# US 2022/0065927 A1 · Sub-sampled based instrument noise correction for jitter measurements

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2022/0065927 A1
- Document ID: US20220065927A1
- Status: Published application
- Kind: A1
- Application number: 17/009,480
- Filed: 2020-09-01
- Published: 2022-03-03
- Earliest claimed: 2020-09-01
- Assignees: Silicon Laboratories Inc.
- Applicants: Silicon Laboratories Inc.
- Inventors: Daniel de Godoy Peixoto; Gregory J. Richmond
- Web page: https://patentgrep.com/patent/US20220065927A1

## CPC

- G01R 31/31709: {Jitter measurements; Jitter generators (measuring jitter, noise figure or signal-to-noise ratio per se G01R29/26; analysis of tester signals G01R31/31901)}
- G01R 29/26: Measuring noise figure; Measuring signal-to-noise ratio
- G01R 31/31725: {Timing aspects, e.g. clock distribution, skew, propagation delay (for tester hardware G01R31/31937)}
- G01R: MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES (indicating correct tuning of resonant circuits H03J3/12)

## Abstract

A time error vector is determined using pairs of two closest points of input-referred noise data that straddle respective crossing times indicating when a clock signal representation crosses a threshold value, a slew rate of the clock signal representation, and the crossing times. A system filter is applied to the time error vector in the frequency domain. A first RMS value is determined indicating a jitter value present in the filtered time error vector. A raw clock signal time error vector of the clock signal under test is generated, the system filter is applied to the raw clock signal time error vector in the frequency domain, and a second RMS value indicating a jitter content of the filtered raw clock signal time error vector is determined. The second RMS value is corrected using the first RMS value to thereby generate a jitter measurement compensated for input-referred noise.

## Claims

### Claim 1

A non-volatile computer readable medium storing instructions to determine jitter in a clock signal under test that, when the instructions are executed by a processor, cause the processor to:
determine two points of input-referred noise data for a crossing time, a first point of the two points being before the crossing time and a second point of the two points being after the crossing time, the crossing time being a time that a clock signal representation crosses a threshold value, the clock signal representation corresponding to the clock signal under test; and
determine a first time error associated with the crossing time using the two points of the input-referred noise data, a clock slew rate of the clock signal representation, and the crossing time.

### Claim 2 (depends on 1)

The non-volatile computer readable medium as recited in claim 1 wherein the instructions further cause the processor to:
determine additional time errors for respective additional crossing times; and
determine a time error vector based on the first time error and the additional time errors.

### Claim 3 (depends on 2)

The non-volatile computer readable medium as recited in claim 2 wherein the respective additional crossing times are based on an average clock period of the clock signal under test.

### Claim 4 (depends on 2)

The non-volatile computer readable medium as recited in claim 2 wherein the respective additional crossing times are based on a nominal clock period of the clock signal under test.

### Claim 5 (depends on 1)

The non-volatile computer readable medium as recited in claim 1, wherein the first point is a closest sample of the input referred noise data before the crossing time and the second point is a closest sample of the input referred noise data after the crossing time.

### Claim 6 (depends on 2)

The non-volatile computer readable medium, as recited in claim 2, wherein the instructions further cause the processor to:
perform a transform to the time error vector to generate a frequency domain time error vector;
apply a system filter to the frequency domain time error vector to generate a filtered time error vector; and
determine a first root mean square (RMS) value of the filtered time error vector using the filtered time error vector.

### Claim 7 (depends on 6)

The non-volatile computer readable medium, as recited in claim 6, wherein the instructions further cause the processor to:
correct a second RMS value using the first RMS value, the second RMS value associated with jitter measurements of the clock signal under test to thereby determine a compensated jitter measurement of the clock signal under test.

### Claim 8 (depends on 7)

The non-volatile computer readable medium, as recited in claim 7, wherein the instructions further cause the processor to:
generate a frequency domain raw clock signal time error vector;
apply the system filter to the frequency domain raw clock signal time error vector to generate a filtered raw clock signal time error vector; and
determine the second RMS value using the filtered raw clock signal time error vector.

### Claim 9 (depends on 1)

The non-volatile computer readable medium, as recited in claim 1, wherein the instructions further cause the processor to:
determine the first time error as:
TIE
noise
⁡
[
n
]
=
(
P
1
⁡
[
n
]
+
(
P
2
⁡
[
n
]
-
P
1
⁡
[
n
]
)
×
SR
×
T
crossing
⁡
[
n
]
-
P
1
⁡
[
n
]
SR
×
T
period
+
P
2
⁡
[
n
]
-
P
1
⁡
[
n
]
)
SR
,
where P1[n] is the first point, P2[n] is the second point, Tperiod is a period of a sampling clock of an analog to digital converter used to generate the input-referred noise data, SR is the clock slew rate, Tcrossing[n] is the crossing time, and n is an index of the crossing time.

### Claim 10

A method of determining jitter in a clock signal under test comprising:
determining two points in input-referred noise data for each crossing time of a plurality of crossing times, a first point of the two points before the crossing time and a second point of the two points after the crossing time, the crossing times indicating when a clock signal representation corresponding to the clock signal under test crosses a threshold voltage value;
determining respective time errors associated with the crossing times using the two points of the input-referred noise data for each of the plurality of crossing times, a slew rate of the clock signal representation, and the crossing times; and
determining a time error vector based on the respective time errors.

### Claim 11 (depends on 10)

The method as recited in claim 10 further comprising:
grounding an input to a probe of an instrument used to measure the clock signal under test; and
generating the input-referred noise data with the input grounded.

### Claim 12 (depends on 10)

The method as recited in claim 10 wherein the crossing times are based on an average clock period of the clock signal under test.

### Claim 13 (depends on 10)

The method as recited in claim 10 wherein the crossing times are based on a nominal clock period of the clock signal under test.

### Claim 14 (depends on 10)

The method, as recited in claim 10, further comprising:
applying a transform to convert the time error vector to a frequency domain time error vector;
applying a first system filter to the frequency domain time error vector and generating a filtered time error vector; and
determining a first root mean square (RMS) value of the time domain representation of the filtered time error vector after the inverse transform.

### Claim 15 (depends on 14)

The method, as recited in claim 14, further comprising:
correct a second RMS value indicative of a jitter measurement of the clock signal under test using the first RMS value to generate a compensated jitter measurement of the clock signal under test.

### Claim 16 (depends on 14)

The method, as recited in claim 14, further comprising:
sampling the clock signal under test in clock signal measuring equipment that includes an analog to digital converter to convert the clock signal under test to sampled digital values.

### Claim 17 (depends on 15)

The method, as recited in claim 15, further comprising:
determining a raw jitter component of the clock signal under test as a raw jitter vector;
converting the raw jitter vector to a frequency domain raw jitter vector;
applying a second system filter to the frequency domain raw jitter vector to generate a filtered raw jitter vector; and
determining the second RMS value using the filtered raw jitter vector.

### Claim 18 (depends on 17)

The method, as recited in claim 17, wherein the first system filter and the second system filter are identical filters.

### Claim 19 (depends on 10)

The method as recited in claim 10,
wherein the first point is a closest point to the crossing time before the crossing time and the second point is a closest point after the crossing time.

### Claim 20 (depends on 10)

The method, as recited in claim 10, further comprising:
determining each of the time errors as:
TIE
noise
⁡
[
n
]
=
(
P
1
⁡
[
n
]
+
(
P
2
⁡
[
n
]
-
P
1
⁡
[
n
]
)
×
SR
×
T
crossing
⁡
[
n
]
-
P
1
⁡
[
n
]
SR
×
T
period
+
P
2
⁡
[
n
]
-
P
1
⁡
[
n
]
)
SR
,
where P1[n] is the first point, P2[n] is the second point, Tperiod is a period of a sampling clock of the analog to digital converter used to generate the input-referred noise data, SR is the slew rate of the clock signal representation, Tcrossing[n] is the crossing time, and n is an index of the crossing time.

### Claim 21

A method of correcting for jitter in a clock signal under test, comprising:
receiving input-referred noise data;
determining for respective crossing times, pairs of two closest points of the input-referred noise data that straddle each of the respective crossing times, the crossing times indicative of respective times a clock signal representation crosses a threshold value; and
determining a time error vector associated with the crossing times using the pairs of two closest points, a slew rate of the clock signal representation, and the crossing times;
performing a transform on the time error vector to generate a frequency domain time error vector;
applying a system filter to the frequency domain time error vector to generate a filtered time error vector;
determining a first root mean square (RMS) value indicating a jitter value present in the filtered time error vector;
generating a raw clock signal time error vector of the clock signal under test;
converting the raw clock signal time error vector to a frequency domain raw clock signal time error vector;
applying the system filter to the frequency domain raw clock signal time error vector to generate a filtered raw clock signal time error vector;
determining a second RMS value indicating a jitter content of the filtered raw clock signal time error vector; and
correcting the second RMS value using the first RMS value to thereby generate a compensated jitter measurement compensated for input-referred noise.
