[Top] [All Lists]

[TowerTalk] problems with 10m COMTEK 4 sq unit

To: <>
Subject: [TowerTalk] problems with 10m COMTEK 4 sq unit
From: (JayTerleski)
Date: Sun, 18 Mar 2001 17:44:28 -0600

The parasitic array as we both understand are not meant for ultimate
F/B.  I like them only on the low bands where one can listen on a
Pennant, or Beverage system.

The other thing you might look at is the Array Solutions triangle array
controller. This is using true phase and magnitude control to steer the
array, not a hybrid coupler.  It does come with engineering assistance
as well, in setting it up.  

The 4 square array controller is still being finalized as I am still
adding control features into it to make it the best a hybrid coupler can
be, which is not so easy as your finding out.

The reason hybrid couplers are attractive to hams is that they always
"seem" to work with the less than broadcast quality arrays we build. 
They also are a "no tune" or plug and play solution. We hams like that
as well.  The broadcast industry approach would be different and would
use LC networks. Not easy to tune up.  Therefor not a mass market type
device since most hams do not have the tools to do this.

You mention that your dipoles have an impedance of 110 ohms, so the 75
ohm 1/4 transformers are transforming this to 150 ohms.  Not good, the
hybrid coupler is looking for 50 ohms and I am sure Comtek is using 50
ohm transformer windings in the coupler, at least that is the design I
have seen in ON4UNs book.  At 30Mhz things get worse than at 3.5 Mhz so
any stray inductance and capacitance in the coupler will magnify the

You may want to try to reduce the dipoles impedance by using a 2:1
broadband xfmr to take it back to 50 ohms first.  But you also state
that resistors plugged into the coupler are also not giving you the low
SWR you expect to see.

If you have a source of ceramic caps try values of 10-22pf and place
them at the antenna ports to ground with the resistor dummy loads in
place. If the SWR is reduced you know the real sign of the J ohms you
see on the MFJ to be inductive.  The MFJ cannot tell sign it guesses. (
Tom JI will probably tell us why but its a single port device) and the
shunt caps are canceling inductance that I bet you have inside the
coupler.  Reduce the J vector by using equal value of caps on all 4
antenna jacks.  They all must be equal since you will be changing the
phase of the feed. You want to change it equally yo all antennas.

We probably can take this off-line from the reflector and deal with each
other from this point on. I do allot of this helping stuff every week.
Even though I don't make the product. Hi.

Jay, WX0B wrote:
> In a message dated 3/18/2001 14:51:27 Eastern Standard Time,
> writes:
> >
> >  If you build the proper feedlines so you have approx 50ohms of +J (3/8WL
> >  feedline 50 ohms) You can make the unpowered dipoles act as reflectors.
> >  This kind of array has its beginings with the old K8UR array which had 5
> >  dipoles pulled out from a single tower.
> >
> I have tried that and have one working right now. If I set the individual
> radiators (3x) for 50 ohm impedance and run the array as 1 radiator with two
> stub tuned reflectors, the impedance is fine, 50 ohms (F/B or pattern is less
> than what I would like to have), but when I switch to other configurations
> (two radiators, one reflector) impedance changes (hard to see any pattern,
> any time). Behaves the same here with soaked (average) ground and by the salt
> water. It seems that parasitic vertical arrays suck (for max F/B).
>     This is why I am trying to get 4 square going, which provides (on the
> screen) much better pattern, but probable problem with my controller and lack
> of explanation how them boxes work, what is exactly required, how to test it
> and how to fix or tune in case of variation in antenna elements is hampering
> installations.
>     We have situation where we can model the stuff, but are presented with
> box that should work, just plug it in (trust me, it works :-). Some
> explanation and specifications would go long way in helping us to use and
> adapt them properly.
>     When I first got instructions for 80m Comtek box and we wired it up at
> VE1ZZ, the numbering in the instruction book was confusing, ports, directions
> and positions on the switch were numbered 1,2,3,4 each meaning different
> thing (and not correlated). My new instructions have corrections by hand
> after I brought it up. (After operating contest beaming wrong ways, still
> good enough for world #1 :-)
>     I have gone through W4RNL material, it is excellent work (on paper) and
> saved me some digging and fooling around, but things are not as easy with
> real life wires and tubing and magic boxes.
>     It looks like my problem might be caused by faulty ACB-10 box and
> possible problem with radiator's impedances of abt 110 ohms being out of
> range for it.
>     I need sharp, vertical switchable array on the beach to complement the
> rest of the antenna artillery, this is why I am persuing 4 sq.
>     I also managed to model 3 el. Quad and I was pleased to find that
> computer says that (I was right) Quads are not as dumb as some say.
>     I think Stackmatch will be ideal for switching and phasing between
> different antenna systems, but this parasitic triangular array is not what it
> looks like on the computer or what what I would like. Anybody else have other
> experiences?
>     Clock is ticking and I am sitting (again) with bunch of (mismatched)
> vertical dummies, but living, learnig and enjoying every (frustrated) minute
> of it :-) Been there before, will do it again.
> Thanks for any light sheding input!
> Yuri
> --
> FAQ on WWW:     
> Submissions:    
> Administrative requests:
> Problems:       

Jay Terleski
WX0B - Array Solutions
Phased Arrays - Horizontal and Vertical
RF Switches, Antennas, Towers

FAQ on WWW:     
Administrative requests:

<Prev in Thread] Current Thread [Next in Thread>