Probe characterization v5 web

Probe Characterization
Amber Precision Instruments, Inc. (API)
Jul. 23, 2014

Our EMI Probes
1
EMI Probes:
- Up to 6 GHz
- Up to 20 GHz
- Up to 40 GHz
Optional EMI Probes;
Choose:
- Size
- Frequency range
- Field Component
* EMI Hx 2 mm

Characterization Structure
up to 15 GHz
2
50 Ohms Microstrip Line (MSL)

Characterization Structure
up to 40 GHz
3
50 Ohms Air-Dielectric* Coplanar Waveguide (CPW)
* To avoid the effect of non-TEM waves on probe characterization at higher freqs (>15 GHz).

Probe Characterization Setup
4

Typical EMI Probe S21
5
1 2 3 4 5 6 7 8 9 10
x 10
9
-30
-35
-40
-45
-50
-55
-60
-65
Frequency [Hz]
|S21| [dB]
* EMI Hx 2 mm: up to 10 GHz

What Are the Specs?
6
Log freq: Low freq with 20 dB/dec slope
10
8
10
9
10
10
-30
-35
-40
-45
-50
-55
-60
-65
Frequency [Hz]
|S21| [dB]
* EMI Hx 2 mm

What Are the Specs?
7
Log freq: Low freq with 20 dB/dec slope
-30
-35
-40
-45
-50
-55
-60
* EMI Hx 2 mm
10
8
10
9
10
10
-65
Frequency [Hz]
|S21| [dB]
20 dB/dec Line

What Are the Specs?
8
High freq: How high the probe measures?
0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
x 10
10
-30
-35
-40
-45
-50
-55
-60
-65
Frequency [Hz]
|S21| [dB]
* EMI Hx 2 mm

What Are the Specs?
9
Unwanted field: Decoupling of unwanted components
10
8
10
9
10
10
-30
-40
-50
-60
-70
-80
-90
-100
Frequency [Hz]
|S21| [dB]
Hx Field,  = 0
Unwanted Field,  = 90
* EMI Hx 2 mm

What Are the Specs?
10
Low freq: How low the probe measures
8
10
10
9
10
10
-30
-35
-40
-45
-50
-55
-60
-65
Frequency [Hz]
|S21| [dB]
in a “practical” setup?

What Are the Specs?
11
10
7
10
8
10
9
10
10
-30
-40
-50
-60
-70
-80
-90
Frequency [Hz]
|S21| [dB]

What Are the Specs?
12
10
7
10
8
10
9
10
10
-30
-40
-50
-60
-70
-80
-90
Frequency [Hz]
|S21| [dB]
60 dB dynamic range

What Are the Specs?
13
10
7
10
8
10
9
10
10
-30
-40
-50
-60
-70
-80
-90
Frequency [Hz]
|S21| [dB]
20 dB/dec slope
60 dB dynamic range
4 MHz

High Freq Probe
14
10
9
10
10
-30
-40
-50
-60
-70
-80
-90
Frequency [Hz]
|S21| [dB]
Hx Field,  = 0
Unwanted Field,  = 90
* EMI Hx 1 mm L: up to 20 GHz

Probe Factor
Probe factor: Provide probe factor
SmartScan imports probe factor
SmartScan measures and calculates system factor
10
8
Measured Probe Factor
Theoretical Open-Circuit Probe Factor
10
9
10
10
50
40
30
20
Frequency [Hz]
|PF| [dB(A/m)/V]
* EMI Hx 2 mm 15

Other Methods of
Data Presentation
• How other people characterize their probes?
• How the characterization data is presented?
• Is their data presentation fair?
16

Example* of 10 kHz (?) to 6 GHz
Probe Data Presentation
17
Let’s analyze the data!
Extract
* Aprel: http://media.wix.com/ugd/36d46a_46d2902bfdd24c25b401582e6a488370.pdf

Is Freq Scale OK?
18

Is Freq Scale OK?
19
Why the freq axis of
extracted data don’t
match?!
Let’s see!

Is Freq Scale OK?
20
Why all ticks are
0.99999 GHz 1.3 GHz 1.3 GHz 1.3 GHz 1.1 GHz
equal?!
Well, not a big deal.
Maybe just a typo!

Lower Frequencies
21
Why all ticks are
equal?!
Maybe just a typo!

Lower Frequencies
22
Why all ticks are
equal?!
Maybe just a typo!
What happened to
lower freqs?!
Is it extrapolated or
really measured?
What is the measurement
instrument setup (IF BW,
span, no. of points, etc.) to
measure down to 10 kHz?
Is it a “practical” setup?

Plot It Differently
23
4
-40
-60
-80
-100
10
10
5
10
6
10
7
10
8
10
9
-120
Frequency [Hz]
|S21| [dB]
What if we plot it
in a different way
(log freq)?

Plot It Differently
24
What if we plot it
in a different way
(log freq)?
-40
-60
-80
-100
10
4
10
5
10
6
10
7
10
8
10
9
-120
Frequency [Hz]
|S21| [dB]
20 dB/dec Line
The probe is NOT
purely inductive
at lower freqs
(as claimed).
Is it probe issue?
Or, measurement
issue?
Or, simply not
extrapolated the
right way?
Note: Simple extrapolation
(even if it is done the right
way) does NOT mean probe
can measure down to that
freq!

How about 10 kHz to
20 GHz (?) Probe?
25
Extract

Flat Response (?)
26

Flat Response (?)
27
The data is from -30 dB
to -60 dB. Why the scale
is from 0 to -180 dB?!
Well, it looks better* this
way! :-)
*“flat response across
the frequency range”

Plot It Differently
28
0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
x 10
10
-30
-35
-40
-45
-50
Frequency [Hz]
|S21| [dB]
What if we plot it
in a different way
(tight scale)?

Plot It Differently
29
What if we plot it
in a different way
(log freq)?
8
-30
-35
-40
-45
10
10
9
10
10
-50
Frequency [Hz]
|S21| [dB]

Plot It Differently
30
What if we plot it
in a different way
(log freq)?
8
-30
-35
-40
-45
10
10
9
10
10
-50
Frequency [Hz]
|S21| [dB]
20 dB/dec Line
The lower freqs seems
to be fine for this prob.
The deviation might be
just data extraction
error.

8
-30
-35
-40
-45
10
10
9
10
10
-50
Frequency [Hz]
|S21| [dB]
20 dB/dec Line
Higher Frequencies
31
How about the
higher freqs?
There is about
20 dB notch @
13 GHz!
A deep notch is usually
a BIG red flag for
“unwanted field” coupling!

Higher Frequencies
32
The question is:
How to make sure this probe
measures the “desired field”
up to 20 GHz (as claimed)?
The “unwanted field” decoupling
data of probe need to be provided.
This probe is suspected
to work <10 GHz,
unless the rejection info
is available!

10 kHz to 40 GHz Probe
33
Extract

Flat Response (?)
34

Flat Response (?)
35
Same question:
The data is from -35 dB
to -55 dB. Why the scale
is from 0 to -120 dB?!

Plot It Differently
Single Probe Solution (?)
36
What if we plot it
in a different way
(tight scale and
log freq)?
10
10.2
10
10.3
10
10.4
10
10.5
10.6
10
-35
-40
-45
-50
Frequency [Hz]
|S21| [dB]

Plot It Differently
Single Probe Solution (?)
37
What if we plot it
in a different way
(tight scale and
log freq)?
10
10.2
10
10.3
10
10.4
10
10.5
10.6
10
-35
-40
-45
-50
Frequency [Hz]
|S21| [dB]
This info is NOT sufficient
to evaluate the probe up
to 40 GHz!
The “unwanted field” decoupling
data of probe need to be provided.
What does happen to lower freqs?
Does this probe have sufficient
sensitivity to measure down to 10
kHz?!

Another Example* of Data
Presentation
* Detectus AB: http://www.detectus.se/products_nearfieldprobes.html 38

Another Example* of Data
Presentation
Again, the data is from -30 dB
to -60 dB. Why the scale is
from 0 to -60 dB?!
Without unwanted field data,
there is no sufficient info to
evaluate the probe.
Log freq scale is required for
low freq evaluation.
* Detectus AB: http://www.detectus.se/products_nearfieldprobes.html 39

What to Look for in
Probe Characterization Data?
• Unwanted field rejection data need to be provided to
determine higher usable freq.
• Is lower usable freq achievable in a “practical” setup?
• Is the probe purely inductive/capacitive at lower freqs?
• Is the characterization structure (MSL, CPW, etc.) suitable for
freq of interest?
• What is the measurement setup/condition?
• Is the scale proportional to data range?
40

41
Questions?
Contact us: amberpi@amberpi.com
www.amberpi.com

Probe characterization v5 web

More Related Content

What's hot (8)

Viewers also liked (7)

Similar to Probe characterization v5 web (20)

Recently uploaded (20)

Probe characterization v5 web