On 1/22/20 10:23 AM, VE6WZ_Steve wrote:
Hi Mike
I have not considered the temperature coefficient of the core. That is a great
suggestion.
I did find some generic plots of permeability vs temperature and there can be a big
change. The data sheet from Fair Rite does not quote any temperature coefficient
values. I’m not sure how to to test that easily other than heating the core in
the field on a cold day. Or perhaps cooling it with an air can upside down in the
summer.
Which material is the core made out of? The Fairrite catalog has
properties vs temperature data in it for some materials. Pages 6 and 7
have the tempco of initial permeability (1.6 %/C for 31 mix, 1.25 for
43 mix, 0.10 for 33 mix, for instance)
The pages for 31 mix (page 17) show mu vs temperature from -50 to 150C,
measured at 100 kHz (about a 1:3 variation from -50 to 25, BTW)
and percent impedance vs temperature at 4 frequencies. (10,25, 50, 100
MHz), where the loss is dominant.
The other mixes are the same.
Some of the mixes that are more intended for power supply usage (79,77)
have power loss density vs temperature at a couple frequencies, for
instance
The other option is to sweep the raw wire in the field, but at -38C that would
be rather difficult.
I do plan to experiment with changing the termination resistor value this
summer to see if I can replicate the cold weather sweep.
73, de steve ve6wz
On Jan 21, 2020, at 6:50 PM, Michael Tope <W4EF@dellroy.com> wrote:
Steve,
Have you eliminated the core material in the 9:1 transformer as a possible
source of the problem?
I don't see how the resistance of the galvanized wire changes the surge
impedance significantly. When viewing the beverage as a transmission line
(single round conductor over ground plane), I would expect the distributed
resistance of the steel wire would to make the transmission line loss/unit
length higher, but other than adding a slightly reactive component to the
characteristic impedance, the steel wire resistance change shouldn't make the
real component of the transmission line characteristic impedance change much.
73, Mike W4EF...............
On 1/21/2020 3:55 PM, VE6WZ_Steve wrote:
A few days ago I posted a video explaining some experiments I did adding short
radials to my feed and termination grounds on the Beverage.
During that week temperatures went as low as -41 C at the station.
The mystery has been solved.
I wondered if the frozen ground was limiting conductivity and upsetting the
ground resistance part of the impedance. The experiments showed that adding
radials to stabilize the frozen ground had no effect.
In the video I show how the real resistance of the 920 foot galvanized steel
wire might drop 10-20 Ohms when the temperature drops 50 deg C from the
summertime. This would mean that I need a LOWER termination resistor to match
the surge impedance.
However, I also show that the ceramic termination resistor I use has a very
high negative temperature coefficient (-1300 C). This results in an INCREASE of
30 to 40 ohms in my 470 Ohm termination resistor. Therefore, the termination
could be 40 to 60 Ohms too low!! I show in the video that the resistor does
indeed increase by 40 Ohms when I chucked it out my back door at -30C for an
hour.
Today I was at the remote and it has warmed up to -1C (35 deg C warmer) and the
analyzer sweep of the wire has mostly returned to it summer time reading.
This seems to indicate that the main cause of the Beverage miss match is the
temperature coefficient of the wire and the termination resistor. Obviously the ground
is still just as frozen today as it was last week at the feed and termination and
under the wire length! Perhaps I will look into changing the term resistor with a
metal foil unit that has a positive temperature coefficient. Carbon composition units
also have a negative TC, but not as severe as the ceramic units. Those ceramic units
are almost like thermistors! (well….not really)
However, as I said in the video, this is like a “solution looking for a
problem” because few Hams experience these extreme temperature changes, and even when
we do, it's just for a short time. Also, Beverage modelling shows that the termination
resistor can effect the F/B, but it has limited effect on the RDF which is what really
matters. This exercise was done mostly out of interest and to try and solve the puzzle.
I made a new video showing the sweeps I made today if interested. The video
includes the original content so if you've seem that just jump to the end.
YouTube video here: https://youtu.be/3Oft826Q8tA <https://youtu.be/3Oft826Q8tA>
73, de steve ve6wz
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|