On Sun, 19 Aug 2007 16:28:32 -0000, wrote:
Hello,

I bought some no name RG-59/U coax (RG-59/U-SP-95 Made in the USA). I
went to cut and tune a 1/4WL matching section for 3.8Mhz. I cut the
coax a little longer because I intended to tune it with my MFJ259B
or.LP100 exactly to the design frequency. I have a 50 Ohm dummyload
that actually is about 54 Ohms @ 3.8Mhz. Plugging the numbers into
ON4UN's program for calculating impedance along a coax line: Given the
coax is 75 Ohms and the load is 54 Ohms, the program reported that if
the line was a 1/4WL long the impedance at the source end would be
about 105+j0 Ohms.

However my actual measurement with both the MFJ259B and LP100 showed
an impedance of 74 + j0 Ohms.

http://remote.wu2x.com:8888/lee/quar...-75-meters.jpg

I plugged in a few numbers into ON4UN program and calculated it would
take coax that had a characteristic impedance of 64 Ohms to see the
transformation that I am seeing.

Is there any error in my logic here? If this coax really is 64 Ohms,
then I'd like to find something that really is closer to 75 Ohms so I
can achieve the 2:1 ratio that I intended.I still have another 100
feet of it and can do more tests with the tools I have on hand.

After looking over some of the followups to your post, I had a thought.

Addressing the errors' introduced by loss, and the possibility that
you're seeing worst case loss at a lower frequency, why don't you go
back and re-measure your matching section at 3X, 5X, 7X 3.8 mcs.?

It's the reverse of a trick I've used to cut n*1/2WL 1:1 coax runs for
VHF -- by measuring at *sub* multiple freqs.

Anyway -- it's a thought.
Jonesy
--
Marvin L Jones | jonz | W3DHJ | linux
38.24N 104.55W | @ config.com | Jonesy | OS/2
*** Killfiling google posts: http://jonz.net/ng.htm