Maximum allowable error when cutting
the phasing lines
The following discussion about the maximum allowable error when
cutting the phasing lines place in the Moon-Net reflector. I found the subject so interesting that I decided to
put together all the messages in this page.
On 7-May-2001 G0RUZ wrote: I am just about to trim some phasing lines for 144 and 432
Mhz. I have a vector voltmeter and a signal generator and a 50 ohm hybrid so I
can measure phase with good accuracy. I was wondering what is considered an
acceptable phase error on phasing lines. I have been told that anything less
than 4 degrees is OK. This equates to 23mm on 144 and 7.7 mm on 432.
This is quite easy to achieve mechanically and in fact my early measurements
indicate that I have achieved a 1 degree error on 144 without even trimming! I
managed this by using a piece of tubing the same length as the line needs to be
cut. I put a connector on one end which could not pass through the tube and then
pulled the cable through the tube and kept the tension on the cable high while I
sawed it flush with the end of the piece of tubing. I then measured the phase
error between the first cable cut and the seven subsequent cables and all were
around 1 degree or so. I was slightly surprised by this and so cut another about
20mm long. This showed a error of + 4.2 degrees which is pretty good when the
velocity factor of the FSJ4 is taken into account.
The 70cms cables will be another matter entirely of course.
I would be interested in peoples opinions of this subject, please post
replies to moon net and not just direct to me as I am sure that others will be
interested.
On 7-May-2001 G8MBI wrote: According to results of my modelling experiments then
'phase' issues are over hype'd
having said that good or perfect, is better than error.
the first really noticable effect is a dip forming in the antennas normally
'smooth' nose pattern, this becomes visible on simulation plots at about 10
degrees of error. But you would never find it (at 144 anyway) in real world
measurement.
2 times 17m2 bayed at dl6wu spacing =4492mm
---phase one antenna with respect to other degrees---
ref -5 -10 -15 -20 -25 -30
gain 17.6 17.59 17.56 17.52 17.46 17.38 17.29
F/B 19.2 19.2 19.2 19.2 19.2 19.2 19.2
1st lobe -10.72 -10.33 -9.96 -9.56 -.05 -8.56
As I wrote last week below about 10 degrees it is insignificant.
these are peak gain, the dip forming in the centre will give a small 'hole'
in this range of error, my notes do not show any measurement of the hole, I
assume then, that at the time I thought it was not significant enough to write
down, it tends to form but then flatten until you go very extreme.
I know for a fact that the most common construction cock up of 180 degree
error is easily found as the hole is actually 'audiable' when the antenna is
moved from side to side....(don't ask me how !!)
don't do it...but below 10 or 15 degrees of error, the world ain't gonna end
as MANY forecast it will on this subject.
Your mileage may vary according to antenna used, segmentation, and modelling
package.
On 7-May-2001 Mark wrote: In broadcast antenna arrays the spec used on most systems is
+-5 Deg error. This is primarily for control of sidelobe response. It is tight
enough to allow for real variations that occur when distribution cables are bent
into position. The antenna pattern will not be significantly disturbed with
errors less than 10 Deg, Aim for a very acheivable 5 deg and you have the
confidence that the array will work properly. Note that if you are unsing foam
filled hardline cables, different batches from the same manufacturer can have
noticibly different velocity factors. Over long lengths it is possible to have 2
cables of similar mechanical length with a different no. of electrical
wavelengths for each. Worst case you can have a cable exactly half wave out, by
most phasing methods. Do your phasing at more than one frequency, and start at a
low frequency to set the ball park length. and move to the operating frequency
or above when you are sure you cannot make a halfwave error.
On 7-May-2001 G0RUZ wrote: Hi Mark, Graham et all your advice echoes what John Denby
G3TSA said. I forgot to point out that I had already started the checking
process at 30Mhz and worked up to 144. The cable was all off the same roll and
it all seems to work very well. You can get very close mechanically with the
method I used. I am quite convinced that I am measuring things properly. So on
144 I reckon if you use cable off the same roll and cut the cable mechanically
accurately you will have little to worry about.
On 7-May-2001 SM5BSZ wrote: The far field of any antenna is completely defined by the
field on a surface that surrounds the antenna. (Maxwells equations)
This means one can calculate the far field radiation pattern as if the field
was generated by radiators on the surface excited proportional to the field on
the surface.
For highly directive antennas like the ones we use for EME the field is small
everywhere except in front of the antenna so the far field is completely defined
by the field on a flat surface in front of the antenna because the field on the
rest of a surface that would be required to contain the antenna in a closed
volume will be so small that the contributions can be neglected.
Imagine a flat surface about 1 stacking distance away from the last director.
Compare it to a flat surface in front of a parabolic dish.
The phase and amplitude should be constant in both cases. The far field 'does
not know' if the antenna is a parabola or something else so start by considering
a parabolic reflector:
There are rules of thumb saying the accuracy of the parabolic surface has to
be better than +/- 0.1wl corresponding to a phase error of +/- 0.2wl = +/- 72
degrees!!!!! (most unfavourable case) There will be some loss of gain, I do not
have any book with design rules for parabolas but I think the expected loss vs
parabola accuracy is well known in this group.
In case the antenna behind the imaginary flat surface is a yagi array we do
not have to make the phase better than we have to in the parabolic reflector
case.
We have to worry about ALL contributions to phase errors!!!!
In order of significance (probably):
1) Boom mounting poistion. All feed points have to be in a plane. If the
support structure sags when the antenna is elevated the phase will differ by the
amount that the feed points deviate from the best fitted flat surface. (Look at
the elevated antenna from the side and place youe eyes in the plane of the
radiators. It is then easy to estimate the deviations from a flat surface if you
also hold a thin tube in the plane a meter away from your eyes to get a straight
line to compare to)
2) Antennas. A yagi antenna is a high Q resonator. The current in the
radiator is not in phase with the drive voltage from the cable unless the
impedance is resistive. Even if the impedance is forced to be 50 ohms resistive
by some tunable matching device the outgoing wave may still differ in phase if
the antennas are different.
Even if the antennas are mechanically identical so they give accurately the
same phase when measured one by one, the phase of the outgoing wave may differ
because of detuning due to improper mounting of the antennas.
a) Understacking. The field from the near neighbours may produce additional
currents on the antennas that twist the phase of the total current (also changes
the feed impedance). In a 4-stack all yagis see the same effect but in a larger
array the effect differs for different antenna positions.
b) Booms or elements from other yagis at other frequency bands couple to the
E-field of some yagis, lowering their resonance frequency while other yagis are
unaffected.
3) Cables. Normally the phase shift from cables is accurately determined by
the physical length. On 144MHz some serious mistake is required to loose gain by
phase errors due to cable length errors but maybe(?) one has to be careful on
432MHz when using open balanced lines.
Do not worry too much about the cables!
(Note: I'm not showing any E-Mail
address here in order to avoid them from being collected by SpamBots. You can
possibly find the E-Mail addresses of the above OM at QRZ.COM.)
This discussion is still not closed. If you have any opinions or additional
related information you would like to be published here, just send
me an E-Mail
German
French
Portuguese
Russian
Read
other Moon-Net discussions
Home page
Site map
Radio Sherlock search engine
Terms of service
Privacy
Cookies