap 3-digit capacitance meter (a $50 ebay item)
and was measuring the C value. I could tell the "rotten" caps - their
capacitance would drift right after I connected them, as they cooled
off from the minimal heat of my fingers! That's part of the battle
with "in stock" caps that I have - the grade is often not marked on them.
Hope this is useful...
73,
Dave W8NF
142 2006-09-12 21:21:13 Wes Hayward Re: 7 MHz Oscillator Capacitors Hi Dave, et al,
Much of this data was abstracted from the January 1983 Tektronix
common parts catalog, the capacitor and resistor volume.
As you noted, NP0 ("negative, positive, zero) is essentially
identical to the C0 designator. But there are a variety of
temperature characteristics within the C0 flag, each represented by
a letter. The letter designates the tolerance on the temperature
coefficient. G is the modifier standing for +/-30 ppm/deg C.
Hence, a COG capacitor will have a temperature coefficient of 0
ppm/deg C, with a +/-30 error. The C0G part can be anywhere
between -30 and +30 parts per million per degree C and still be C0G.
Some of the codes with their tolerance limits are:
R +/-10 ppm/deg C
F +/-15
Z +/-20
G +/-30
H +/-60
W +/-100
S +/-150
The various positive and negative temperature characteristic curves
have designators of their own. C0 is the "gold" standard already
mentioned. U2 is the designator for negative 750 ppm/deg C, or
N750. So a U2S part would be a TC of -750 ppm/deg C, +/-150
ppm/deg C. Anything with a TC from -600 to -900 ppm/deg C would
fit in this slot.
This is more detail than we normally need or even want. The
important thing to remember is that it is not enough to know the
temperature coefficient on a standard part, but you also need to
know the tolerance in that TC. The system outlined applies
predominantly to ceramic capacitors where the TC is ultimately
controlled by the chemistry behind the process. We have not even
mentioned the grossly unstable capacitor types that we use for
blocking and, sometimes, bypass capacitors.
There is an uncertainty related to our inductors. We may wind them
on a -6 toroid core where the manufacturer tells us that the TC is
+35 ppm/deg-C. Seeing the nature of capacitors, we can't help but
wonder if there may not be a similar spread related to the
toroids. Many of us have observed much worse temperature
characteristics when the wire on a toroid is not wound firmly
against the core. The bottom line conclusion that I've reached is
that the only real answer to building a really stable oscillator is
to put the circuit in a thermal chamber, measure the behavior with
temperature, and then adjust the compensation. Having done it, I
was amazed at just how simple (and even cheap) it is to do. Yet
it is virtually impossible to get folks to do the experiment. So
much for that soap box.
One capacitor type of interest is the polystyrene unit mentioned in
another posting. I've found them to be useful parts with a
negative TC of -150 ppm/degree C. The typical error is +/-50
ppm/degree C. I have used them for two applications. One is for
places where a reasonable capacitor is needed, but where extreme
temperature stability or knowledge thereof is unimportant. The
other place where they have been very useful is as a negative TC
part for use in temperature compensation.
Mica capacitors are historically interesting, for they were
capacitors for use in the HF region that had very high Q. As such,
they handle a lot of current without much heating. This can be
very important in an oscillator, for it is a circuit where there can
be very large circulating current, even when the DC power into the
circuit is very small. Some mica capacitors were very stable, and
others were very poor. Here are some of the mica cap designators
pulled from the old Tek book:
B +/-500 ppm/degC
C +/-200 ppm/degC
D +/-100 ppm/degC
E -20 +100 ppm/degC
F 0 to +70 ppm/deg C
The range depicted in this chart will explain some of the mixed
results we have obtained with silver mica capacitors in the past.
Some are great while others are poor at best.
Not all NP0 capacitors are equal. I got used to some of the great
caps that we could purchase from the Tektronix company store or from
the engineering stock rooms. Or we could often find them on the
floor. (The term "floor sweepings" is itself a proper designator
for electronic components.) When I first became interested in
building low cost crystal filters (early 1980s) I purchased some low
cost disc ceramic capacitors that were sold as NP0 parts. They
worked fine for crystal filters, but were lousy in oscillators.
Further investigation showed that the caps had low Q. When they
heated, they changed the temperature of other parts within the
oscillator. They were nothing like the wonderful old floor
sweepings.
I rediscovered this problem in recent times with some attempts at
building some VHF bandpass filters with SMT components that I had
purchased. Following the sloppy thinking of earlier times, I had
assumed that if the caps had been purchased from one of the two
major distributors that we use these days and they were sold as NP0,
they would be high quality (whatever that means) capacitors. Not
so. The parts will, I'm sure fall into one of the many
classifications that we can legitimately call NP0. Perhaps they
are C0G, or perhaps C0S or worse. The catalog is vague in this
regard. But the thing I discovered is a sometimes terrible Q, even
at 100 MHz and lower.
The VHF measurements I was doing used air core coils. I've used
these in some Tek situations and knew the Q to be good, although I
didn't always remember the numeric Q results. The inductance was
calculated with relative ease from a standard formula in the ARRL
Handbook and could also be measured with an LC meter. I got
suspicious when I measured especially low Q in a resonator with
physically small coils and a 47 pF chip cap. I rewound a new coil
with the same L, but with much larger wire and a larger coil
diameter and saw only a slight Q improvement. Q measurements were
done with the scheme from Chapter 7 of EMRFD.
Then I remembered some parts that were hidden in the back recesses
of the junk box. These were some porcelain chip parts from a kit
from Transitor Electronics, Inc. I pulled the little 47 pF NP0 chip
from the board and dropped a porcelain cap in the same slot. The
Q immediately jumped up. Changing the inductors now made sense.
This prompted me to purchase a variety of capacitors from Digi-
Key. I have yet to evaluate them, but it's on the list.
One final comment while on the subject: There is a flaw in the Q
measurement scheme presented in EMRFD Ch 7. The fundamental
concept is OK and it's a handy method. But upon looking at the
equations and a program to use them, I noticed that a small error in
the depth of a notch could yield quite an uncertainty in Q. I
have yet to do a careful error analysis, so be forewarned: there
could be a problem. Be sure to do measurements of notch depth
with a good step attenuator or a high resolution power meter. And
be sure that the system really is a 50 Ohm environment.
If you examine a traditional Q meter, you find that the internal
circuit develops a very low impedance drive. This is then applied
to a series LC where the L is the unknown Q part. The higher the
Q, the higher the voltage that appears at the junction of the L and
C. It is that voltage that is measured.
Anyway, there is always fun stuff to be found as we peel back the
layers of the onion.
73, Wes
w7zoi143 2006-09-13 08:37:10 Roy J. Tellason Re: 7 MHz Oscillator Capacitors On Wednesday 13 September 2006 12:20 am, Wes Hayward wrote:
> Hi Dave, et al,Interesting stuff...
>
> Much of this data was abstracted from the January 1983 Tektronix
> common parts catalog, the capacitor and resistor volume.
That's the first I've heard of that book, do you know offhand if it's
available in electronic form anywhere?
--
Member of the toughest, meanest, deadliest, most unrelenting -- and
ablest -- form of life in this section of space, a critter that can
be killed but can't be tamed. --Robert A. Heinlein, "The Puppet Masters"
-
Information is more dangerous than cannon to a society ruled by lies. --James
M Dakin144 2006-09-13 10:57:59 Wes Hayward Re: 7 MHz Oscillator Capacitors Hi Roy, et al,
I know nothing about the book being available in electronic form. I
would be very surprised if it was. Generally it is pretty boring
stuff.
The decades old books that I have kicking around are of little value
to folks in the "outside world." I'm sure that the general purpose
data from the book is available in other places.
Sorry that I have no more to offer in this area.
73, Wes, w7zoi145 2006-09-13 11:13:49 Michael Neverdosk... Re: 7 MHz Oscillator Capacitors A better place to ask for Tektronix books and information is the TekScopes list on yahoo;
http://tech.groups.yahoo.com/group/TekScopes/?yguid=226809086
I hope that link works, yahoo is strange sometimes.
michael N6CHVOn 9/13/06, Roy J. Tellason < rtellason@verizon.net> wrote:On Wednesday 13 September 2006 12:20 am, Wes Hayward wrote:
> Hi Dave, et al,
>
> Much of this data was abstracted from the January 1983 Tektronix
> common parts catalog, the capacitor and resistor volume.
Interesting stuff...
That's the first I've heard of that book, do you know offhand if it's
available in electronic form anywhere?
146 2006-09-13 14:09:01 Roy J. Tellason Re: 7 MHz Oscillator Capacitors On Wednesday 13 September 2006 02:10 pm, Michael Neverdosky wrote:
> A better place to ask for Tektronix books and information is the TekScopesYeah it is. :-)
> list on yahoo;
> http://tech.groups.yahoo.com/group/TekScopes/?yguid=226809086
>
> I hope that link works, yahoo is strange sometimes.
>
> michael N6CHV
It worked, though, so I just joined the group, and am pending. Generally
something like groups.yahoo.com/group/(groupname) seems to work. I didn't
know they had a "tech." portion going in front of that, maybe it's a recent
addition.
I realize that the thread here is about specific applications, but component
selection is still sometimes puzzling to me. I salvage one heck of a lot of
stuff (if you're looking for anything at all feel free to contact me
off-list :-) and sometimes the choices I see don't make a whole lot of sense.
Like using electrolytics down to 0.1uF? When is it sensible to use mylar,
or polystyrene, vs. ceramic, when the values overlap? Stuff like that...
If this in a more general sense isn't necessarily appropriate for here, we
can take it somewhere else. Including a group that I started (called
roys-tech-chat on yahoo) for things that don't quite fit in other places.
Looking forward to your further comments on this subject...
> On 9/13/06, Roy J. Tellason <rtellason@verizon.net> wrote:--
> > On Wednesday 13 September 2006 12:20 am, Wes Hayward wrote:
> > > Hi Dave, et al,
> > >
> > > Much of this data was abstracted from the January 1983 Tektronix
> > > common parts catalog, the capacitor and resistor volume.
> >
> > Interesting stuff...
> >
> > That's the first I've heard of that book, do you know offhand if it's
> > available in electronic form anywhere?
Member of the toughest, meanest, deadliest, most unrelenting -- and
ablest -- form of life in this section of space, a critter that can
be killed but can't be tamed. --Robert A. Heinlein, "The Puppet Masters"
-
Information is more dangerous than cannon to a society ruled by lies. --James
M Dakin147 2006-09-13 17:25:35 topossibilities Re: 7 MHz Oscillator Capacitors > I prefer polystrene caps in VFO circuits.I have had great success in building very stable oscillators using a
type 6 toroid core (TC of +35 ppm/deg-C) and a combination of NPO
(i.e. COG) and polystyrene capacitors (TC of -150 ppm/degree C).
By making 1/3 of the total tank capacitance with polystyrene
capacitors, from start up, the oscillators had a brief (less than 5
minutes) and small frequency change (less than 100Hz). Thereafter, it
stabilized to no more than +/- 5Hz.
Notes: There were Hartley circuits with the feedback tap made at 17%
and used air variable type trimer and tuning capacitors.
Ed, W1AAZ
PS
Another trick to TCing an oscillator is to determine if a + or -
coefficient capacitor is needed. Then if you have a capacitor of the
approximiate value and TC, install it in series with a variable
capacitor (e.g. such that total series C and TC are somewhat near the
required range). Now you can set the required compensation.
Lots of books, parts and kits for sale at:
http://mysite.verizon.net/topossibilities/index.html