MW Phaser #2.pdf

(600 KB) Pobierz
MW Phaser #2
Dallas Lankford
6/27/05, rev. 6/12/07
12/11/2010: This is the return of the MW Phaser #2 article which resided in The Dallas Files for many
years. A few typos have been corrected; a few sentences have been deleted.
Attention: 7/21/07 The main supplier of 200 and 50 ohm pots, Newark InOne, has recently raised the
price of their 200 ohm Type J pots (now manufactured by Honeywell, formerly Clarostat, formerly Allen
Bradley) to about $50 each, and discontinued their 50 ohm Type J pots. They have a few (34 as of today)
50 ohm Type J pots for about $20 each (buy them while you can). Surplus Sales Of Nebraska currently
has 250 ohm Allen Bradley Type J pots (NOS = new old stock) for $6 each. I have been using the 250
Ω AB Type J pots for several weeks in a new high performance phaser, and they seem to be suitable
substitutes for the specified 200 ohm Type J pots. (Yep, 12/11/2010.) NTE may also have suitable 50 Ω
pots, NTE number 501-0001, Mouser 526-501-0001 for $12.50 each. (Nope, 12/10/2010. They quickly
became “noisy.”)
MW Phaser #2 is a slight variation of one I built about 10 years ago The previous one used twinax
antenna lead in inputs and baluns to convert from balanced to unbalanced. I also used physically larger
toroids in the first one. Previously Russell Scotka had built a copy of Misek's phaser which motivated
me to make the modifications described below. I have used virtually every variable phaser which I could
build, buy, or invent since the early 1980's, but none of them before or since MW Phaser #1 worked
anywhere nearly as well. Other phasers generally suffer from one or more serious defects, including (1)
they generally do not work well throughout the entire MW band on sky wave (nighttime) signals,
especially at the higher MW frequencies, (2) they often do not work well on ground wave (daytime)
898282811.078.png 898282811.089.png 898282811.100.png 898282811.111.png 898282811.001.png
signals; it may be quite difficult, and sometimes impossible, to generate and maintain deep, stable nulls,
(3) they may cover up weak signals with internally generated noise (due to noisy active combiners and/or
noisy amps), and (4) they may cover up weak signals with intermodulation distortion. MW Phasers #1
and #2 are modifications of the phaser designed by Misek; see The Beverage Antenna Handbook, Second
Edition, by Victor Misek, 1987. The schematic above is a variation of MW Phaser #2.
Here are some of the modifications I made to Misek's original design. In the phasing branch of the
circuit I used a DPDT toggle switch as a band switch for approximately the upper half and lower half of
the MW band. Actually, MW Phasers #1 and #2 use rotary switches to provide nulling from at least 150
kHz to 30 MHz. The "bands" and values of inductors and capacitors are given in the table below. The
values may be rounded off to the nearest standard values. Of course, at SW frequencies the sky wave
nulls are generally not very deep or stable. but the phaser can
generate very deep and stable nulls for ground wave shortwave
signals (power line noise, neighbor's electronic noise makers, etc.).
There is considerable overlap in the bands above; for example, the
700 kHz band produces good nulls in all directions over a frequency
range of at least 350 - 1400 kHz. For the inductors I used an Amidon
T106-1 powdered iron toroid wound with #24 enameled copper wire
tapped at 3T, 5T, 7T, 11T, 16T, and 25T (T = turns), with the first 7
turns wide spaced and the remaining turns close spaced (the tapped inductances were checked with an
inductance meter, but this was probably not necessary). The AL value of the T106-1 is 325, i.e., L(uH) =
.0325 T^2. As can be seen, the inductances calculated from the formula were somewhat less than the
measured inductances. I do not know if this is true in general for T106 toroids, or if I happened to buy a
batch of T106 toroids
substantially out of spec. Other powdered iron toroids can be used instead of the T106-1; I would
recommend using a mix which is optimal for the MW band, and a reasonably large toroid (about 1 inch
outside diameter). Perhaps I should have used off the shelf inductors, but I decided to use a large
powdered iron toroid to minimize the possibility of intermodulation distortion products. If you use off
the shelf inductors, do not use tiny subminiature inductors. In the phasing branch I used 50 ohm pots as
vernier phase controls which makes it easier to maximize nulls depths. Do not buy cheap 50 ohm and
200 ohm pots. Cheap pots will wear out quickly, becoming noisy, making it difficult to generate and
maintain deep, stable nulls. In my opinion there is only one type of pot to use, Allen Bradley (previously
marked AB) type J. Allen Bradley has not made pots for many years, but one can sometimes still find
AB type J pots at hamfests. Clarostat bought out the pot division of AB and continued to make type J
pots. I have used the Clarostat type J pots and they seem to be as good as the AB's. The only supplier I
know who sells both the 50 ohm and 200 ohm pots is Newark InOne . It appears that
Honeywell recently bought out Clarostat. Whether or not they will continue to make a good quality type
J pot remains to be seen. Possibly some other make of pot rated at 100,000 rotations (or more) would be
satisfactory. I don't know. Misek used a single FET amp in his phaser, but Russell Scotka told me it
caused intermodulation distortion products in high RF environments (Miami, FL). That motivated me to
898282811.011.png 898282811.022.png 898282811.031.png 898282811.032.png 898282811.033.png 898282811.034.png 898282811.035.png 898282811.036.png 898282811.037.png 898282811.038.png 898282811.039.png 898282811.040.png 898282811.041.png 898282811.042.png 898282811.043.png 898282811.044.png 898282811.045.png
develop the dual 2N5109 push-pull Norton amp used in MW Phasers #1 and #2 and elsewhere. The 100
ohm pots in the Norton amp are small 25 turn 1/2 watt pots which are adjusted so that each 2N5109
draws 16 mA (160 mV measured across the 10 ohm resistors). The 2nd order intercept of the amp can be
increased by adjusting one of the 100 ohm pots to minimize 2nd order intermod while doing a 2 tone
intermodulation distortion measurement. The 2nd order optimization is frequency dependent, so I just
pick two of my favorite frequencies, say 600 + 700 = 1300 kHz, to optimize on. The 2N5109's should
probably be heat sinked. I made heat sinks from some small rectangular sheets of copper formed into
open cylinders around some drill bits, held onto the 2N5109 cases by tension. Ferrite beads (FB-101-61)
are used on the collector leads of the 2N5109's for parasitic oscillation suppression. I have also used
MRF-581A's, like KIWA does for their version of this amp. MRF581A's drawing 19 mA do not require
heat sinks. The Phasing dots on T8 and T9 must be observed for correct negative feedback; otherwise
the amplifier will probably oscillate. The 22 ohm and two 270 ohm resistors in the non-phased path are a
3.5 dB attenuator. The attenuator is necessary to prevent running out of phasing range (at one end of the
phasing pots range or the other) for some sky wave signals and some antennas. Later (7/25/05) rear
pannel switched 3.5 and 7 dB attenuators were found useful for some antennas arrays; see Field Change
1 above. To recombine the unphased signal with the phased and amplified signal I used a hybrid
combiner based on Hayward's design. This completes the description of the major modifications I made
to Misek's phaser.
How should you use this phaser (or, for that matter, any other phaser) in the MW band? If, like Misek
did, you use it with a two wire steerable beverage antenna, then you should do what Misek did.
Unfortunately, his Beverage Antenna Handbook is copyrighted, so it is not appropriate for me to
reproduce the relevant parts of his book here. Fortunately, the handbook is available from Universal
Radio and from other sources. Just do a Google on the net. Like many people, I do not have enough real
estate for a beverage. But I do have a 200 foot by 300 foot lot, and that is sufficient for a pair of noise
reducing inverted L and a pair of noise reducing vertical antennas spaced about 160 feet apart, 1/10
wavelength apart at the low frequency end of the MW band. The 1/10 wavelength condition is required
for adequate residual signals when nulls are formed using two antennas of the same kind. My inverted
L's have 15 foot vertical legs and 30 foot horizontal legs. The L's have 4:1 turns ratio step down
transformers at the knees of the L, and the ground level end of the L is grounded with a 4 foot industrial
grade ground rod (i.e., they look like the letter L rotated 90 degrees clockwise). The verticals are L's
hung from tall trees. For best nulls the L's should be parallel and point in the same direction; the best
nulls are in arcs about the line perpendicular to such L's. For example, my L's point East, and my best
nulls are towards the North and South. Nulls to the East and West are good, but not as good as to the
North and South. 160 foot spacing is not adequate for good nulling throughout the VLF band, especially
al lower frequencies. Yes, I can null some VLF signals with the two 160 foot spaced inverted L's, but I
generally can't null them completely, and in some cases there are few, if any, residual signals after
nulling the primary signal. For VLF + MW nulling I use an inverted L and a big air core loop. Little
separation between the inverted L and loop is necessary; you should not run the inverted L within about
20 feet of the loop. I also put up a 2nd big air core loop spaced about 160 feet from the 1st big air core
loop (lying in the same plane) to see if they provided better nulling of undesired signals. It seems to be
the case, but is not as dramatic as I had hoped for. In any case, all of my antenna arrays produce excellent
ground wave and sky wave nulls. Some people have reported excellent nulling results using two random
wire antennas and other phasers. I doubt this. Yes, they obtained some nulls. But it is unlikely that they
obtained anything resembling the nulls obtained with a two wire beverage antennas, two spaced noise
reducing inverted L's or verticals, or a noise reducing inverted L or vertical and a big air core loop using
the MW phaser described here. But note that nulls may be unstable when the geomagnetic field is
disturbed, as Bjarne Mjelde pointed out to me, and skywave nulls may be unstable during sunset and
sunrise transitions.
To null a signal, select the frequency "band" closest to the desired frequency and set the vernier phase
pots to mid range (12 o'clock when the knobs are calibrated). Next rotate first one phase pot (200 ohms),
and then the other, watching the receiver S-meter for a dip in signal level. It may be a good idea to first
null a ground wave because they are more stable. After a dip is found with one phase pot, then try to
deepen the dip with the other phase pot. If little or no dip is found with either phase pot, use the reverse
antennas toggle switch to reverse the antennas and start over. Alternate back and forth between the phase
pots until the deepest null is found. Then alternate back and forth between the vernier phase pots, trying
to make the null deeper. If the null is not very deep, use the antenna reverse toggle switch to reverse the
antennas, and start over, resetting the vernier pots to mid range, followed by alternating between
adjusting the phase pots, etc. Sometimes good nulls may be found with the antenna reverse toggle switch
in both positions, but one of the nulls may be somewhat better than the other. So it is a good idea to
check both. Once a null is generated at a particular frequency in a particular direction, there is usually an
almost as good null at nearby frequencies in the same direction. This characteristic is sometimes useful
for nulling strongly fading signals is there is a more stable signal in the same direction at a nearby
frequency; null the stable signal, then tune the strongly fading signal, and finally adjust the phaser
controls to try to deepen the null on the strongly fading signal.
Below is a parts list for MW Phaser #2 (MWP2) and other phasers which have evolved from it, including
MW Phaser #3, and MW Phaser #4 (A and B), all of which will be referred to as MWPx (to be described
in another article). Most of the work on the parts list was done by Bjarne Mjelde, with me filling in some
of the blanks. We hope that the parts list will be helpful for those who wish to build one of these
phasers. However, this is not a step-by-step construction article, nor is it intended to be. We assume that
anyone who attempts to build MWPx will have considerable experience building other similar
MWPx Parts List
Component Value ELFA part # Other part #, remarks Qty
22 ohm
see remark below:
270 ohm
50 ohm
resistor, amp
4700 ohm
resistor, amp
1000 ohm
resistor, amp
10 ohm
resistor, amp
22 ohm
or 47 ohm
for all fixed resistors use 0.25 watt
metal film, Mouser part #: 271-
VALUE where VALUE is the
resistor value per Mouser table of
pot, amp
100 ohm,
25 turn, 0.5
64-347-16 (15T
0.25 watt, which
is adequate)
Mouser part #: 594-64Y101
ferrite bead
Amidon Part #: FB-61-101
1 uF
Mouser part #: 80-C330C105K5R
4500 pF
Mouser part #: 39-GM472K
2100 pF
Mouser part #: 39-GM222K
898282811.046.png 898282811.047.png 898282811.048.png 898282811.049.png 898282811.050.png 898282811.051.png 898282811.052.png 898282811.053.png 898282811.054.png 898282811.055.png 898282811.056.png 898282811.057.png 898282811.058.png 898282811.059.png 898282811.060.png 898282811.061.png 898282811.062.png 898282811.063.png 898282811.064.png 898282811.065.png 898282811.066.png 898282811.067.png 898282811.068.png 898282811.069.png 898282811.070.png 898282811.071.png 898282811.072.png 898282811.073.png 898282811.074.png 898282811.075.png 898282811.076.png 898282811.077.png 898282811.079.png 898282811.080.png 898282811.081.png 898282811.082.png 898282811.083.png 898282811.084.png 898282811.085.png 898282811.086.png 898282811.087.png 898282811.088.png 898282811.090.png 898282811.091.png 898282811.092.png 898282811.093.png 898282811.094.png 898282811.095.png 898282811.096.png 898282811.097.png 898282811.098.png 898282811.099.png 898282811.101.png 898282811.102.png 898282811.103.png 898282811.104.png 898282811.105.png 898282811.106.png 898282811.107.png 898282811.108.png 898282811.109.png 898282811.110.png 898282811.112.png 898282811.113.png 898282811.114.png 898282811.115.png 898282811.116.png 898282811.117.png 898282811.118.png 898282811.119.png 898282811.120.png 898282811.121.png
Component Value ELFA part # Other part #, remarks Qty
0.01 uF
Mouser part #: 39-GM103K
1000 pF
Mouser part #: 39-GM102K
500 pF
Mouser part #: 39-GM471K
220 pF
Mouser part #: 39-GM221K
inductor, amp
330 uH
Mouser part #: 542-77F331
inductor, amp
1 mH
Mouser part #: 434-02-102J
pot, phase
200 ohm
Newark part #: 01F6262
pot, vernier
50 ohm
Newark part #: 01F6259
rotary switch
2 pole, 6
Mouser part #: 690-D4C0206N
T106-1 or
58-723-11 (=
Amidon part #: T106-1or T106-2
2N5109 or
BHIAB Electronics, Sweden
(2N5109 only), or Mouser part #:
610-2n5109, or contact RF Parts
(USA) for MRF581A
BNC (f)
Mouser part #: 31-2221
toggle switch
Mouser part #: 611-7201-081
12VDC input
connector, 2.1
Mouser part #: 163-4302
gnd binding post
power indicator
No specific part numbers are listed
for this.
screws, nuts, flat
washers, split ring
lock washers,
ground lugs, etc.
I like stainless hardware, but plated
steel will do, and nickel plated brass
ground lugs. (DL) No specific part
numbers are listed for these.
#24 enameled
copper wire
No specific part numbers are listed
for this.
insulated stranded
I like #22 stranded silver plate with
Teflon insulation, but less fancy will
do. (DL) No specific part numbers
are listed for this.
220 ohm
Mouser 271-VALUE; see above
47 ohm
see above
898282811.002.png 898282811.003.png 898282811.004.png 898282811.005.png 898282811.006.png 898282811.007.png 898282811.008.png 898282811.009.png 898282811.010.png 898282811.012.png 898282811.013.png 898282811.014.png 898282811.015.png 898282811.016.png 898282811.017.png 898282811.018.png 898282811.019.png 898282811.020.png 898282811.021.png 898282811.023.png 898282811.024.png 898282811.025.png 898282811.026.png 898282811.027.png 898282811.028.png 898282811.029.png 898282811.030.png
Zgłoś jeśli naruszono regulamin