Mark Connelly, WA1ION - revision 20121128
Date: Wed, 14 Jun 2000 17:33:01 GMT
From: "Mark Connelly" <[email protected]>
Subject: more phased array ideas
Hello to Dallas, Ron, and anyone else interested in this continuing phased-arrays discussion. I'm wondering about the Misek circuit versus other phasers such as the JPS-Timewave ANC-4, the MFJ-1026 (per "http://www.hard-core-dx.com/nordicdx/antenna/special/mfj1026.html" modifications), and my homebrew broadband designs such as Superphaser-1, Superphaser-2, and DL-2. Smoothness of phase adjustment, I suppose, is the main difference, at least regarding the ANC-4. I was not that impressed with the ANC-4 I tried; actually I found nulling with the modified MFJ-1026 easier.
Phased broadband loops may give wider nulls, as Dallas seems to have indicated by reference to the article by O. G. Villard. I think that Tom Rauch (W8JI), a leading 160-m contest operator, uses an array of small broadband active receiving loops with good success. These have matching baluns to go from oddball-impedance balanced to something close to 50-ohm unbalanced to drive a low-noise high-intercept point moderate-gain push-pull amplifier using either CATV bipolars (2N5109 / 2N3866 type) or VMOS power FET's (VN10KM or similar). Some of the new high-speed op-amps (e.g. Analog Devices AD9631/9632) could also be used for the required low-noise / low-distortion gain. Two broadband loops spaced at 50 m / 164 ft. gave good reduction of WQEW-1560, 50 kW at 200 miles, to permit logging India-1566 at the Rowley salt-marsh. WQEW is one of the nastier stations (along with WPTR-1540) to null here in the Boston area because of a huge night signal and a jumping-around skip angle that's almost like shortwave.
I think that two K9AY loops (which, when adjusted, already have a tendency toward a cardioid pattern) could be spaced 50 m and phased to create a cardioid null of exceptional width and depth to take out high-angle skip and multiple targets within a 20 or 30 degree swing off the main null line. (The K9AY loop is covered in detail in Al Merriman's article at "http://home.inforamp.net/~funk/termloop.html" and on the Wellbrook page "http://www.wellbrook.uk.com/K9AY.html".)
The set-up I usually use on mobile DXpeditions consists of an active whip (MFJ-1024 or equivalent) phased against a broadband 6 ft. per side square loop. This creates the classic loop-sense cardioid array on which Ron Schatz did pioneering work for MW use back in the '60s / '70s. See Reprint A20 which may be ordered from the National Radio Club. The use of a loop versus a whip allows co-location of the antennas on the car roof: no physical separation is necessary. The loop-sense concept produces a broad single null; the idea has been around since the earliest aircraft and ship RDF (radio direction finder) systems.
The loop structure I use is comprised of two 6' [1.83 m] whips at ends of a 6' wooden board spring-clamped to heavy-duty Thule roof-racks. 3' wires from the base of each whip go to the amplifier's balanced terminals and a 6' clip lead connects the tops of the two telescoping whips. See "badxph99.htm" for a link to a picture of the set-up and "bbl-1a.pdf" for construction data. I have recently had good luck using the Wellbrook Communications ALA-1530 Amplifier with this loop: nice low noise performance, moderate gain, no IMD. The Kiwa broadband amplifier should also be usable if a ferrite-core balun transformer is placed between the two loop wires and the single-ended amplifier input. Construction details for another suitable amplifier, the BA-1 Balun/Amplifier, are available at "bbl-1b.pdf".
For an array of more than two antennas, I think I would go with four broadband loops, one on each corner of a square about 150' on a side. The northeast and southwest corner loops could both be oriented for maximum NE/SW pick-up (minimum NW/SE pick-up) and the northwest and southeast corner loops could both be oriented for maximum NW/SE pick-up (minimum NE/SW pick-up). To finish things off, a fifth antenna - an omnidirectional active whip, would be positioned in the middle of the square. This set-up would give the operator a lot of variety in both single-null solutions and for multiple nulls created by using two or more phasing units.
73 / good DX ... Mark
Mark Connelly - WA1ION - Billerica, MA, USA
(following is from Dallas Lankford)
Thank you very much for forwarding the very interesting note from Ron Hardin about using multiple ANC-4 phasing accessories to phase "multiple + 1" Dymek McKay DA100E active verticals. I was especially interested in the following comments by Ron.
"With 3 antennas and 2 ANC-4's deep nulls become child's play instead of hyper-sensitive adjustments; a 4 antennas and 3 ANC-4's lets you pretty easily null out two stations and hear a third. Beyond that it's hard to find a justification."
When Russell Scotka and I were doing our work on phased antennas about 1992 - 1995, Russ observed that the ANC-4 would not generate as deep nulls and not as stable nulls as our variation of Misek's phasing circuit. Apparently by using a third antenna and a second ANC-4 this disadvantage of the ANC-4 is overcome.
Equally interesting is Ron's experiences using 4 antennas and 3 ANC-4s to routinely null two signals and hear a third. This has been talked about by some people I know, but to the best of my knowledge no one has succeeded until Ron.
My current phased antennas are two 30 feet long 15 feet high noise reducing inverted L antennas, separated about 150 feet, on a north-south line, with the horizontal parts of the inverted L antennas perpendicular to the line connecting the bases of the vertical parts, and oriented in the same direction. Lesser separations give lower "left over" signals after nulling. Greater separations were not deemed necessary or desirable.
I tried vertical noise reducing antennas, but nighttime MW skywave signals were greatly reduced with verticals. My antennas are not active, but passive, so apparently one gets higher signal levels with active whips. I call this to everyone's attention because it might be possible to use noise reducing inverted L antennas (perhaps longer than the 45 footers I use) with ANC-4s and save on the cost of Dymek McKay DA100Es.
Best regards, Dallas
How wonderfully simple and elegant. Now why didn't I think of that? This would be useful not only for nulling two signals, but also for nulling a signal and a noise source.
Yep. But I believe your 4 antennas / 3 ANC-4s is undoubtedly the best way to go.
I have to admit that setting up or moving your whips is a lot easier than setting up or moving my inverted Ls. I've spent many an hour hoisting even my small inverted Ls up into trees and down again with ropes tossed over high branches.
I gather from what you said above that you mainly use your phased arrays to listen to daytime MW signals. Russ and I found that for nighttime listening, the Misek phasing circuit and our variations of it could not be equaled. It is relatively easy to get deep, stable, nighttime, skywave nulls at the lower end of the MW band with most phasing circuits. But it is not at all easy to get such nulls at upper end of the MW band. In our experience, only Misek's circuit suffices. Of course, we never imagined or tried systems like you have described. I, for one, would be very interested to hear about your experiences nulling nighttime, skywave, MW signals at the upper end of the MW band (say > 1200 kHz). I guess what I'd like to know is whether I could buy 3 ANC-4s, or if I am going to have to put out another 200 man hours to build two more Misek phasing control boxes in order to obtain dual signal nulling capability and retain my nighttime top end nulling capability.
You didn't explain why with 3 antennas and 2 ANC-4's deep nulls become child's play instead of hyper-sensitive adjustments. This is as interesting to me as your dual nulling capability.
Best regards, Dallas
Ron - I really can't thank you enough for taking the time to share the results of your multiple phaser experiments.
Ron Hardin wrote:
Again, why didn't I think of that? Duuuhhh!!!
You may be right when you point to the granularity of the pots as one of the limiting factors for the nulls. Russ and I spent considerable time finding suitable pots. At first Russ was using Clarostat 2 W linear model "380s," if I remember correctly. These were fine until some time around 1993 or 1994, when (apparently) Clarostat made some kind of production change, and their "new 380s" went bad (scratchy... didn't hold deep nulls... etc.) after only a few weeks. Russ was able to get a few Allen Bradley Type J pots, which worked fine, but he learned that the Allen Bradley pot division had been bought out by Clarostat and wasn't making pots at that time. About 6 months later Clarostat Type J pots appeared in the Newark catalog. Russ bought some, they were excellent, and we have used them ever since.
When I finally built my modified version of Misek's unit, I also used the AB (er, Clarostat) type Js. The 200 ohm, which we use, currently sell in Newark for $18.22: Newark 01F6262, Clarostat (AB) type JA1N056S201UA, linear taper, hot-molded carbon potentiometer.
I have tried "bandspread pots," but they did not enable me to get deeper nulls with the Misek circuit. They did facilitate nulling in some cases, but bandspreading is hardly worth the effort for a Misek type phasing circuit. I can't say if bandspreading would be useful for other circuits, such as the ANC-4.
The Misek circuit with inverted L antennas also does not permit SW nulling. According to O. G. Villard in his 1993 WRTH article, "More compact directional SW antennas" (based on propagation research during the 1950s), a SW signal is usually composed of multiple arrival (vertical) angle components. To null out all of these skywave "modes" at once, a directional antenna must have a fan shaped null, at least in the lower part of the vertical plane. Villard used pairs of phased loops to generate fan shaped nulls. I never tried any of his designs.
Again, thanks for all the information.
Best regards, Dallas
Date: Tue, 13 Jun 2000 19:55:53 GMT
From: "Mark Connelly" <[email protected]>
Subject: phased arrays
Hi Al (/ others). Thanks for forwarding Ron Hardin's write-up on the phased active whips. At home, and at most other house QTH's I've used, active whips pick up too much electrical noise to be of much use. On field trips to electrically-quiet beach sites, I have phased two MFJ-1024's (and/or homebrew equivalents) at an approximate spacing of 50 m / 164 ft. Nulls with this simple set-up have been good.
Optimum spacing should be between 1/12 wavelength and 1/3 wavelength over the frequency range of interest. 50 m was chosen for good overall medium-wave and 160-m ham performance because it is 1/12 wavelength at 500 kHz and 1/3 wavelength at 2 MHz.
With antenna separation along an east-west line, a null of a station to the west corresponds to maximum signal from a station to the east, and vice versa. Nulls can be rotated about the compass, but it should be noted that, with the above array, if you null a station to the north, since there is no arrival time difference between the two whips, you also null signals from the south simultaneously, thereby producing a figure-of-8 pattern with lobes east and west. You would need to have the two whips on something within 30 or so degrees of a north-south separation axis if true cardioid nulls north or south are desired.
Regarding the use of 3 or more whips and 2 or more phasers to get multiple movable nulls, it certainly sounds like a grand scheme but one that would be difficult to implement at a field site (at least if it's land you don't own) and not likely to be low- noise enough in average residential settings.
Arrays with two (or more) broadband loops might even be better than ones comprised of broadband active whips. With the 50-m separated east-west case above, and the 2 loops aligned for max. east-west pick-up, some of your north-south QRM would be reduced by being off the sides of the two loops, then you'd kick in the phasing to kill off the stuff from the west. I think one of my India-1566 receptions from the Rowley, MA salt-marsh used a set-up like this. Who knows what you could do with a big array of broadband loops spaced at different distances and pointing different directions ?
Two loops, if at a right angle, don't even require separation distance to allow rotatable nulls to be produced with a phasing unit. Also, unlike the whips, the loops aren't going to pick up that much local electrical noise; hence, a system that is more useful in a suburban residential setting.
73 / good DX ... Mark
Mark Connelly - Billerica, MA, USA
(following is from Al Merriman / Ron Hardin)
Subject: Fw: MW active antenna's
Date: Sun, 11 Jun 2000 09:39:59 -0400
I forward this for what it is worth.
---Subject: Re: MW active antenna's
Added note: the July 2000 issue of QST magazine has a very good article on Flag, Pennant, and Ewe antennas used on the 160-meter band. There are some related articles available on the Web at "http://www.angelfire.com/md/k3ky/page37.html".
Low Angle Antennas
Date: Fri, 12 May 2000 05:10:16 GMT
(response to recent hard-core-dx discussion)
I've had a couple of vertical and sloper experiences. In 1975, when I rented a house at Sudbury, MA (about 30 km west of Boston) I had a 400 m Beverage towards Europe and a 30 m vertical wire (pine tree supported). At or before sunset, on initial fade-ins, the vertical actually brought in clearer low band European MW stations (Daventry, UK on 647 was common in those days). Within an hour, the Beverage with its tighter directional pattern became the superior antenna (since it rejected increasing skip from stations to the west as evening progressed). Phasing the vertical against the Beverage sometimes gave results superior to either antenna by itself.
More recently ... summer of 1999 to be exact ... my family stayed a week at my brother-in-law's house in East Harwich, MA on Cape Cod. The house is about 3 km inland. A loop used right at the shore picks up very impressive Canadian Maritime Provinces AM stations on groundwave during the day and strong Europeans (especially Spain) at local sunset. The performance of the small ferrite loop at the E. Harwich house is far less impressive because the 3 km of sandy soil between the house and the shore really eats up signal strength on an antenna so close to the ground. At such a location, it's a good idea to elevate any antenna being used (other than a Beverage) at least 10 m / 33 ft.
Just like Tom's area in NJ, the East Harwich house has many pitch pine trees (indicators of sandy soil) near it. When I spend a few days there, I put up two slopers of about 25 m each from the house to nearby trees. One sloper is oriented for northeast pick-up (Europe) and the other for southeast pick-up (Brazil, eastern Caribbean). These antennas provide excellent reception in their respective directions, thoroughly beating any ferrite or small air-core loop, especially at local sunset when medium wave signals from Europe, the Middle East, Africa, and eastern Brazil are just starting to fade up. Last July, the southeast sloper brought me some sunset-period low-band Brazilians that I have never heard from any other US location.
The Europe-oriented sloper provides reception quality that a loop can only achieve if operated right at the seaside. Some of the better catches have been Italy-900 overriding CKDH/other domestics and Qatar-954 over Spain.
The E. Harwich slopers worked reasonably well straight to the receiver or phasing unit. To reduce electrical noise, I tried 4:1 baluns with a sloper to one primary side and a 100 m radial wire (lying on the ground and not connected to mains earth) to the other primary side. Coaxial cable to the receiving position came off balun transformer secondary windings.
When I really wanted to crush the domestic interference, I connected the two slopers to inputs of a phasing unit and dialled up nulls that put even pesky New York stations like WBBR into the "mud" to allow co-channel Latin American and adjacent channel European / African DXing. The best null results are on an east-west line, i.e. the bisector of the angle between the two right-angle (NE, SE) slopers. Two loops placed orthogonally in the vertical plane would also have best phasing-unit generated nulls on the bisector of the angle between the loops. The same principle applies to two horizontal longwires run out from the receiving position at a right angle.
A vertical phased against a broadband loop is another good solution for limited-space installations.
Two verticals spaced at about 50 m and sent to a phasing unit can work well if local electrical noise isn't too bad.
The above discussion, and the recent K9AY Loop experiences by the Grayland, WA medium wave DXpeditioners, show that good low angle pick-up and directivity do not always require a Beverage antenna.
The Top-Band (160-m) ham forum on contesting.com also discusses this subject in considerable detail.
Mark Connelly, WA1ION
(following is from Tom Sundstrom on Hard-Core DX list)