Christine Vollinger, Erk Jensen

Material characterization for the 18 MHz to 40 MHz

sweep-tuneable RF system

Measurement support by Fritz Caspers

2

RF tuning cavity with magnetic bias

Example for perpendicular bias (from Friedrichs, 1991)

Example for parallel bias (from Gardner, 1991)

3

From Talk of Erk in PS2 Review:Parallel plus perpendicular biasing possible?

• Smythe suggested in 1983 to use– transverse bias to get into saturation (lower losses)– add’l. parallel bias for tuning (less bias).

• Could this work? (I’d like to try!)

4

Leads to two primary goals:1. Characterization of ferrite samples with different saturation and

line widths under different bias conditions in the required f-range. (This required a dedicated test set-up, since supplier data does not exist – PART 1).

2. Verification of Smythe’s claims (PART 2).3. First requirement for an f-range of 18-40 MHz is a -range of

approx. 5 (since f 1 / Sqrt[] ).4. Ratio of maximum to minimum f is given by Sqrt[max/ min]; max

is the value at minimum bias field.

5

Measurement Set-up• Problematic: samples are available in sizes 1” x 1” (or smaller);• Small sample size led to strip-line set-up on ferrite; • 1-port frequency swept measurement with NWA chosen to obtain s11. From s11, the DUT impedance is determined and the µ can be calculated;• Measurement taken with magnetic bias field applied perpendicular to signal propagation;• Carries the assumption that the magnetic RF-field left/right of the strip-line can be neglected (there the bias field would be parallel).

Hbias

Coaxial connection

Ferrite

Strip-line

6

μ’ for Material #1 (Y36)

μ’-ratio:

μ’(20 MHz,300 G)/μ’(40 MHz, 1005 G)=4.0

μ’-ratio:

μ’(20 MHz,145 G)/μ’(40 MHz, 1005 G)=11.1

7

μ’’ for Material #1 (Y36)

μ’’-values:

μ’’(20 MHz,145 G)= 6.3μ’’(20 MHz,300 G)= 0.2μ’’(40 MHz,1005 G)= 0.074

8

μ’ for Material #2 (G300)

μ’-ratio:

μ’(20 MHz,145 G)/μ’(40 MHz, 1005 G)=9.0

μ’-ratio:

μ’(20 MHz,300 G)/μ’(40 MHz, 1005 G)=2.8

9

μ’’ for Material #2 (G300)

μ’’-values:

μ’’(20 MHz,145 G)= 10.1μ’’(20 MHz,300 G)= 0.12μ’’(40 MHz,1005 G)= 0.076

10

μ’ for Material #3 (G810)

μ’-ratio:

μ’(20 MHz,558 G)/μ’(40 MHz, 1008 G)= 5.8

μ’-ratio:

μ’(20 MHz,153 G)/μ’(40 MHz, 1008 G)= 8.5

11

μ’’ for Material #3 (G810)

μ’’-values:

μ’’(20 MHz,153 G)= 11.9μ’’(20 MHz,558 G)= 1.2μ’’(40 MHz,1008 G)= 0.03

12

μ’ for Material #4 (RG3)

μ’-ratio:

μ’(20 MHz,145 G)/μ’(40 MHz, 1004 G)= 12.5

μ’-ratio:

μ’(20 MHz,349 G)/μ’(40 MHz, 1004 G)= 3.14

13

μ’’ for Material #4 (RG3)

μ’’-values:

μ’’(20 MHz,145 G)= 17.01μ’’(20 MHz,349 G)= 0.1μ’’(40 MHz,1004 G)= 0.05

14

Summary (1/3)

• Four ferrite samples have been ordered from industry and measured in an external magnetic bias field. This allowed the determination of their available -range and the corresponding values of ’ as a function of frequency and Bias-field.• Requirement for an f-range of 2.2 from 18-40 MHz is a ’-ratio of approx. 5.

Y36 G300 G810 RG3

’-ratio, [145 – 1005 Gauss]

11.1(’’,145 G= 6.3)

(’’,1005 G = 0.074)

9.0(’’,145 G= 10.1)

(’’,1005 G = 0.076)

8.5(’’,153 G = 11.9)(’’,1005 G= 0.03)

12.5(’’,145 G = 17.01)(’’,1005 G= 0.05)

’-ratio, [300 – 1005 Gauss]

4.0(’’,300 G= 0.2)

(’’,1005 G = 0.074)

2.8(’’,300 G= 0.12)

(’’,1005 G = 0.076)

-- --

’-ratio, [558 – 1008 Gauss]

-- -- 5.8(’’,558 G = 1.2)

(’’,1005 G= 0.03)

--

’-ratio, [349– 1004 Gauss]

-- -- -- 3.14(’’,349 G = 0.1)

(’’,1004 G = 0.05)

15

Summary (2/3)

• Measurement set-up is found that appears robust and stable (1-port-measurement with NWA).• More ferrite samples from different suppliers are already ordered for measurement. • Next step will be to investigate potential non-linearity of the materials, if exposed to moderate power (IP3 measurement).

16

Leads to two primary goals:1. Characterization of ferrite samples with different saturation and

line widths under different bias conditions in the required f-range. (This required a dedicated test set-up, since supplier data does not exist – PART 1).

2. Verification of Smythe’s claims (PART 2).

17

Verification of Smythe’s Claims (PART 2)• Smythe suggested in 1983 to use

– transverse bias to get close to saturation (reduces losses)– additional parallel bias for tuning (less bias required for same μ-range).

Happlied

Measurement Set-upHapplied

Htune

Hbias

19

μ’ for Material #2 (G300) in 2-Directional Magnetic Bias

Change of 55 G (parallel) with 200 G perpendicular -> μ’-ratio > 6

μ’-ratio with Hperp=200G:

μ’(20 MHz,HparBias=18 G)/μ’(40 MHz, HparBias=200 G)= 6.4

20

μ’’ for Material #2 (G300) in 2-Directional Magnetic Bias

μ’’-values with Hperp=200G :

μ’’(20 MHz, Hpar=18 G) = 1.23μ’’(20 MHz, Hpar=73 G)= 0.106μ’’(40 MHz, Hpar=200 G)= 0.105

21

Contour-plots “Smythe” and G300 Material

RF field direction

Hpa

r/4

Ms

Hperp/4Ms

Plot of Smythe Measured Data

22

Summary (3/3)• The idea of Smythe to use a 2-directional magnetic bias field appeared in 1983, but (to my knowledge) has never been used or further investigated.• Smythe had no measurement data to support his claims.

• From the measurement done on the G300 sample, it appears as if the method is promising and could work; we will continue further tests with other samples and higher parallel bias fields.

G300 with perpendicular bias

G300 with 2-directional bias

’-ratio, [300 – 1005 Gauss]

4.0(’’,300 G= 0.2)

(’’,1005 G = 0.074)

--

’-ratio, [ 200 Gauss perpendicular] &[18 – 200 Gauss parallel]

-- 6.4(’’,18 G= 1.23)

(’’,200 G = 0.105)

23

I would like to thank Fritz Caspers for his strong and unfailing support with all kind of measurement problems!