SIGNAL MIXING IN FERRITES ON MICROWAVES

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP80T00246A007100270002-2
• Release Decision: RIPPUB
• Original Classification: K
• Document Page Count: 11
• Document Creation Date: December 22, 2016
• Document Release Date: July 15, 2009
• Sequence Number: 2
• Publication Date: August 3, 1959
• Content Type: REPORT

Approved For Release 2009/07/15: CIA-RDP80T00246AO07100270002-2
Next 1 Page(s) In Document Denied
Iq
Approved For Release 2009/07/15: CIA-RDP80T00246AO07100270002-2
 Jl I
Approved For Release 2009/07/15: CIA-RDP80T00246A007100270002-2
SIGNAL MIXING IN FERRITES ON MICROWAVES
by A.L. Mikaelyan and V.J. Antonyanz
In this paper the main formulae are derived for the mixing
of microwaves in ferrites, and the results of experiments are
described, which were carried out to check the theory.
Let us examine the case of a spheroid magnetized along the
z-axis, which is the axis of symmetry. If the ferrite is sub-
jected to high-frequency oscillations of frequencies U) , and
w2, which are close to each other, the difference frequency com-
pament formed from the Landau - Lifshita equation is as follows:
dM
dt
1P,
(ca, - i.)
~1 d v W, w ~,
oV N -~
(1)
where PI and P2 are the powers absorbed in the ferrite sample
at the frequencies W1 and W2 ;
V is the volume of the sample;
2aH is the width of the resonance curve.
If the difference frequency power is drawn off by a coil
wound on the ferrite sample, which is connected in the resonant
circuit, the voltage amplitude in the circuit is as follows:
ns d R pl ~' ~~~' ? rz,S CI-1~~
117- ~
d t o f1 J~}l Q V (~)
Here Q is the quality factor of the circuit (at the differ-
ence frequency);
hl is number of turns of the coil on the ferrite;
Approved For Release 2009/07/15: CIA-RDP80T00246A007100270002-2
 Approved For Release 2009/07/15: CIA-RDP80T00246AO07100270002-2
All quantities in this formula can be determined experi-
mentally.
'The test not up which was used allowed us to measure the
power absorbed by the ferrite at the frequencies CUo and W2, and
also the voltage at a difference frequency equal to 30 mo.
The ferrite sample is installed in a rectangular waveguids
at a distance of one half a wave length from the short-circuiting
plug, that is, where the magnetic field bunches.
The main purpose of the experiment was to compare the ex-
perimental values for the difference frequency with the theoretical
ones computed from formula (2). We used spherical samples, on
which two turns of wire (of 0.1 im dia.) were wound.
The ends of the wire were connected to a circuit tuned to
the difference frequency, which had a quality factor Q equal
to 15.
The plane of the turns was perpendicular to the direction
of the do magnetic field (fig. 1), while their position with
respect to the do field could be varied by turning dielectric
rod, on which the ferrite was attached.
The measurement results and also the calculated values
of the difference frequency signal for a manganese mono crystal
and yttrium garnets are given in table I. The quantity b in the
table gives the discrepancy between measured and calculated
values. These values compare well enough for monocrystals of
different diameters since the value of 6 lies within the limits
of measurement accuracy.
Approved For Release 2009/07/15: CIA-RDP80T00246AO07100270002-2 T - _
 Approved For Release 2009/07/15: CIA-RDP80T00246AO07100270002-2
-3-
Table 2 contains similar results for two types of poly-
crystalline ferrites. We see that the discrepancy between the
experimental and calculated values for the signals amounts to
20 decibels on the average for the type HM-2 ferrite and 30 deci-
bels for the type M-50 ferrite. This discrepancy cannot be attri-
buted to measurement errors however.
We believe this discrepancy between theory and experiment
may be explained if we consider the polycrystalline ferrite as a
first approximation to be a system of monocrystals coupled to
each other, where formula (t) is applicable to each of the latter.
In this case the quantity off in formula (I) will correspond
to the width of the absorption curve for a single monocrystal in
the polycrystalline sample.
Since the crystallographic axes of the individual monocrys-
tale are directed randomly in the sample, the do magnetic field
applied to the sazgale will not be the resonant field for all mono-
crystals. This is one of the main factors determining the large
width of the absorption line for polycrystalline ferrites.
The width of the absorption line for a polyorystalline sample
is determined by other factors, also (e.g. stresses, internal non-
uniformities in the magnetic moment, etc.) None the less, it
will not be narrower than the difference between the resonance
fields for the direction in which it is easy and hard to magne-
tise the sample.
Thus, formula (I) should be applied to each of the monocrys-
Approved For Release 2009/07/15: CIA-RDP80T00246AO07100270002-2
 Approved For Release 2009/07/15: CIA-RDP80T00246A007100270002-2
-4-
tale forming the polycrystal, and then all the signals should
be summed up.
In doing no, the quantity pH may be taken as a first appro-
ximation to be constant. Then, the summing process amounts to
determining the powers absorbed by the sample as a whole. This
quantity is just what was determined in the experiments (see
table 2).
Thus, we come to the conclusion, that for polycrystalline
samples we should use in formula/ (I)/and consequently in for-
mula (2)/the unknown quantity AH, which corresponds to the
width of the resonance curve for a single monocrystal.
If this reasoning is true, the value of o H may be deter-
mined from measurements of the difference frequency signal.
Using formula (2) and data of table 2 we obtain:
for the type HK-2 ferrite 2 AH - 32 oersteds and for the type
14-50 ferrite 2A H - 14.5 oersteds. It is entirely possible that
such a method of determining ?H for different monocrystals
may be used in investigating ferrites.
Another series of experiments were carried out to investi-
gate certain particularities of mixing phenomena in ferrite*.
Only some of the results obtained are given below.
Curves for the signal and absorption in ferrite are given
in fig. 2 as a function of the dc magnetizing field for differ-
ent diameters of the ferrite rods. The latter were placed along
the axis of a square waveguide, and were magnetized by a longi-
tudinal field (see the drawing).
As we see from the drawing, the maximums for the ferrite
Approved For Release 2009/07/15: CIA-RDP80T00246A007100270002-2
 Approved For Release 2009/07/15: CIA-RDP80T00246AO07100270002-2
-5-
absorption curves shift towards stronger magnetic fields as
the diameter of the rod is increased (its length is constant
and equal to 10 mm). This is connected with changes in the
demagnetizing factors. Here, the maximum of the signal coin-
cides with the corresponding maximums for absorption in ferrite.
This confirms that formula (2) is correct. It follows from
this formula that the dc field only influences the magnitude
of the signal through the power absorbed in the ferrite (this
is true for polycrystalline as well as for monocrystalline
f errites) .
As the rod diameter is increased above a certain value,
the signal ceases to increase and then at a diameter of 2 mm it
even begins to decrease. This is associated with the fact that
a rod 2 ma in diameter already absorbs almost all of the power;
further increase in the diameter only iapairs the structure of
the field in the ferrite, and results in wider absorption curves
and, consequently, in wider signal curves.
A decrease in the a H of the material results in the
appearance of additional peaks in the absorption and the
signal. The peaks increase in number as 0 H decreases and the
ferrite diameter increases. This can be clearly seen from
fig. 3, which presents curves for the type HM--2 polycrystalline
ferrite (2AH - 320 oersteds), and from fig. 4, which is for mono-
crystalline ferrite (24H - 60 oersteds).
Similar phenomena are observed when ferrite spheres are
placed in a resonator, which resonated at both frequencies.
Approved For Release 2009/07/15: CIA-RDP80T00246AO07100270002-2 -,-- ?
 Approved For Release 2009/07/15: CIA-RDP80T00246AO07100270002-2
-6-
Finally, let us examine the losses involved with one possible
design of a ferrite mixer. It is connected in a balanced circuit
as shown on fig. 5. A heterodyne is connected to the input channels
of the short-slot hybrid junction and a signal is applied. Two
short-circuited cnctions are connected to the output channels.
They have type Y1-50 ferrite rods 3 nu in diameter and 10 mm
long. The conversion loss amounted to 58 decibels for a hetero-
dyne power of 50 milliwatts. In order to reduce this figure to
8 - 10 decibels (as in the crystal mixer), it is necessary to in-
crease the heterodyne power by 48 to 50 decibels, that is, up
to 3 - 5 kw. However, regeneration will appear even at much lower
power levels. This will most likely strengthen the mixing effect
and, consequently, will reduce the required heterodyne power.
In closing we should like to state that similar eaperimsnts
were carried out at shorter and longer wave lengths, however, no
new phenomena were discovered.
Approved For Release 2009/07/15: CIA-RDP80T00246AO07100270002-2
 xi:n++y~sw .r+.~e.....,~irs^errY _:._...r... , -:._.'.-. -.F - _.T._
Approved For Release 2009/07/15: CIA-RDP80T00246AO07100270002-2
Xeeictzdc suppatt
Approved For Release 2009/07/15: CIA-RDP80T00246AO07100270002-2 -
 Approved For Release 2009/07/15: CIA-RDP80T00246AO07100270002-2
pt f f e tepee f tequenc y
The
output
Fttti,te
omelet
Of sphtze:-
2pH
p
mtasuzed
caecueated
(H0 = rtes)
f zom (1)
mm;
(Q)
(d a)
(de,
(ds)
0179
51cle
2,4
619
5,5
* 14
SLnygee
l
f
1, X77
42ot
12,0
16, 1
18,5
- 2,4
s of
a
cttt
ese
1,27
58a
15,5
17,3
16,4
+0,9
mangan
ite
',4
Ma
25,0
21,1
-
20
1,1
fett
2,15
600e
41,3
20,3
20,2
+03
Singer ctyst
0,5
Tale
1,5
10,3
16, 3
-6
of yttz4um
0,85
12Q
3,6
2366
20,6
3,2
i.ton yaznet
1,28
26 at
14,5
30
25
*5
Tu6ee 2
The
mtt
J) f f etence f tequency
output
7ezzlte
ez
Of sphezes
mm)
2AH
(0e)
p
tin
- aes~
o H -H
%neasu t ed
(da)
caecu~at,eu
1
db,
(td:
1,4
55
10,'
-77
11 2
1,73
8L0 _
13,3
-6,9
1rM
-2800
2,12
3200e
14,8
18
- 3,2
*
S
25
200
.
---
19,8
-1,6
2,8
=28,5
21
0
21
1.31
3
21
-10
+31
1j51
63
23,2
-6
---
+29,2-
M-30
1,65
7,8
25
-4 9
#2R9
-'-'
1,92
4600e
11,3
27,4__
-26
#30
47rMj-0100
2,16
14,0
29,1
-23
*31,4
242
,
18
308
-1,6
#32,6
2,76
22,5
32,5
0
+3165
OZ
313
335
19
+316
Approved For Release 2009/07/15: CIA-RDP80T00246AO07100270002-2
 Approved For Release 2009/07/15: CIA-RDP80T00246AO07100270002-2
W-j
Fy z
Approved For Release 2009/07/15: CIA-RDP80T00246AO07100270002-2
 oMtpit
Atte rttim
Approved For Release 2009/07/15: CIA-RDP80T00246AO07100270002-2
Fig
Singee czystae feizite
(26H = 6Oot)
16
'I,
14
1
if~
12
i ~~ ~
L
L-
0 300
S ignae
Lot A, : ~-O ~.- I I
f zom 9etezodyne
I
Atttnuat~on
re t zite sampees
Approved For Release 2009/07/15: CIA-RDP80T00246AO07100270002-2
Dgf.UM f